YT YouTube2Text

YouTube2Text → Transcripts → Six Sigma Full Course in 7 Hours | Six Sigma Green Belt Training | Six Sigma Training | Simplilearn

Six Sigma Full Course in 7 Hours | Six Sigma Green Belt Training | Six Sigma Training | Simplilearn — Transcript

by Simplilearn · 56,888 words · 10,288 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00[Music]
0:04hello guys welcome to this six sigma
0:06full course video by simply learn in
0:09this video we'll cover everything you'll
0:11need to know about six sigma and cover
0:13the major concepts of lean six sigma we
0:16will start out with an animated video
0:18covering each of these concepts and go
0:20into more detail on the dmeic process we
0:23will also cover the different six sigma
0:25builds and much more so without further
0:27ado let's jump right into it before we
0:30do that don't forget to get subscribed
0:31to our youtube channel and don't forget
0:33to hit that bell icon so that you can
0:35never miss an update from simply learn
0:37now over to our training experts imagine
0:40you've been tasked with a really
0:41important project at work the company
0:44you're working for produces luxury cars
0:47the production numbers are going down
0:49and a lesser number of cars are getting
0:50manufactured each day there also seems
0:53to be an issue with the quality of the
0:54windshield wipers that go on these cars
0:57the question you are faced with is there
0:59a way for the company to stop the stall
1:01in production and increase the
1:03production per day from 1 000 to 2 000
1:07also is there a way to find out what's
1:10causing the drop in the wiper quality
1:12there is
1:13six sigma
1:15six sigma gives you the tools and
1:17techniques to determine what's making
1:18the manufacturing process slow down how
1:21you can eliminate the delays improve the
1:23process and fix further issues along the
1:26way the concept was introduced in 1980
1:29by bill smith while working for motorola
1:32since then six sigma has seen worldwide
1:35adoption
1:36six sigma aims to reduce the time
1:39defects and variability experienced by
1:42processes in an organization thanks to
1:44six sigma you can produce a defect-free
1:47product
1:4999.9996 percent of the time allowing
1:52only 3.4 errors per 1 million
1:54opportunities
1:56six sigma also increases customer
1:58loyalty towards the brand improves
2:01employee morale leading to higher
2:03productivity
2:04six sigma has two major methodologies
2:07dmaic and dmadv
2:10let's look at the first methodology
2:13dmaic is an acronym for define measure
2:17analyze improve and control
2:20let's have a look at each of these
2:22stages individually and how it relates
2:24to your earlier problem
2:26in the define phase you determine what
2:28issues you're facing what your
2:30opportunities for improvement are and
2:33what the customer requires of you
2:35here you look at the process as a whole
2:37and determine the issues with the
2:39manufacturing process
2:41in this case finding out why the cars
2:43had varying windshield wiper quality and
2:46how to optimize the current process to
2:48manufacture more cars in the measure
2:50phase you determine how the process is
2:53performing currently in its unaltered
2:55state you determine the current number
2:57of cars that are manufactured in a day
3:00in the current scenario 1000 cars are
3:02manufactured in a day and each of these
3:05cars are outfitted with a pair of
3:06windshield wipers by one of 30 machines
3:09used some of the metrics measured are
3:11how many cars are produced in a day time
3:13taken to assemble a car how many
3:16windshield wipers were attached in a day
3:18time that takes them to do so defects
3:21detected from each machine on assembly
3:23completion and so on
3:25following this in the analyze phase you
3:28determine what caused the defect or
3:30variation
3:31on analyzing previous data you find out
3:34that one of the machines that installed
3:36the windshield wiper was not performing
3:38as well as it was supposed to production
3:41was taking longer since the car chassis
3:43was being moved across the different
3:45locations slower as cranes had to
3:47individually pick and drop the frame
3:49this was because the wheels were
3:51attached to the car only in the last
3:53stage
3:55next in the improve phase you make
3:57changes to the manufacturing process and
3:59ensure the defects are addressed
4:02you replaced the faulty machines that
4:04installed the windshield wiper with
4:05another one you also find a way to save
4:08time by attaching wheels on the frame in
4:10the initial stages of the manufacturing
4:12process unlike how it was done earlier
4:15now the car can be moved across the
4:17assembly area faster and finally in the
4:20control phase you make regular
4:22adjustments to control new processes and
4:25future performance based on the changes
4:27made the company was able to reduce
4:29production time and manufacture about 2
4:32000 cars a day with a higher quality of
4:35output dmaic is one of the most commonly
4:38used methodologies in the world it
4:40focuses on improving the existing
4:42products of the organization
4:44the second methodology is dmadv which is
4:48short for define measure analyze design
4:52and verify
4:53it is used when the company has to
4:55create a new product or service from
4:57scratch it is also called dfss or design
5:01for six sigma
5:02let's take the scenario where the
5:04company decides to build a new model a
5:07sports car
5:08in the define phase you define the
5:10requirements of the customer based on
5:12inputs from customers historical data
5:15industry research you determine what you
5:17need to ensure your car becomes a
5:19success the data collected indicates
5:21customers are drawn to cars which can
5:23achieve more than 150 miles per hour
5:27customers are also more inclined towards
5:29cars which have v6 engines and an
5:31aerodynamic frame
5:33then in the measure phase you use the
5:36customer's requirements to create a
5:37specification
5:39this specification helps define the
5:41product in a measurable method so that
5:43data can be collected and compared with
5:45specific requirements
5:47some of the major specifications that
5:49you focus on are the top speed engine
5:52type and type of frame
5:54in the analyze phase you analyze the
5:57product to determine whether there are
5:58better ways to achieve the desired
6:00results areas of improvement are
6:03determined and tested
6:05based on the analysis of the prototype
6:07created in this phase you find that the
6:09product satisfies just about all of the
6:11customer requirements except the top
6:13speed
6:15so research begins on an aluminum alloy
6:18that could possibly meet the speed
6:19requirements of the customer following
6:22this the design phase based on the
6:24learnings from the analysis phase the
6:26new process or product is designed
6:29revisions are made to the model and the
6:31car is manufactured with the new
6:32material the analysis phase is repeated
6:35based on the new design you also bring a
6:38focus group and see how they receive it
6:40based on their feedback further changes
6:43are made and finally in the verify phase
6:46you check whether the end result meets
6:48or exceeds customer requirements
6:51once you launch your brand new sports
6:53card you collect customer feedback and
6:56incorporate it into future designs
6:58and guess what your customers are loving
7:02the new design and that is dmadv for you
7:06six sigma has also found success in a
7:09number of different industries the
7:11petrochemical healthcare banking
7:14government and software are some of the
7:16industries that have utilized the
7:18concepts of six sigma to achieve their
7:21business goals
7:22another commonly used methodology
7:24adopted by companies around the world is
7:26lean lean is a methodology that aims to
7:29remove any part of the process that does
7:31not bring value to the customer it means
7:34doing more with less while doing it
7:37better the philosophy behind lean comes
7:39from the japanese manufacturing industry
7:41by bob hartman who at the time was part
7:43of toyota
7:45since then across the world services and
7:47manufacturing organizations have
7:49incorporated lean within their
7:51businesses
7:52but what if you could have the best of
7:54both worlds a combination of both six
7:57sigma and length
7:59that's lean six sigma before we get into
8:02lean six sigma we've got a quiz for you
8:05what methodology is used by companies
8:08when they want to release a new product
8:10created from scratch a
8:12design measure analyze define verify
8:16b
8:17define measure analyze design verify
8:22c
8:23demand measure analyze improve control
8:28d
8:29define measure analyze improve control
8:33leave your answers in the comments
8:35section below for a chance to be one of
8:37three people to win an amazon voucher
8:40now let's get back to lean six sigma
8:43lean six sigma is a methodology that
8:45focuses on eliminating problems removing
8:48inefficiencies and waste while improving
8:50the working conditions to ensure the
8:52customer's needs are better satisfied it
8:55combines the tools methods and
8:57principles of lean and six sigma
8:59we'll have another video detailing the
9:01process of lean six sigma very soon
9:04imagine you're the manager of a
9:06supermarket chain
9:07you've noticed that two things need your
9:09immediate attention
9:11the first issue is how to handle the
9:13different kinds of waste that you
9:14encounter at your supermarket
9:17the next one requires you to address the
9:19supply chain issues at the supermarket
9:21which are causing delays to the morning
9:23delivery of milk
9:25leading to customer dissatisfaction and
9:27attrition
9:28these problems can be solved by
9:30incorporating two of the most popular
9:32quality management methodologies in the
9:34world lean and six sigma
9:37one famous for its ability to handle
9:39waste and another known for process
9:41improvement
9:42but what if there was a methodology that
9:45combined the concepts of both six sigma
9:48and lean
9:49one that could solve all your issues
9:52well there is
9:53lean six sigma
9:55before we dive into lean six sigma let's
9:58take a closer look at its parent
10:00methodologies
10:01first off lean is a methodology that
10:04focuses on providing value to the
10:06customer
10:07eliminating waste
10:09continuous improvement
10:11reducing cycle time lean and six sigma
10:15both aim to handle waste
10:17but what is this waste
10:19waste is any step or action in the
10:21process that a user does not gain any
10:24value from in short things that users
10:27wouldn't want to pay for
10:29why would a consumer want to pay extra
10:31for the additional truck that was
10:32required to deliver milk to the
10:34supermarket just because the other one
10:36broke down
10:38this waste can be divided into eight
10:40categories let's have a look at each of
10:42them
10:42one transportation this waste refers to
10:46the excess movement of people tools
10:48inventory equipment and other components
10:51of a process than it is required
10:54two inventory
10:55this waste occurs due to having more
10:57products and materials than required
11:00this can cause damage and effects to
11:02products or materials greater time for
11:04completion inefficient allocation of
11:06capital and so on
11:09three
11:10motion
11:11this refers to the time and effort
11:13wasted due to unnecessary movement of
11:15people equipment or machinery
11:18this could be sitting through inventory
11:20double data entry and so on
11:234. waiting
11:25this can be time wasted waiting on
11:27information instructions materials or
11:30equipment
11:315. over-production
11:34this is the waste created due to
11:36producing more products than required
11:396.
11:41it refers to more work more components
11:44or more steps in a product or service
11:46than required
11:48seven defects
11:50this is the waste originating from a
11:52product or service that fails to meet
11:54customer expectations
11:568. skills
11:58this waste refers to the waste of human
12:00potential under utilizing capabilities
12:03and delegating tasks to people with
12:05inadequate training
12:07for years now many systems have emerged
12:09that use the lean methodology to
12:11identify and handle the different kinds
12:14of waste
12:15some of the more popular and effective
12:16ones are
12:18jit or just in time
12:205s and
12:22kanban the jit methodology focuses on
12:26reducing the amount of time the
12:27production system takes to provide an
12:29output and the response time from
12:31suppliers to customers
12:335s is another methodology that focuses
12:36on cleanliness in organization while
12:38improving profits and efficiency
12:41kanban is also another popular
12:43methodology to achieve lean
12:46it is a visual method to manage tasks
12:48and workflows
12:49kanban enables users visualize the
12:52workflow to identify issues in the
12:54process and fix them
12:56these methodologies help in optimizing
12:58the waste production and are often used
13:01together to maximize results
13:03so that's the first problem solved
13:06now let's have a look at how you can
13:08improve the supermarket supply chain
13:10efficiency
13:11for that let's have a look at the other
13:13part of lean six sigma six sigma
13:16six sigma is a set of tools and
13:18techniques that are used for process
13:20improvement and removing defects
13:22let's see how six sigma makes that
13:25possible
13:26six sigma has two major methodologies
13:29dmaic and dmadv
13:32you can learn more about these two
13:34methodologies by checking out our six
13:36sigma in nine minutes video by clicking
13:38on the top right corner
13:40let's have a closer look at dmaic since
13:43lean six sigma uses the dmaic
13:46methodology of six sigma dmaic is an
13:49acronym for define measure analyze
13:53improve control
13:55it is used to improve existing products
13:57and processes so that it can meet the
13:59customer's requirements
14:01in the define phase you determine what
14:03the goals of the project are
14:05in this case you want to reduce the
14:07amount of time taken to deliver milk
14:09from the warehouse to the supermarket so
14:12that it is stocked on the supermarket
14:13shelves before 8 30 in the morning
14:16in the measure phase you measure the
14:18performance of the current unaltered
14:20process
14:21the milk truck leaves at 7 30 a.m in the
14:23morning and can take one of three routes
14:26a b and c
14:29route a is currently the preferred one
14:31as it takes only 60 minutes to reach the
14:33supermarket compared to the routes b and
14:36c which takes 70 and 80 minutes
14:38respectively
14:40in the analyze phase you find out why
14:42the defects exist since routes b and c
14:45were school bus routes
14:47by reducing the starting time by one
14:49hour at 6 30 instead of 7 30 meant
14:52avoiding the traffic routes b and c now
14:54take 40 to 45 minutes to reach the
14:56supermarket route a still takes the milk
14:59truck one hour to get to the supermarket
15:01even when the truck leaves at 6 30 a.m
15:04in the improved phase performance can be
15:06improved by addressing and eliminating
15:08the root causes
15:10now that you've realized that advancing
15:12the milk pickup by an hour and changing
15:14the route to route b can save time you
15:17change the process accordingly
15:19providing your workers with ample time
15:21to stock the milk into the shelves
15:23before the morning rush
15:24and finally in the control phase you
15:28make regular adjustments to control new
15:30processes and future performance
15:32you continue to monitor the delivery
15:34times and try out alternate routes to
15:36continually improve the process and
15:38ensure even faster delivery
15:41this process change led to reduced man
15:44hours and cost enhanced sales and
15:47customer retention
15:48the lean six sigma methodology offers
15:50many such benefits to businesses
15:53let's take a look at some of them
15:55one increase in profits
15:57two standardized and simplified process
16:01three reduced errors
16:03four employee development
16:06five
16:07value to customers
16:08and that is lean six sigma for you
16:11six sigma is a set of tools and
16:13techniques that have helped several
16:14companies around the world achieve
16:16business success hi guys i'm raut from
16:18simply learn and let's get started with
16:20our introduction to six sigma now let's
16:22understand this better with an example
16:24here let's talk about how things were
16:26before six sigma was introduced here
16:28jenny and james are having a
16:29conversation with each other jenny is
16:31james's manager and she's not happy at
16:33all she says james is in a lot of
16:35trouble this is because she found out
16:37that the customers weren't happy with
16:39the organization's service and the
16:40operational costs were way too high and
16:43as manager james had to make sure that
16:45this did not happen now let's have a
16:47look at the same scenario in present day
16:48here we have jenny congratulating james
16:51she's very impressed with his work but
16:52james says it's all thanks to six sigma
16:55methodology so jenny asks she wants to
16:57know more about six sigma so to
16:59understand six sigma here's what you
17:01need to know firstly we'll have to
17:02understand what is six sigma what its
17:04advantages are some of its methodologies
17:07what are the different roles in a six
17:08sigma team what is lean what is a lean
17:11process and what is lean six sigma so
17:13now let's get started with understanding
17:15what exactly is six sigma the six sigma
17:17methodology makes sure to find as well
17:20as eliminate any sort of defect or
17:22variation that could be affecting your
17:24product service or process now this
17:26methodology is statistics based is data
17:28driven and focused on continuous
17:30improvement now this means that there's
17:32no end goal in the horizon there's
17:34always another goal to reach there are
17:36three core ideologies behind six sigma
17:39the first one states that for any
17:40business to be successful there's
17:42continuous efforts that are required so
17:44that you can achieve stable as well as
17:46predictable process results the second
17:48ideology states that in any business or
17:50manufacturing process there are certain
17:52characteristics that can be defined
17:54measured analyzed and controlled the
17:56final ideology says that along with the
17:58rest of the organization the top level
18:00management plays a very important role
18:02to making sure that there's sustained
18:04quality now let's talk about the
18:06advantages of six sigma six sigma can
18:09help produce a road map or a path
18:11through which you can easily find and
18:13reduce any sort of organizational risk
18:15and reduce the operational costs another
18:18advantage is that it helps improve the
18:20efficiency of the process and making
18:22sure that it works in a timely manner it
18:24decreases defects improves the overall
18:26tracking and monitoring process and
18:28ensures that the products are aligned
18:30with the company's policies it is also
18:32reported that it greatly helps improve
18:34customer as well as vendor satisfaction
18:36it helps improve the cash flow and
18:38ensures that the products are complying
18:40with the regulations of the organization
18:43now let me tell you about the process of
18:45six sigma now six sigma projects are of
18:47basically two methodologies the dma ic
18:50and the dma dv now let's talk about
18:52dmaic in detail that's short for define
18:56measure analyze improve and control this
18:59is one of the most commonly used
19:00methodologies in the world this is
19:02commonly used by companies when they
19:04have to fix or improve an already
19:06existing product or process that does
19:08not meet the company's standards now
19:10let's have a look at the process first
19:11phase is the define phase in this phase
19:14you define the problem that the
19:15customers are facing you find out where
19:17you can improve and you clearly
19:19understand what the customers require of
19:21you the second phase is the measure
19:23phase now in this phase you actually
19:25identify how well the process is doing
19:28in its current unaltered state in the
19:31analyze phase you process the data that
19:33you get from the measure phase and
19:34determine what exactly is the cause of
19:36the delay or variation in the improve
19:39phase you start by making small changes
19:41to the business process and make sure
19:43that the problem you identified earlier
19:45is being taken care of and finally in
19:47the control phase you control the new
19:49process so that it doesn't go wrong and
19:52use the same knowledge for future
19:53processes now let's have a look at dmadv
19:57this is short for define measure analyze
20:00design and verify now this is also
20:02commonly known as dfss or the design for
20:04six sigma now this is commonly used by
20:06companies around the world when they
20:08have a new product that needs to be
20:09created all the way from scratch in the
20:12first phase which is the define phase
20:14you define what the goal of the project
20:16is and what the customers require of you
20:18in the measure phase you measure as well
20:20as determine what the customer needs and
20:22how they respond to your products in the
20:25analyze phase you perform analyses to
20:27determine how you can improve your
20:28product or service so they can better
20:30serve your customers in the design phase
20:32you set up process details and make
20:35optimizations to the design to make sure
20:37your customer is satisfied and finally
20:39in the verify phase you check how well
20:41the design is working out and how well
20:43it meets the customer's needs now before
20:46we go on let's talk about how six sigma
20:48was used in reference to the earlier
20:50example the situation that james was
20:52facing a survey conducted by the
20:54organization james was working for
20:56indicated that the customers weren't
20:58very happy with the organization so they
21:00decided to fix that with the help of six
21:02sigma so they decided that the dmaic
21:04methodology would be best suited to
21:06solve their problem so let's have a look
21:08at what they did firstly in the define
21:10phase they used a tool called the voice
21:12of the customer this tool represented
21:14the needs as well as requirements of the
21:15customer this showed that the customers
21:17expected prompt delivery the correct
21:19product selection and a knowledgeable
21:21distribution team from the company and
21:23now on to the measure phase the company
21:25wanted to know why the customers didn't
21:27like them so they performed some data
21:29collection from there they found out
21:31that they took 56 percent longer than
21:34other companies to deliver their product
21:36so they decided to reduce the amount of
21:38time it takes between order entry and
21:40the delivery of the product and now in
21:42the analyze phase here they knew what
21:44the issue was but they wanted to know
21:46what exactly made their products
21:48delivery so slow why were the customers
21:50receiving the product so late then they
21:53performed some analysis their analysis
21:55showed the possible causes it could have
21:57been inaccurate sales plans issues with
21:59their safety stock issues with their
22:01vendors delivery performance and falling
22:03behind on the manufacturing schedule
22:05further analysis also indicated that
22:07most of their sales almost 80 percent
22:10came from 30 of their products the issue
22:12was that they didn't have enough safety
22:14stock to satisfy the customers who
22:16required that 30 of products and now on
22:19to the improved phase so now that they
22:20knew what was causing their problem they
22:23wanted to solve it they began to have
22:24monthly reviews and tried to make sure
22:27that their in-demand products stayed in
22:29demand another thing that they wanted to
22:31focus on was to make sure that they
22:33could order and provide the customer
22:34with the products that they wanted and
22:36finally onto the control phase they
22:38began to set up plans so that they could
22:40monitor the sales of that 30 of products
22:43that were selling the most each year
22:44they would review how well a product was
22:46selling and replace it if it had fallen
22:48out of favor now let me tell you what a
22:50six sigma team consists of let's talk
22:53about the roles in a six sigma team
22:55first up is level seven now these are
22:57individuals who are at the novice level
23:00now these individuals don't know in
23:02great detail about what the project is
23:04but they have a basic understanding of
23:06the principles and the methodology
23:08behind the program now they usually
23:10support with smaller projects and with
23:12smaller issues but these individuals
23:14found the foundation for the people who
23:16decide where the program is going and
23:18now we're at level six now these are
23:21individuals who have a yellow belt
23:22certification now they are core members
23:24of the six sigma team who have an
23:26understanding of how the basic metrics
23:28work and how they can perform some sort
23:30of improvement now they have their own
23:32areas of expertise and they're required
23:34to determine certain processes that need
23:36to improve at the same time they're also
23:39in charge of smaller improvement
23:40projects now level five these are people
23:43who have a green belt certification now
23:45these individuals are usually part-time
23:47professionals who have a number of
23:48different duties to fulfill they focus
23:50on smaller six sigma projects they are
23:53usually involved with gathering data
23:54performing some sort of experiment and
23:56analyzing information they also assist
23:59with black belt projects and now we're
24:01at level four these are individuals who
24:03have a black belt certification they're
24:05usually team leaders of a six sigma
24:07project they complete four to six
24:09projects a year and are experts in the
24:11principles methodologies and lean
24:13concepts thanks to their understanding
24:15of statistical experimental design they
24:17can also understand the hidden reasons
24:19behind why a particular product failed
24:21and now we're at level three these are
24:24individuals who have a master black belt
24:26certification now these are individuals
24:28who are experts when it comes to
24:29resources the practices and the
24:31methodologies that are employed in six
24:33sigma their main emphasis is to coach
24:35train and certify black belts they also
24:38are involved with other six sigma
24:40leaders to ensure a company's goals are
24:42met now level two these individuals are
24:45called champions so they work really
24:47closely with the executives and usually
24:49have a role like a senior or a middle
24:52executive level role they also have a
24:54clear understanding of what exactly is
24:56the company's vision and mission they
24:58also understand metrics so that they can
25:00set up a six sigma project that lines up
25:02with the company's goals they're
25:03responsible for removing any sort of
25:06roadblock that could hamper the success
25:07of a project and finally we're at level
25:10one these are the executives now these
25:13individuals represent the highest level
25:15when it comes to a six sigma team now
25:16they have training as well as experience
25:18through which they can set up six sigma
25:20projects that clearly line up with the
25:22company's goals their main emphasis is
25:25to ensure that the project is able to
25:27add value to the organization and at the
25:29end of the day is successful now this is
25:31when jenny interjects she wants to know
25:33about lean john tells her that lean just
25:36like six sigma is another methodology so
25:38what exactly is lean now lean is a
25:41methodology that has a very important
25:44ideology to make sure that there's
25:46continuous optimization of the processes
25:48and there's an elimination of waste so
25:51what's waste so waste is basically any
25:53part of the process that the customer
25:56doesn't want to pay for it is a process
25:58that does not add any value to the
26:00customer now coming back to lean here
26:02are some of its characteristics whenever
26:04decisions are being made in a lean team
26:06the main emphasis is to understand how
26:08it exactly adds value to the customer
26:11every member in a lean team has a clear
26:13understanding of what exactly are the
26:15goals of the organization it also
26:17encourages employees to push for further
26:19success even if the organization is in a
26:21good place or is already doing well
26:23there's also an emphasis on
26:24cross-functional collaboration and
26:26communication lean focuses on answering
26:29the difficult question or the complex
26:31ones rather than employing short-term
26:33fixes and with lean you can easily
26:35prepare for issues that can come up in
26:38the future or improvise in unexpected
26:40circumstances so let's talk about how
26:42lean and six sigma are different from
26:44one another the lean methodology aims to
26:47reduce waste it does so by analyzing the
26:50workflow it also emphasizes on
26:52minimizing resource usage and improving
26:55customer value now let's talk about six
26:57sigma the aim of six sigma is to provide
27:00near-perfect results it wants to reduce
27:02costs and improve customer satisfaction
27:05basically both of them are moving
27:06towards the same goal to reduce the
27:08amount of waste and to create efficient
27:10processes now let's talk about the
27:12process of lean now there are five
27:14different steps let's start with the
27:16first one identifying value you need to
27:18identify value by determining what
27:20exactly is the problem you're trying to
27:22solve for the customer the second step
27:24is to map your value stream you need to
27:26map the workflow of your company you
27:28need to focus on the different actions
27:30and the people that are involved with
27:32the process you need to be able to
27:34identify which parts of the process are
27:36able to add value and the ones that
27:38don't the third step is to create a flow
27:40you need to break up your work into
27:42smaller silos and visualize the workflow
27:44so that you can easily identify problems
27:47that might show up later the next step
27:49is to establish pull you need to set up
27:51a system through which products are
27:53created only when there is a demand or a
27:55requirement for it through this you can
27:57optimize resource capacity and finally
27:59we're at the fifth step which is
28:00continuous improvement you need to
28:02ensure all your employees at all levels
28:05are involved in the continuous
28:07improvement of the process so what
28:09exactly is lean six sigma what if you
28:11could combine the best of both worlds
28:14the combination of six sigma and lean
28:16methodology led to the creation of lean
28:19six sigma lean six sigma is a
28:21methodology that aims to solve problems
28:23removes any form of waste or
28:25inefficiency and improving the working
28:27conditions of employees to make sure
28:29that they can serve the customers better
28:31now this is a combination of the tools
28:34methods and principles that are employed
28:36in lean and six sigma let's talk about
28:38some of its advantages it aims to
28:40provide customers with a better
28:41experience by streamlining the process
28:43with efficient power flows it aims to
28:46drive higher results it can reduce cost
28:48remove waste and prevent effects it can
28:50help the organization handle day-to-day
28:52problems the decreased lead times help
28:55increase capacity and profitability and
28:58finally it helps with people development
29:00and improving the morale of the
29:02organization
29:03this lesson provides an overview of the
29:05certified six sigma greenbelt or cssgb
29:09course
29:10a process is a series of steps designed
29:13to produce a product and or service
29:16according to the requirement of the
29:17customer
29:18a process mainly consists of four parts
29:21input process steps output and feedback
29:26input is something put into a process or
29:28expended in its operation to achieve an
29:31output or a result
29:33for example man material machine and
29:36management
29:37output is the final product delivered to
29:40an internal or external customer
29:42for example product or services
29:45it is important to understand that if
29:47the output of a process is an input for
29:49another process the latter process is
29:52the internal customer
29:54each input can be classified as
29:56controllable represented as c
29:59non-controllable represented as nc
30:03noise represented as n
30:06and critical represented as x
30:09the most important aspect of the process
30:11is the feedback
30:13as can be inferred from the image any
30:16change in the inputs causes change in
30:18the output
30:19therefore y equals f of x
30:22feedback helps in process control
30:25because it suggests changes to the
30:27inputs
30:28let us learn about the process of six
30:30sigma in the next screen
30:34let us understand how six sigma works in
30:37this screen
30:38six sigma is successful because of the
30:40following reasons
30:42six sigma is a management strategy it
30:45creates an environment where the
30:47management supports six sigma as a
30:49business strategy and not as a
30:51standalone approach or a program to
30:54satisfy some public relations need
30:56six sigma mainly emphasizes the dmacc
30:59method of problem solving
31:01focused teams are assigned well-defined
31:03projects that directly influence the
31:06organization's bottom line with customer
31:08satisfaction and increased quality being
31:11byproducts
31:13six sigma also requires extensive use of
31:15statistical methods
31:17the next screen will focus on some key
31:19terms used in six sigma
31:23let us look at the sigma level chart in
31:26this screen
31:27as discussed earlier the six sigma
31:29quality means 3.4 defects in 1 million
31:32opportunities or process with a
31:3699.99966 percent yield
31:38the sigma level chart given on the
31:40screen shows the values for other sigma
31:43levels
31:44please take a look at the values
31:45carefully
31:46let us understand the benefits of six
31:48sigma in the next screen
31:51the organizational benefits of six sigma
31:53are as follows
31:55a six sigma process eliminates the root
31:58cause of problems and defects in a
32:00process
32:01sometimes the solution is creating
32:03robust products and services that
32:05mitigate the impact of a variable input
32:08or output on a customer's experience for
32:10example many electrical utility systems
32:13have voltage variability up to and
32:16sometimes exceeding a 10 percent
32:18deviation from nominal value
32:20thus most electrical products are built
32:23to tolerate the variability
32:25drawing more amperage without damage to
32:28any components or the unit itself
32:31using six sigma reduces variation in a
32:34process and thereby reduces waste in a
32:37process
32:38it ensures customer satisfaction and
32:40provides process standardization
32:43rework is substantially reduced because
32:46one gets it right the very first time
32:49further six sigma addresses the key
32:51business requirement
32:53six sigma can also be used by
32:55organizations to gain advantage and
32:57become world leaders in their respective
32:59fields
33:00ultimately the whole six sigma process
33:03is to satisfy customers and achieve
33:05organizational goals
33:07in the next screen let us understand six
33:09sigma and quality
33:14taking a process to six sigma level
33:16ensures that the quality of the product
33:18is maintained
33:19the primary goal of improved quality is
33:22increased profits for the organization
33:25in very simple terms quality is defined
33:27as the degree of excellence of a product
33:29or a service provided to the customer
33:32it is conformance to customer
33:34requirement
33:35if the customer is satisfied with the
33:37product or service then the product or
33:39service is of the required quality
33:41let us look at the history of quality in
33:44the next screen in the mid 1930s
33:46statistical process control spc was
33:50developed by walter shoehart and used
33:52extensively during world war ii to
33:54quickly expand the us's industrial
33:57capabilities
33:58spc is the application of statistical
34:01techniques to control any process
34:04walter shoehart's work on the common
34:06cause of variation and special cause of
34:08variation assignable has been used
34:11proactively in all six sigma projects
34:14the approach to quality has varied from
34:17time to time
34:18in the 1960s there were quality circles
34:21which originated in japan
34:23it was started by kaoru ishikawa
34:26quality circles were self-improvement
34:28groups composed of small number of
34:30employees belonging to a single
34:32department
34:33quality circles brought in improvements
34:35with little or no help from the top
34:37management
34:38in 1987 iso 9000 was introduced
34:43iso stands for international
34:45organization for standardization
34:47iso 9000 is a set of international
34:50standards on quality management and
34:52quality assurance to help organizations
34:55implement quality management systems
34:57iso 9000 is still in effect
35:01baldrige award now known as the malcolm
35:04baldrige national quality award was
35:06developed by the u s congress in 1987 to
35:10raise awareness of quality management
35:12systems as well as recognize and award u
35:15s companies that have successfully
35:17implemented quality management systems
35:20in 1988 another quality approach was
35:23developed known as benchmarking
35:26in this approach an organization
35:28measures its performance against the
35:30best organizations in its field
35:32determines how such performance levels
35:34were achieved and the information is
35:37used by the organization to improve
35:39itself
35:40then in the 1990s there was the balanced
35:43scorecard approach
35:45it is a management tool that helps
35:46managers of all levels to monitor their
35:49results in their key areas so that one
35:52metric is not optimized while another is
35:54ignored
35:56during the year 1996 through 1997 an
36:00approach known as re-engineering was
36:02developed
36:03this approach involved the restructuring
36:05of an entire organization and its
36:08processes
36:09integrating various functional tasks
36:11into cross-functional processes is one
36:14of the examples of re-engineering
36:17in the next screen let us find out about
36:19the quality gurus and their contribution
36:22to the field of quality
36:25let us focus on six sigma and the
36:27business system in this screen
36:30business systems are designed to
36:32implement a process or a set of
36:34processes
36:35a business system ensures that process
36:38inputs are at the right place and at the
36:40right time so that each step of the
36:42process has the resource it needs
36:46a business system design should be
36:48responsible for collecting and analyzing
36:50data so that continual improvement of
36:53its processes products and services is
36:56ensured
36:57a business system has processes sub
37:00processes and steps as its subsets
37:04human resources manufacturing and
37:06marketing are some examples of processes
37:09in a business system
37:12six sigma improves a business system by
37:15continuously removing the defects in its
37:18processes and also by sustaining the
37:20changes
37:21a defective item is any product or
37:23service that a customer would reject
37:26a customer can be the user of the
37:28ultimate product or service or can be
37:31the next process downstream in the
37:33business system
37:35let us learn about six sigma projects
37:37and organizational goals in the
37:39following screen
37:43let us understand the structure of the
37:45six sigma team in this screen
37:47there are totally five levels in the six
37:50sigma screen
37:51the first level consists of the top
37:53executives of the organization
37:55these people lead change and provide
37:57direction as they own the vision of the
37:59organization
38:00for any improvement initiative to work
38:03it is important that top management of
38:05the organization be actively involved in
38:08its propagation
38:09the top executives own the six sigma
38:12initiatives
38:13next in the level are six sigma
38:15champions
38:16they identify and scope projects develop
38:19deployment and strategy and support
38:22cultural change
38:23they also identify and coach master
38:26black belts
38:27three to four master black belts work
38:29under every champion
38:31six sigma master black belts train and
38:34coach black belts green belts and
38:36various functional leaders of the
38:38organization
38:40they usually have at least three to four
38:41black belts under them
38:44the fourth level in six sigma structure
38:46is six sigma black belts
38:48they apply strategies to specific
38:50projects and lead and direct teams to
38:53execute projects
38:55finally there are six sigma green belts
38:58they support the black belt employees by
39:01participating in project teams
39:03green belts play a dual role
39:06they work on the project and perform
39:08day-to-day jobs related to their work
39:10area
39:11in the next screen we will understand
39:13the drivers and metrics of
39:15organizational strategy
39:18while financial accounting is useful to
39:21track physical assets the balanced
39:23scorecard or bsc
39:25offers a more holistic approach to
39:27strategy implementation and performance
39:30measurement by taking into account
39:32perspectives other than the financial
39:34one
39:35for an organization traditional
39:37strategic activities that concentrate
39:39only on financial metrics are not
39:41sufficient to predict future performance
39:44they are not sufficient to implement and
39:46control the strategic plan either
39:49bsc translates the organizational
39:51strategy into actionable objectives that
39:54can be met on an everyday basis and
39:57provides a framework for performance
39:59measurement
40:00the balanced scorecard helps clarify the
40:03organizational vision and mission to
40:05workable action items to be carried out
40:07and measured it also provides feedback
40:10on both internal business processes and
40:13external outcomes
40:14by doing so it enables continuous
40:17improvement in strategic performance
40:19toward achieving organizational goals
40:22the balanced scorecard achieves all this
40:25by integrating the organizational
40:26strategy with a limited number of key
40:29metrics from four major areas of
40:31performance
40:32finance customer relations internal
40:35processes and learning and growth
40:38many organizations in the world use
40:41balanced scorecard approaches and the
40:43number is increasing every day
40:46in the next screen we will describe the
40:48balanced scorecard framework
40:50we will learn about developing a
40:52balanced scorecard in this screen
40:55while applying the balanced scorecard in
40:57an organization care must be taken to
40:59account for interactions between
41:01different perspectives or strategic
41:03business units and avoid optimizing the
41:06results of one at the expense of another
41:09to outline the strategy a top-down
41:12approach is followed by determining the
41:14strategic objectives measures targets
41:16and initiatives for each perspective the
41:20strategic objectives refer to the
41:22strategy to be achieved in that
41:24perspective
41:25three or four leading objectives are
41:27agreed upon the progress towards
41:30strategic objectives is assessed using
41:32specific measures these measures should
41:35be closely related to the actual
41:37performance drivers
41:38this enables effectively evaluating
41:40progress
41:42high level metrics are linked to lower
41:44level operational measures
41:46the target values for each measure are
41:48set the initiatives required to achieve
41:50the target value are identified
41:53as already mentioned this exercise is
41:56carried out for all the perspectives
41:59finally the scorecard is integrated into
42:01the management system
42:03in the next screen let us understand the
42:05change in the approach to the balance
42:07score card from the four box model of
42:09bsc to strategy maps
42:14in earlier approaches to the balanced
42:16scorecard the perspectives were
42:18presented in a four box model
42:21this kind of scorecard was more a
42:23comprehensive glance at the key
42:25performance indicators or metrics in
42:27different perspectives
42:29however the key performance indicators
42:31or metrics of different perspectives
42:33were viewed independent of each other
42:36which led to a silo based approach and
42:38lack of integration
42:40however modern scorecards place the
42:43focus on the interrelations between the
42:45objectives and metrics of different
42:47perspectives and how they support each
42:49other
42:50a well-designed balanced scorecard
42:53recognizes the influence of one
42:55perspective on another and the effect of
42:57these interactions on organizational
43:00strategy
43:01to achieve the objectives in one
43:03perspective it is necessary to achieve
43:06the objectives in another perspective
43:08in short the four perspectives form a
43:11chain of cause and effect relationships
43:14a map of interlinked objectives from
43:16each perspective is created
43:19these objectives represents the
43:21performance drivers which determine the
43:23effectiveness of strategy implementation
43:26this is called a strategy map
43:29the function of a strategy map is to
43:31outline what the organization wants to
43:33accomplish and how it plans to
43:35accomplish it
43:36the strategy map is one page view of how
43:39the organization can create value
43:42for example financial success is
43:44dependent on giving customers what they
43:47want which in turn depends on the
43:49internal processes and learning and
43:51growth at an individual level
43:54in the next screen we will look at the
43:56impact of the balanced scorecard on the
43:58organization
44:01the balance scorecard and strategy map
44:03force managers to consider cause and
44:05effect relationships which leads to
44:08better identification of key drivers and
44:10a more rounded approach to strategic
44:12planning the bsc enables the
44:15organization to improve in the following
44:17ways
44:18being a one-page document a strategy map
44:21can easily be communicated and
44:23facilitates understanding at all levels
44:25of the organization
44:27an organization is successful in meeting
44:29its objectives only when everyone
44:32understands the strategy
44:34the balance scorecard also forces an
44:36organization to measure what really
44:38matters and manage information better so
44:41that quality of decision making is
44:43higher
44:44creating performance reports against a
44:46balanced scorecard allows for a
44:48structured approach to reporting
44:50progress
44:51it also enables organizations to create
44:53reports and dashboards to communicate
44:56performance transparently and
44:58meaningfully
44:59as expected a balanced scorecard helps
45:01an organization to better align itself
45:04and its processes to the strategic goals
45:06outlined in the bsc
45:09the overall objectives of the bsc can be
45:11cascaded into each business unit to
45:14enable that unit to work toward the
45:16common organizational goal
45:19all the activities of the organization
45:21such as budgeting or risk management are
45:23automatically aligned to the strategic
45:25objectives
45:27to conclude the balance scorecard is a
45:29simple and powerful tool that when
45:32implemented correctly equips an
45:34organization to perform better
45:37let us proceed to the next topic of this
45:39lesson in the following screen
45:41in this topic we will look at what lean
45:44is and how lean is applied to a process
45:48let us start with the lean concepts in
45:50the next screen
45:51let us look at the process issues in
45:54this screen
45:55lean focuses on three major issues in a
45:58process known by their japanese names
46:01muda mura and marie
46:04muda refers to non-value-adding work
46:06mura represents unevenness and marie
46:09represents overburden
46:11together they represent the key aspects
46:14in lean
46:15let us look at the types of waste in the
46:17next screen
46:22there are seven types of muda or waste
46:25as per lean principles
46:27let us understand these seven types of
46:29muta
46:30overproduction
46:32this refers to producing more than is
46:34required
46:35for example a customer needed 10
46:38products and 12 were delivered
46:40inventory in simple words this refers to
46:44stock
46:45the term inventory includes finished
46:47goods semi-finished goods raw materials
46:50supplies kept in waiting and some of the
46:53work in progress for example test
46:56scripts waiting to be executed by the
46:58testing team
46:59defects repairs rejects
47:01any product or service deemed unusable
47:04by the customer or any effort to make it
47:06usable to the original customer or a new
47:08customer for example errors found in the
47:11source code of a payroll module by
47:14quality control team
47:15motion
47:16a waste due to poor ergonomics of the
47:18workplace for example finance and
47:21account teams sit on the first floor but
47:23invoices to customers are printed on the
47:26ground floor causing unnecessary
47:28personnel movement
47:30over processing
47:32additional process on a product or
47:34service to remove unnecessary attribute
47:36or feature is over processing
47:39for example a customer needs a bottle
47:41and you deliver a bottle with extra
47:43plastic casing a customer needs abec3
47:47bearing and your process is tuned to
47:49produce more precise abec 7 bearings
47:53taking more time for something the
47:54customer doesn't need
47:56waiting when a part waits for processing
47:59or the operator waits for work the
48:01wastage of waiting occurs for example
48:04improper scheduling of staff members
48:06transport when the product moves
48:09unnecessarily in the process without
48:11adding value
48:13for example a product is finished and
48:15yet it travels 10 kilometers to the
48:17warehouse before it gets shipped to the
48:19customer
48:20another example an electronic form is
48:23transferred to 12 people some of them
48:25seeing the form more than once that is
48:28the form is traveling over the same
48:30space multiple times
48:32next we will look at lean wastes other
48:34than the seven types of waste discussed
48:36some lean experts talk about additional
48:39areas of waste
48:41under utilized skills
48:43skills are underutilized when the
48:45workforce has capabilities that are not
48:47being fully used toward productive
48:49efforts
48:50people are assigned to jobs in which
48:52they do not fit
48:54underperforming processes
48:56automation of poorly performing process
48:59improving a process that should be
49:01eliminated if possible for example the
49:04product returns department or product
49:06discounts process
49:07a symmetry in processes that should be
49:10eliminated for example two signatures to
49:12approve a cost reduction and six
49:14signatures to reverse a cost reduction
49:17that created higher costs in other areas
49:20in the next screen we will look at an
49:22exercise on identifying the waste type
49:25we will cover each step of the lean
49:27process in the next few screens in this
49:30screen we will learn about the first
49:31step identify value to implement lean to
49:35a process
49:36it is important to find out what the
49:38customer wants
49:40once this is done the process should be
49:42evaluated to identify what it needs to
49:45possess to meet customer requirements
49:48the next screen will focus on the next
49:51step of the lean process
49:53value stream mapping
49:56in this screen we will discuss the
49:58differences between push and pull
50:00processes
50:01an organization can adopt either of
50:03these processes depending on the
50:05requirement
50:07contrary to a pull process in a push
50:09process the first step is to forecast
50:12the demand for a product or service
50:15the production line then begins to fill
50:17this demand and produced parts are
50:19stocked in anticipation of customer
50:21demand
50:22for example a garments manufacturer
50:25produces 200 shirts based on expected
50:27demand and then waits for customer
50:29orders for them
50:31note that the demand is expected and not
50:33actual
50:34discounts offered to customers by big
50:37retailers are examples of the push
50:39process
50:40if the garment company adopts a pull
50:42process instead it would start making
50:45the shirts only after receiving a
50:47confirmed demand from customers
50:50note that although the pull approach
50:52seems better it is not applicable to all
50:54situations for example a pharmacy uses a
50:58push process
50:59in the next screen we will learn about
51:01theory of constraints
51:04let us look at an example for the toc
51:07methodology in this screen
51:09the three sub processes in the packing
51:12process are coding or printing filling
51:15and sealing
51:16the data for the three sub processes are
51:19observed and collected as number of
51:21units produced in an hour
51:23the data is as follows
51:25coding or printing is 900 units per hour
51:29filling is 720 units per hour and
51:32ceiling is 780 units per hour
51:35how can you implement the toc
51:37methodology in this example
51:40let us build the toc map for this
51:42example
51:43the first step in the toc methodology is
51:46to identify the constraint
51:48looking at the data the output per hour
51:51from the filling process is 720
51:54this is the constraint in the system
51:57in the second step the constraint is
51:59exploited by analyzing the performance
52:02using data
52:03to break the constraint a repair and
52:05maintenance personnel can be assigned to
52:07maintain the filling machine on a daily
52:09basis
52:11in the third step the other fixes in the
52:13repair and maintenance function are made
52:15as subordinate decisions to the one
52:17taken in step two
52:19in this example carry out the
52:21maintenance of the filling machine
52:23in the fourth step the constraint is
52:25elevated by implementing the decisions
52:28in this example remove the damages from
52:30the filling machine the next step is to
52:33go back to step one and identify the
52:35next system constraint
52:37as per the data collected after
52:39implementation of the first cycle of the
52:41toc
52:43ceiling can be identified as the next
52:45system constraint
52:48let us now analyze the data before and
52:50after toc implementation in this example
52:54the number of units produced per hour
52:56before implementing the toc encoding or
52:58printing process was 900 units filling
53:02process was 720 units and sealing
53:05process was 780 units
53:08after implementing the toc the number of
53:10units produced per hour for the filling
53:12process increased to 840 from 720 units
53:19let us proceed to the next topic of this
53:21lesson
53:22in this topic we will discuss the
53:24concepts in design for six sigma or dfss
53:28let us first understand dfss in the next
53:32screen
53:34dfss or design for six sigma is a
53:38business process methodology that
53:40ensures that any new product or service
53:43meets customer requirements and the
53:45process for that product or service is
53:47already at six sigma level
53:50dfss uses tools such as quality function
53:53deployment or qfd and failure mode and
53:57effects analysis or fmea
54:01dfss can help a business system to
54:04introduce an entirely new product or
54:06service for the customer
54:08it can also be used to introduce a new
54:11category of product or service for the
54:13business system
54:14for example an fmcg company plans to
54:18make a new brand of hair oil a type of
54:20product already in the market
54:23dfss also improves the product or
54:26service and adds to the current product
54:28or service lines
54:30to implement dfss a business system has
54:33to know its customer requirements
54:36dfss can be used to design a new product
54:39or service a new process for a new
54:41product or service or redesign of an
54:44existing product or service to meet
54:46customer requirements
54:48let us learn about processes for dfss in
54:51the next screen
54:55the two major processes for dfss are
54:58idov and dmadv
55:02idov stands for identify design optimize
55:06and verify
55:08dmadb stands for define measure analyze
55:13design and verify
55:15in the idov process
55:18the first step involves identifying
55:20specific customer needs based on which
55:22the new product and business process
55:24will be designed
55:26the next step involves design which
55:28involves identifying functional
55:30requirements developing alternate
55:32concepts evaluating the alternatives
55:35selecting a best fit concept and
55:38predicting sigma capability
55:40tools such as fmea are used here
55:44the third step optimize uses a
55:47statistical approach to calculate
55:49tolerance with respect to the desired
55:51output
55:52when idov is implemented to design a
55:55process expected to work at six sigma
55:58level this is checked in the optimize
56:00phase if the process does not meet
56:03expectations the optimize phase helps in
56:06developing detailed design elements
56:08predicting performances and optimizing
56:11design
56:12the last stage of idov is to verify that
56:16is to test and validate the design and
56:18finally to check conformance to six
56:20sigma standards
56:22the other process dmadv has five stages
56:27the first stage is to define the
56:28customer requirements and goals for the
56:30process product or service
56:33next measure and match performance to
56:35customer requirements
56:37the third stage involves analysis and
56:40assessment of the design for the process
56:42product or service the next step is to
56:45design and implement the array of latest
56:47processes required for the new process
56:50product or service
56:52the final stage is to verify results and
56:54maintain performance
56:57in the next screen we will look at the
56:59differences between idov and dmadb
57:05the primary difference between idov and
57:08dmadb is that while i
57:11is used only to design a new product or
57:14service
57:15dmadb can be used to design either a new
57:18product or service or redesign an
57:21existing product or service
57:23idov involves design of a new process
57:27while dmadv involves redesigning an
57:30existing process
57:31in idov no analysis or measurement of
57:34existing process is done and the whole
57:37development is new
57:39the design step immediately follows the
57:41identification of customer requirements
57:44in contrast dmadv the existing product
57:48service or process is examined
57:50thoroughly before moving to the design
57:52phase the design stage comes only after
57:56defining requirements and analyzing the
57:58existing product service or process
58:02in the following screen we will learn
58:04about tool quality function deployment
58:06or qfd which is one of the dfss tools
58:13qfd also called voice of customer or voc
58:17or house of quality
58:19is a predefined method of identifying
58:21customer requirements
58:23it is a systematic process to understand
58:26the needs of the customer and convert
58:28them into a set of design and
58:30manufacturing requirements
58:32qfd motivates business to focus on its
58:35customers and design products that are
58:38competitive in lesser time and at lesser
58:40cost
58:41the primary learning from qfd includes
58:45which customer requirements are most
58:47important what the organization
58:49strengths and weaknesses are
58:51where an organization should focus their
58:53efforts and where most of the work needs
58:56to be done
58:58to learn from qfd the organization
59:00should ask relevant questions to
59:02customers and tabulate them to bring out
59:05a set of parameters critical to the
59:07design of the product
59:09apart from understanding customer
59:11requirements it is also important to
59:13know what would happen if a particular
59:15product or service fails when being used
59:18by a customer
59:20it is necessary to understand the
59:21effects of failure on the customer to
59:24ensure preventive actions are taken
59:26and to be able to answer the customers
59:28in the event of failure
59:31in the next screen we will look at
59:32another dfss tool
59:34failure modes and effects analysis or
59:37fmea
59:41failure modes and effects analysis or
59:44fmea is a preemptive tool that helps any
59:47system to identify potential pitfalls at
59:50all levels of a business system
59:53it helps the organization to identify
59:55and prioritize the different failure
59:57modes of its product or service and what
1:00:00effect the failure would have on the
1:00:01customer
1:00:02it helps in identifying the critical
1:00:05areas in a system on which the
1:00:06organization's efforts can be focused
1:00:09note that while fmea enables
1:00:12identification of critical areas it does
1:00:15not offer solutions to the identified
1:00:17problems
1:00:18we will look at the varieties of fmea
1:00:21such as dfmea and pfmea in the next
1:00:26screen pfmea stands for process failure
1:00:31mode and effects analysis
1:00:33and df-mea stands for design failure
1:00:37mode and effects analysis
1:00:40pfmea
1:00:41is used on a new or existing process to
1:00:44uncover potential failures
1:00:46it is done in the quality planning phase
1:00:48to act as an aid during production
1:00:51a process fmea can involve fabrication
1:00:54assembly transactions or services
1:00:58dfmea is used in the design of a new
1:01:01product service or process to uncover
1:01:03potential failures
1:01:05the purpose is to find out how failure
1:01:07modes affect the system and to reduce
1:01:09the effect of failure on the system
1:01:12this is done before the product is sent
1:01:14to manufacturing
1:01:15all design deficiencies are sorted out
1:01:17at the end of this process
1:01:20in the following screen we will
1:01:21understand fmea risk priority number
1:01:26fmea risk priority number or rpm is a
1:01:31measure used to quantify or assess risk
1:01:33associated with the design or process
1:01:36assessing risk helps identify critical
1:01:39failure modes
1:01:40higher the rpn higher the priority the
1:01:43product or process receives
1:01:46rpn is a product of three numbers
1:01:49severity of a failure
1:01:51occurrence of a failure and the
1:01:53detectability of a failure
1:01:55all these numbers are given a value on a
1:01:58scale of one to ten
1:01:59the minimum value of rpn is one and the
1:02:03maximum value is one thousand a failure
1:02:06mode with a higher currency rating means
1:02:08the failure mode occurs very frequently
1:02:11a mode with a high severity rating means
1:02:14that the mode is really critical to
1:02:16ensure safety of operations
1:02:18a mode with a high detection rating
1:02:20means that the current controls are not
1:02:22sufficient
1:02:23in the next screen we will look at the
1:02:25fmea table
1:02:30the fmea table helps plan improvement
1:02:33initiatives by underlining why and how
1:02:35failure modes occur and helps
1:02:37organizations plan for their prevention
1:02:40typically fmea is applied on the output
1:02:43of root cause analysis and is a better
1:02:46tool for focus or prioritization as
1:02:48compared to multi-voting
1:02:51one important aspect of fmea is that it
1:02:54does not need data
1:02:56experts in a particular area can form
1:02:59the fmea table without having to look at
1:03:02data from any source
1:03:04in functions such as human resources the
1:03:07fmea table is very useful as there might
1:03:11not be much data available to the
1:03:13problem-solving team
1:03:15the sample fmea table is given on the
1:03:18screen
1:03:19please go through the contents for
1:03:21better understanding
1:03:22in the following screen we will discuss
1:03:25severity of risk priority number and
1:03:27scale criteria
1:03:31let us first discuss severity
1:03:33severity refers to the seriousness of
1:03:35the effect of the failure mode
1:03:38or how critical the failure mode is to
1:03:40the customer or the process
1:03:42the severity of a failure mode is rated
1:03:45on a scale of 1 to 10
1:03:47using a severity table
1:03:50different industries follow different
1:03:52structures for the severity table
1:03:54a high severity rating indicates a mode
1:03:57is critical to operational safety
1:04:00for example a team working on fmea of a
1:04:03radioactive plant may insert fatal as
1:04:06the effect with rating 10.
1:04:09another example is the severity table
1:04:11for a sports team
1:04:13the team manager wants to rate the
1:04:15severity of failure of the team in an
1:04:17upcoming game
1:04:18she might rate it at nine given that the
1:04:21team would lose a big sponsorship should
1:04:23they face defeat
1:04:25which could in turn be hazardous to the
1:04:27team's future
1:04:29shown here is a generalized table of
1:04:31severity
1:04:32the severity rating can never be changed
1:04:36for example if a mode has a rating of
1:04:38nine before improvement it will continue
1:04:40to have a rating of 9 after improvement
1:04:432.
1:04:44let us look at occurrence of rpm and
1:04:47scale criteria
1:04:50occurrence is the probability that a
1:04:52specific cause will result in the
1:04:54particular failure mode
1:04:56as with severity occurrences rated on a
1:04:59scale of one to ten based on a table
1:05:02like the severity table higher the
1:05:04occurrence of a failure higher is its
1:05:06rating
1:05:07again this table might vary depending on
1:05:09the industry and scenario
1:05:12sometimes the project team can use data
1:05:14here if available
1:05:16based on past data the probability of
1:05:19occurrence of a failure can easily be
1:05:21rated
1:05:22shown here is a generalized table let us
1:05:25next look at detection of rpn and scale
1:05:28criteria
1:05:31detection is the probability that a
1:05:33particular failure will be detected
1:05:36the table shown here is again a
1:05:38generalized one
1:05:40the rating here is a bit different from
1:05:42severity or occurrence
1:05:44higher the detectability of a failure
1:05:46lower is its rating
1:05:48this is because if the failure can
1:05:51easily be detected then everyone would
1:05:54know of it and therefore there would be
1:05:56less or no damage
1:05:58for example if detection is impossible
1:06:01the failure is given a rating of ten
1:06:03please note that at the start of a six
1:06:05sigma project the failure mode is given
1:06:08a relatively high detection rating
1:06:10let us look at an example of fmea and
1:06:14rpn in the next screen
1:06:18in this example a bank wants to
1:06:20recognize and prioritize the risks
1:06:22involved in the process of withdrawing
1:06:24cash from an atm
1:06:27it can be observed from the table that
1:06:29not having a control in place for
1:06:31network issues has the highest rpm
1:06:35this is due to the detectability for a
1:06:37network issue being very low
1:06:40the next set of information in the table
1:06:42shows the action taken by the bank's
1:06:44management to address the failure modes
1:06:47following the implementation the new rpn
1:06:50is calculated retaining the security
1:06:52level at nine
1:06:54this is because the actions were not
1:06:56directed at reducing the severity but at
1:06:59the causes of failure
1:07:01it can be seen that the new rpn is much
1:07:04lower and the risk for both causes has
1:07:07been mitigated
1:07:09this lesson will cover the details of
1:07:11the define phase six sigma can be
1:07:14applied to everything around
1:07:16it can be applied across almost 70
1:07:19different sectors however it cannot be
1:07:21applied to all problems
1:07:24the first step is to check if the
1:07:26project qualifies to be a six sigma
1:07:29project
1:07:30the questions that need to be asked are
1:07:32as follows
1:07:34is there an existing process
1:07:36to implement the dmac methodology of
1:07:38problem solving a process needs to exist
1:07:42the process should be in operation for
1:07:44the development of the product or
1:07:45service
1:07:47is there a problem in the process
1:07:50ideally the process should not have any
1:07:53problem
1:07:54if there is a problem in the process
1:07:56performance the process needs to be
1:07:59improved
1:08:01is the problem measurable
1:08:03the problem has to be measurable in
1:08:05order to assess the root cause and the
1:08:08impact of the problem on the process
1:08:12does the problem impact customer
1:08:14satisfaction
1:08:16if the problem affects customer
1:08:18satisfaction an action needs to be taken
1:08:20immediately
1:08:22else the customer may start finding
1:08:24alternate products or switch to
1:08:26competitors products
1:08:28does working on the problem impact
1:08:30profits of the company
1:08:32it is very essential to assess the
1:08:34impact of the project on the profits of
1:08:37the company
1:08:38if the project affects the profits of
1:08:40the company adversely then such a
1:08:43project is not feasible
1:08:46is the root cause of the problem unknown
1:08:49if the root cause of the problem is
1:08:51visible then a six sigma project is not
1:08:54required
1:08:56other problem solving techniques can be
1:08:58used in this case
1:09:00is the solution unknown
1:09:03if the solution to the problem is
1:09:05already known then there is no need for
1:09:08any project
1:09:09the company can directly implement the
1:09:11solution
1:09:13the define phase of dmacc will be
1:09:15introduced in the next
1:09:18screen the six sigma project process is
1:09:22known as a dmacc process
1:09:24define is the first phase in the six
1:09:27sigma project process
1:09:29in the define phase the problem is
1:09:32defined and the six sigma team is formed
1:09:36the objectives of the define phase are
1:09:38as follows
1:09:40clearly define the problem statement
1:09:42through customer analysis
1:09:45understand customer requirements
1:09:47and ensure that the six sigma project
1:09:50goals are aligned to these requirements
1:09:53define the objectives of the six sigma
1:09:56project
1:09:57plan the project in terms of time budget
1:10:00and resource requirements
1:10:03define the team structure for the
1:10:05project
1:10:06and establish roles and responsibilities
1:10:09in the next screen let us learn about
1:10:12benchmarking
1:10:14benchmarking is the process of comparing
1:10:16an organization's business processes
1:10:19practices and performance metrics with
1:10:22that of industry leaders
1:10:24there are various types of benchmarking
1:10:27let us briefly look at each type process
1:10:30benchmarking entails comparing specific
1:10:33processes to a leading company or an
1:10:35industry standard
1:10:37this is useful to obtain a simplified
1:10:39view of business operations
1:10:41and enables a focused review of major
1:10:44business functions
1:10:46process benchmarking includes
1:10:48comparisons of production processes
1:10:51data collection processes
1:10:53performance indicators and productivity
1:10:56and efficiency reviews
1:10:58financial benchmarking is performed to
1:11:01assess overall competitiveness and
1:11:03productivity
1:11:04it is done by running a detailed
1:11:06financial analysis and analyzing the
1:11:09result
1:11:10performance benchmarking involves
1:11:12comparison of products and services with
1:11:14those of competitors
1:11:16with the intention of evaluating the
1:11:18organization's competitiveness
1:11:21product benchmarking involves designing
1:11:23new products or services or upgrading
1:11:26existing products or services
1:11:29this can involve reverse engineering a
1:11:31competitor's products to study the
1:11:33strengths and weaknesses and modeling
1:11:35the new product on these findings
1:11:38strategic benchmarking refers to
1:11:40studying strategies and problem-solving
1:11:42approaches in other industries
1:11:45functional benchmarking is the focused
1:11:48analysis of a single function with the
1:11:50aim of improving it
1:11:52complex functions may need to be divided
1:11:54into processes before benchmarking is
1:11:57done
1:11:58competitive benchmarking includes
1:12:00standardizing organizational strategies
1:12:03process products services and procedures
1:12:07against the competitors in the same
1:12:08industry
1:12:10collaborative benchmarking is a type of
1:12:13benchmarking where the standardization
1:12:15of various business parameters is
1:12:17carried out by a group of companies and
1:12:19the information is shared if the
1:12:22subsidiary units of a company or its
1:12:24various branches carry out the
1:12:26benchmarking it is called collaborative
1:12:28benchmarking
1:12:30let us take a look at best practices for
1:12:33benchmarking in the next screen
1:12:37best practice is a method that ensures
1:12:39continuous improvement leading to
1:12:41exceptional performance
1:12:43it is also a method to sustain and
1:12:45develop the process continuously
1:12:48some of the best practices in
1:12:50benchmarking are as follows
1:12:53increase the objectives or scope of
1:12:55benchmarking
1:12:57set the standards and path to be
1:12:58followed at the initial stage
1:13:01reduce unnecessary effort and comply
1:13:04with the scope
1:13:05recognize the best in the industry to
1:13:08set a benchmark
1:13:10share the information derived from
1:13:12benchmarking
1:13:14in the next screen we will discuss
1:13:16project selection
1:13:20let us understand process business
1:13:23process and business system in this
1:13:25screen
1:13:27a process is a series of steps designed
1:13:30to produce a product or service to meet
1:13:33the requirement of a customer
1:13:35a process mainly consists of three
1:13:37elements
1:13:39input
1:13:40process and output
1:13:42a business process is a systematic
1:13:44organization of objects such as people
1:13:47machinery and materials
1:13:49into work activities designed to produce
1:13:52a required product or service
1:13:55as shown on the screen a process is a
1:13:58subset of a business process
1:14:00a business process is in turn a part of
1:14:03a business system
1:14:05a business system is a value-added chain
1:14:08of various business processes such as
1:14:11sales or finance
1:14:13for example
1:14:14payroll calculation is a process in the
1:14:17hr business process of an i t company
1:14:20which is a business system
1:14:23in the next screen we will look at the
1:14:25process elements
1:14:28let us discuss the challenges to
1:14:30business process improvement in this
1:14:32screen
1:14:33the improvement to a business process of
1:14:36an organization faces challenges due to
1:14:38the traditional business system
1:14:40structure
1:14:41because it is generally grouped around
1:14:43the functional aspect
1:14:45the main problem in a functionally
1:14:47grouped organization is the movement or
1:14:50flow of a product or service
1:14:52a product or service has to go through
1:14:54various functions and their functional
1:14:57elements to reach the customer or end
1:14:59user
1:15:00the other problem is management of the
1:15:02flow of products or services across
1:15:05various functional elements
1:15:07this is difficult as usually there is no
1:15:09one in charge
1:15:11these business process improvement
1:15:13problems can be solved using the project
1:15:16management approach to produce the
1:15:18product or service
1:15:20in the next screen we will learn about
1:15:22process owners and different
1:15:24stakeholders of a project
1:15:28the representation of where the process
1:15:30owner and stakeholders are placed in the
1:15:32organizational hierarchy is on the
1:15:34screen
1:15:35the process owner is a person at a
1:15:37senior level in the hierarchy
1:15:40he is the one who takes responsibility
1:15:42for the performance and execution of a
1:15:44process
1:15:45and also has the authority and the
1:15:47ability to make necessary changes
1:15:50on the other hand a stakeholder is a
1:15:53person group or organization which is
1:15:55affected or can affect an organization's
1:15:58actions
1:16:00businesses have many stakeholders like
1:16:02stockholders customers suppliers company
1:16:05management
1:16:06employees of an organization and their
1:16:08families the society etc
1:16:11let us discuss the effects of process
1:16:14failure on various stakeholders in the
1:16:16next screen
1:16:18while it is an absolute business
1:16:20necessity to keep one stakeholder
1:16:22satisfied at all times
1:16:23failure to meet one or more process
1:16:25objectives may result in negative
1:16:27effects on them
1:16:29in such situations for the stockholders
1:16:32the perceived value for the company gets
1:16:34reduced
1:16:35customers may seek other competitors for
1:16:37their deals while imposing penalties and
1:16:40finding recourse in legal action against
1:16:42the company
1:16:43suppliers may be on the losing front
1:16:46with delayed payments or not being paid
1:16:48at all
1:16:49company management may require cost cut
1:16:52down
1:16:52employees will receive diminishing wages
1:16:55the community and society will be
1:16:58affected due to pollution created by the
1:17:00organization
1:17:02in the next screen we will understand
1:17:04the relationship between business and
1:17:06the stakeholder
1:17:09in the diagram shown on the screen
1:17:11each stakeholder is both a supplier as
1:17:14well as a customer
1:17:15forming many closed loop processes that
1:17:18must be managed controlled balanced and
1:17:20optimized for the business to thrive
1:17:23communication is the key in such
1:17:25situations and is facilitated through
1:17:27internal company processes
1:17:30the next screen covers the importance
1:17:32and relevance of stakeholder analysis
1:17:35stakeholder analysis is an important
1:17:38task to be completed before doing a six
1:17:40sigma project
1:17:42a business has many stakeholders and any
1:17:46change to a business process affects
1:17:48some or all of them
1:17:50when a process does not meet its
1:17:52objectives
1:17:53it results in the stakeholders being
1:17:55negatively affected
1:17:57which in turn affects the organization's
1:17:59performance
1:18:01the six sigma team must factor in the
1:18:03reasons why a stakeholder may oppose the
1:18:06change effort
1:18:08let us proceed to the next topic of this
1:18:10lesson in the following screen
1:18:14in this topic we will discuss voice of
1:18:17the customer
1:18:18let us start with how to identify the
1:18:20customer in the following screen
1:18:25customers are the most important part of
1:18:27any business
1:18:29a customer is someone who decides to
1:18:31purchase pays consumes and gets affected
1:18:34by a particular product or service
1:18:37it is critical to identify and
1:18:39understand the customer requirements
1:18:42the products or services can be designed
1:18:44according to these requirements
1:18:46consequently the company is able to
1:18:49provide products or services the
1:18:51customers are willing to purchase there
1:18:54are two types of customers
1:18:56internal and external customers
1:18:59in the next screen we will learn about
1:19:01internal customers
1:19:05an internal customer can be defined as
1:19:07anyone within the business system who is
1:19:10affected by the product or the service
1:19:12while it is being developed
1:19:15most often internal customers are the
1:19:17employees of the company
1:19:20for example
1:19:21let us assume that there is a series of
1:19:23processes in a particular business
1:19:25system
1:19:27in such a scenario
1:19:29the second process is the internal
1:19:31customer for the first process
1:19:34the third process is an internal
1:19:36customer for the second process and so
1:19:39on
1:19:40the basic needs of an internal customer
1:19:42are to be provided proper tools and
1:19:45necessary equipment imparted proper
1:19:47training and given specific instructions
1:19:50to carry out their responsibilities
1:19:53however the needs are not limited to
1:19:55these alone
1:19:57other needs include the provision of
1:19:58company newsletters projects storyboards
1:20:01to display the letters etc team meetings
1:20:03to share business news and announcements
1:20:06staff meeting to share information and
1:20:08quality awards from suppliers
1:20:11an internal customer is important
1:20:14first of all the activities of an
1:20:16internal customer directly affect the
1:20:18final or ultimate customer
1:20:20secondly the activities of an internal
1:20:23customer affect the next process in the
1:20:25system
1:20:27finally an internal customer also
1:20:29affects the quality of the product
1:20:30developed or service provided
1:20:33when the needs of the internal customers
1:20:36in most cases employees are met
1:20:39they are more likely to have higher
1:20:41perceptions of quality and also
1:20:43contribute to greater productivity
1:20:46the satisfaction levels of the internal
1:20:48customers can be improved in various
1:20:51ways these include a higher amount of
1:20:53internal communication through company
1:20:56newsletters and team meetings
1:20:58recognitions for work quality awards etc
1:21:01constant training on how to be ahead and
1:21:03well equipped in a competitive
1:21:05environment is very essential too
1:21:08in the next screen we will learn about
1:21:10external customers
1:21:13this screen focuses on the positive
1:21:15effects of a project on the customers
1:21:18the most important aspect of any process
1:21:21improvement project is the customers
1:21:24internal customers are the ones who
1:21:26drive the project
1:21:27hence the effect of the project on
1:21:29internal customers is a critical factor
1:21:32that needs to be considered
1:21:34the positive impact of a project on the
1:21:36internal customers is as follows
1:21:40the project is driven by highly
1:21:42motivated individuals or internal
1:21:44customers who are aware of the project
1:21:46objectives and scope
1:21:49individuals belonging to a credible
1:21:51project understand the project
1:21:53deliverables and display high levels of
1:21:56job satisfaction
1:21:58these individuals go the extra mile to
1:22:00take up tasks beyond their job
1:22:02description
1:22:04such individuals make a highly motivated
1:22:06team focused on delivering their
1:22:08responsibilities in order to meet the
1:22:10customer requirements
1:22:13working together in a positive
1:22:14environment also improves team spirit
1:22:17and bonding
1:22:19the positive impact of a project on the
1:22:21external customers is as follows
1:22:25process improvement projects analyze the
1:22:27problems and come up with an effective
1:22:30solution
1:22:31consequently ensuring a better product
1:22:34a successful process improvement project
1:22:36assists the organization in effectively
1:22:39meeting customer expectations or
1:22:41requirements
1:22:42there is visible improvement in customer
1:22:45service
1:22:46good quality product and service ensures
1:22:48high customer satisfaction
1:22:51let us learn about different methods of
1:22:52customer data collection in the
1:22:54following screen
1:22:58once you begin to identify the customer
1:23:01types you need to look forward to
1:23:03collecting customer data
1:23:05collecting data from customers is very
1:23:07essential as it helps consider the
1:23:09levels at which these customers affect
1:23:11the business
1:23:12begin by collecting feedback from both
1:23:15internal as well as external customers
1:23:18customer feedback helps fill the gaps
1:23:20and improve the various business
1:23:22processes in the organization
1:23:25it helps define a good quality product
1:23:27as perceived by the customer
1:23:29and identify qualities that make the
1:23:31competitors products or service better
1:23:34it also helps identify factors which
1:23:37provide a competitive edge to the
1:23:38product or service on offer
1:23:41there are various methods to collect
1:23:43feedback from the customers many of you
1:23:45might be involved in a similar activity
1:23:47at some point or the other
1:23:49popular and common methods are surveys
1:23:52conducted through questionnaires focus
1:23:54groups individual interviews with the
1:23:56customers
1:23:57customer complaints received via call
1:24:00centers emails and feedback forms are
1:24:03also quite prevalent
1:24:05feedback received in this form are from
1:24:07dissatisfied customers
1:24:09in the next screen we will learn about
1:24:11questionnaires
1:24:15let us discuss the advantages and
1:24:17disadvantages of questionnaires in this
1:24:19screen
1:24:21the advantages of a questionnaire are
1:24:23that it costs less
1:24:24the phone response rate is high anywhere
1:24:27from around 70 to 90 percent and it
1:24:30produces faster results
1:24:32also analysis of male questionnaires
1:24:35requires few trained resources
1:24:38questionnaire is a method used to gather
1:24:40data however there are a few
1:24:42disadvantages associated with it
1:24:46there may be incomplete results and
1:24:48unanswered questions leading to a lack
1:24:50of clarity
1:24:52the response rate of mail surveys is
1:24:54about 20 to 30 percent only
1:24:57at times phone surveys can produce
1:24:59undesirable results as the interviewer
1:25:02can influence the person being
1:25:04interviewed
1:25:06we will differentiate between telephone
1:25:08survey and web survey in the next screen
1:25:12there are different methods to collect
1:25:14data for a survey
1:25:16the methods need to be based on the
1:25:18requirements and needs of the
1:25:19organization
1:25:21the popular methods of survey are the
1:25:23telephone survey and web survey both
1:25:26have their own drawbacks and benefits
1:25:28which are given on the screen
1:25:30the organization needs to choose a
1:25:32method of collecting data according to
1:25:34the situation
1:25:35it is recommended to go through the
1:25:37content for a better understanding in
1:25:40the next screen we will learn about
1:25:42focus groups
1:25:45let us now discuss the advantages and
1:25:47disadvantages of using a focus group for
1:25:50data collection
1:25:51the interaction in a focus group
1:25:53generates information provides in-depth
1:25:56responses and can address more complex
1:25:59questions or qualitative data
1:26:01it is an excellent platform to get
1:26:03critical to quality or ctq definitions
1:26:06as well
1:26:07on the other hand the disadvantages of
1:26:10focus groups are that the learning only
1:26:12applies to those within the group and it
1:26:14is not possible to generalize
1:26:17the information collected is more
1:26:19qualitative than quantitative which is
1:26:21another drawback
1:26:23additionally they can also generate a
1:26:25lot of information from anecdotes and
1:26:27incidents experienced by the individuals
1:26:30in the group
1:26:31in the next screen we will discuss the
1:26:33interview technique
1:26:36this screen discusses advantages and
1:26:39disadvantages of using the interview
1:26:41technique for data collection
1:26:43interviews have a capability to handle
1:26:45complex questions and a large amount of
1:26:48information
1:26:49they also allow us to use visual aids in
1:26:52the process
1:26:53it is a better method to be employed
1:26:55when people do not respond willingly and
1:26:57or accurately by phone or email
1:27:00however there are some shortcomings as
1:27:02well
1:27:03interviews are time consuming and the
1:27:05resources or interviewer needs to be
1:27:07trained and experienced to carry out the
1:27:10task
1:27:11let us discuss the importance and
1:27:13urgency of these inputs in the next
1:27:15screen
1:27:19the table shows the importance and
1:27:21urgency of different kinds of input
1:27:24to understand the kind of input to be
1:27:26chosen
1:27:27different kinds of methods for
1:27:28collecting data are identified
1:27:31telephone survey
1:27:32web survey and interview are the data
1:27:34collection methods identified
1:27:37to select the best methods the criteria
1:27:40or the factors which are important to
1:27:42the organization are listed
1:27:44the criteria are the factors based on
1:27:46which an organization is going to make
1:27:48decisions
1:27:49the list of factors is then given
1:27:51weightage based on the importance of
1:27:53each factor in decision making
1:27:55as seen cost is the most important
1:27:58criterion for which the weightage is
1:27:59given twenty
1:28:01response rate of the customer is next
1:28:03important factor and the list follows
1:28:06visualizing feature and compiling and
1:28:08analyzing data are the factors which
1:28:10have the lowest impact on the decision
1:28:12of selecting the methods for data
1:28:14collection
1:28:16each of the data collecting methods is
1:28:17rated between 1 and 10
1:28:19based on its impact on the listed
1:28:21factors
1:28:22with 10 being highly favorable to the
1:28:24organization and one being least
1:28:27favorable
1:28:28after rating all the methods with the
1:28:30factors listed the sum or total is
1:28:33calculated
1:28:34the calculation of the total involves
1:28:36multiplying each methods rating with the
1:28:38factor weightage and adding all the
1:28:40multiplied values of the column that is
1:28:43for telephone survey the rating is
1:28:45multiplied with factors rating
1:28:48eight multiplied by twelve plus eight
1:28:50multiplied by six plus three multiplied
1:28:53by twenty plus five multiplied by five
1:28:57plus three multiplied by five plus seven
1:29:00multiplied by fifteen plus one
1:29:03multiplied by ten plus seven multiplied
1:29:06by three plus zero multiplied by two
1:29:09plus three multiplied by two plus one
1:29:13multiplied by ten plus seven multiplied
1:29:16by five plus eight multiplied by five
1:29:20and the total of this is four hundred
1:29:22seventy one in a similar way calculate
1:29:26the total value for the remaining two
1:29:28methods
1:29:29the total of other two methods are 744
1:29:32and 522 respectively looking at the
1:29:36overall total of the methods 744 is the
1:29:39highest
1:29:40hence web survey is the best method for
1:29:43the organization to use for data
1:29:45collection
1:29:46let us look at the pros and cons of
1:29:48customer complaints data in the next
1:29:50screen
1:29:52there are pros and cons in gathering
1:29:54information from customer complaints
1:29:56advantages include availability of
1:29:58specific feedback directly from the
1:30:00customer
1:30:01and ease in responding appropriately to
1:30:04every customer on the contrary feedback
1:30:07in this method does not provide an
1:30:09adequate sample size and may lead to
1:30:11processed changes based on one or two
1:30:14inputs from the customer
1:30:16the next screen will discuss the
1:30:17difference between product complaint and
1:30:19expedited service request
1:30:24product complaints and expedited service
1:30:26requests can act as inputs to the
1:30:28company for improving their process
1:30:31these details address the needs of the
1:30:33customer in an indirect way
1:30:35a product complaint means that the
1:30:37customer is not happy with the product
1:30:39that he has purchased from the company
1:30:41an expedited service request means a
1:30:44service request is being rushed if the
1:30:47customer requires the items immediately
1:30:49then an expedited service request is
1:30:52raised from the customer and the
1:30:54organization tries to fulfill it to
1:30:56please the customer
1:30:58product complaint implies that a product
1:31:00is not meeting customers specification
1:31:02hence it has to be improved
1:31:04expedited service request implies that
1:31:07service timeliness are not meeting
1:31:09customer requirements hence service has
1:31:11to be improved
1:31:13product complaint also implies that the
1:31:15customer needs for product are not
1:31:17completely identified whereas expedited
1:31:20service request implies that the
1:31:22customer timings need to be recalculated
1:31:25let us discuss the importance and
1:31:28urgency of these inputs in the next
1:31:30screen
1:31:33the table shows the importance and
1:31:35urgency of different kinds of input
1:31:38to select the best methods the criteria
1:31:40or the factors which are important to
1:31:42the organization are listed
1:31:45the criteria are the factors based on
1:31:46which organization is going to make
1:31:48decisions
1:31:50these factors are then given weightage
1:31:52based on the importance of each factor
1:31:53in decision making
1:31:55as seen cost involved and identification
1:31:58of customer need are the most important
1:32:01criteria for which the weightage given
1:32:03is 15 and the list follows
1:32:06time consumption and compiling and
1:32:08analyzing data are the factors which
1:32:10have the least impact on the decision of
1:32:12selecting the methods for data
1:32:13collection
1:32:15each method is rated between 1 and 10
1:32:17based on its impact of the listed
1:32:19factors with 10 being highly favorable
1:32:22to the organization and one being least
1:32:24favorable to the organization
1:32:27after reading all the methods with the
1:32:29factors listed the summer total is
1:32:31calculated
1:32:32calculation of the total is derived by
1:32:34multiplying each methods rating with the
1:32:36factor weightage and adding all the
1:32:38multiplied values that is for product
1:32:41complaint the rating is multiplied with
1:32:43the factors rating
1:32:45eight multiplied by fifteen plus four
1:32:47multiplied by fifteen plus three
1:32:50multiplied by two plus one multiplied by
1:32:53ten plus one multiplied by ten plus one
1:32:56multiplied by ten plus one multiplied by
1:33:00eight plus one multiplied by ten plus
1:33:03four multiplied by two plus one
1:33:05multiplied by eight plus one multiplied
1:33:08by ten
1:33:09and the total of this is two hundred
1:33:11sixty
1:33:12in the similar way calculate the total
1:33:15value for expedited service request
1:33:18the total of expedited service request
1:33:20is eight hundred seventeen and hence it
1:33:22is effective to the organization
1:33:25let us discuss the key elements of data
1:33:27collection tools in the next screen
1:33:31data collection tools will be selected
1:33:33based on the type of data to be
1:33:35collected
1:33:36the key elements that make these tools
1:33:38effective are as follows
1:33:40data is collected directly from the
1:33:42primary source or customer hence there
1:33:45is no scope for miscommunication or loss
1:33:47of information
1:33:49data is collected exclusively for the
1:33:51stated purpose
1:33:53hence data is highly reliable
1:33:55the data is captured is after
1:33:57understanding the organizational purpose
1:34:00this makes the data exclusively relevant
1:34:02and serves the purpose of the
1:34:04organization
1:34:05data is collected instantaneously when
1:34:08there is a requirement
1:34:10this ensures that the data is up to date
1:34:12hence the data is valid
1:34:14the tools accurately define customer
1:34:17requirements the customer requirements
1:34:19could be current needs or improvement to
1:34:21the product or service that they are
1:34:23currently using
1:34:25the tools help to get enough information
1:34:26about customer requirement through which
1:34:29the process for improving or creating
1:34:31the product or service that the customer
1:34:33requires can be developed
1:34:35in the next screen we will discuss how
1:34:37the collected data can be reviewed
1:34:41collated data must be reviewed to
1:34:43eliminate vagueness ambiguity and any
1:34:45unintended bias
1:34:47nutra worldwide buys laptops for its
1:34:50employees from a company that is into
1:34:52manufacture and sales of laptops
1:34:54the company also provides servicing and
1:34:56repairs for their products to the
1:34:58customers
1:34:59to understand the level of customer
1:35:01satisfaction in neutral worldwide and to
1:35:04improve its process the laptop company
1:35:06is conducting a survey
1:35:08the questionnaire which was prepared
1:35:10initially questionnaire before review
1:35:13had questions that led to ambiguity
1:35:15vagueness and unintended bias
1:35:18let us look at each item on the survey
1:35:21to understand the level of usage of the
1:35:24laptop and to know their customer better
1:35:27the survey is raising a question related
1:35:29to the occupation of the customer
1:35:31it gives the option of student or
1:35:33professional
1:35:34but with this low amount of information
1:35:36the company is neither able to gather
1:35:38the information nor will the given
1:35:40option cover the entire possible
1:35:42occupation in the market
1:35:44including an option of other please
1:35:46specify would help the customer to
1:35:48choose and provide the information if he
1:35:51does not belong to one of the two given
1:35:53groups
1:35:54hence the same is added in the review so
1:35:57that the customer will not be in any
1:35:59ambiguity while filling the
1:36:01questionnaire
1:36:02the question whether the sales executive
1:36:04was supportive with an option of yes or
1:36:06no
1:36:07is a question which leads again to
1:36:09ambiguity and unintended bias
1:36:12the customer might be partially happy or
1:36:15partially not happy
1:36:17but the choice does not let them inform
1:36:19their exact feeling
1:36:21if the customer selects no as the option
1:36:23then the company does not get enough
1:36:25information to understand where their
1:36:27sales executive went wrong
1:36:29hence in the reviewed questionnaire the
1:36:31customer is asked to rate the qualities
1:36:33of their sales executive which will
1:36:35provide better data to the company in
1:36:37order to improve the process
1:36:40next we will discuss a technique named
1:36:42voice of customer
1:36:46the voice of customer is a technique to
1:36:48organize analyze and profile the
1:36:50customer's requirements
1:36:52voice of the customer is an expression
1:36:54for listening to the external customer
1:36:57the table shows the customer
1:36:59requirements while purchasing an air
1:37:01conditioner
1:37:02in all cases the customer is purchasing
1:37:04for his or her domestic usage
1:37:07each customer is further categorized
1:37:09according to his needs and requirements
1:37:12when the customer says that he needs a
1:37:14silent air conditioner he needs sound
1:37:16sleep at night in the bedroom
1:37:18this is primarily to remain fresh the
1:37:21next morning and to get rid of the noisy
1:37:23ceiling fan being used currently
1:37:26in case the customer says that he needs
1:37:28an efficient ac
1:37:29he needs a machine which provides good
1:37:31cooling at night in the bedroom this is
1:37:34mainly because it gets extremely hot in
1:37:36summer
1:37:37also he currently uses a ceiling fan
1:37:40which is not so effective in summers
1:37:43on the other hand when the customer
1:37:44wants to buy an ac which is not too
1:37:47costly he has limited cash for the
1:37:49purchase
1:37:50he wants to purchase a low-cost ac
1:37:53let us discuss the importance of
1:37:55translating customer requirements in the
1:37:57next screen
1:38:00customer requirement is the data
1:38:02collected from customers that gives
1:38:04information about what they need or want
1:38:06from the process
1:38:08customer requirements are often high
1:38:10level vague and non-specific
1:38:12some customers may give you a set of
1:38:14specific requirements to the business
1:38:16but broadly customers requirements are a
1:38:19reflection of their experience
1:38:21customer requirements when translated
1:38:24into critical process requirements that
1:38:26are specific and measurable are called
1:38:28critical to quality ctq factors
1:38:32a fully developed ctq has four major
1:38:35elements
1:38:36output characteristic
1:38:38y metric target and specification or
1:38:41tolerance limits
1:38:43we will discuss the meaning of ctq in
1:38:46the next screen
1:38:49let us understand what quality function
1:38:51deployment is in this screen
1:38:54qfd is a process to ensure that the
1:38:56customers wants and needs are heard and
1:38:59translated into technical
1:39:00characteristics
1:39:02it is also known as the voice of the
1:39:04customer or house of quality
1:39:07qfd is a process to understand the
1:39:09customer's needs and translate them into
1:39:12a set of design and manufacturing
1:39:14requirements
1:39:15while motivating businesses to focus on
1:39:18their customers
1:39:20it also helps companies to design and
1:39:22build more competitive products in less
1:39:24time and lesser costs
1:39:26qfd helps in prioritizing customer
1:39:29requirements recognizing strengths and
1:39:32weaknesses of an organization and
1:39:34recognizing areas that need to be worked
1:39:36on and areas that need immediate focus
1:39:38of efforts
1:39:40qfd is carried out by asking relevant
1:39:42questions to the customers and
1:39:44tabulating them to bring out a set of
1:39:46parameters critical to the product
1:39:48design
1:39:49let us discuss phases of qfd in the next
1:39:53screen
1:39:55quality function deployment involves
1:39:58four phases
1:39:59phase one product planning
1:40:01in this phase the qfd team translates
1:40:04the customer requirements into product
1:40:06technical requirements
1:40:08phase two
1:40:09product design
1:40:11in this phase the qfd team translates
1:40:14the identified technical requirements
1:40:16into key part characteristics or systems
1:40:20phase three process planning
1:40:23in this phase the qfd team identifies
1:40:26the key process operations necessary to
1:40:29achieve the identified key part
1:40:31characteristics
1:40:32phase four
1:40:34production planning or process control
1:40:37in this phase the qfd team establishes
1:40:40process control plans maintenance plans
1:40:43and training plans to control operations
1:40:46next we will understand the structure of
1:40:48qfd
1:40:50let us see what happens after completing
1:40:53the hoq matrix
1:40:55completing one hoq matrix is not the end
1:40:58of the qfd process
1:41:01the output of the first hoq matrix can
1:41:04be the first stage of the second qfd
1:41:07phase
1:41:08as shown in the image the translation
1:41:10process is continued using linked hoq
1:41:13type matrices until the production
1:41:16planning targets are developed
1:41:18let us proceed to the next topic of this
1:41:20lesson in the following screen
1:41:23in this topic we will discuss the basics
1:41:26of project management
1:41:28let us start with a discussion on
1:41:30problem statement
1:41:32every six sigma project targets a
1:41:35problem that needs to be resolved the
1:41:37first step of project initiation is
1:41:39defining the problem statement
1:41:42a problem statement needs to describe
1:41:44the problem in a clear and concise
1:41:46manner
1:41:47a problem statement needs to identify
1:41:49and specify the observed gap in
1:41:51performance
1:41:53it should indicate the current
1:41:54performance state of a process and the
1:41:56required performance state completely
1:41:58derived from customer requirements
1:42:01a problem statement should be
1:42:02quantifiable
1:42:04this means it should have specified
1:42:06metrics including the respective units
1:42:09please note that the problem statement
1:42:11cannot contain solutions or causes for
1:42:13the problem
1:42:15in the next screen we will discuss the
1:42:17is or is not template
1:42:21the is or is not technique was first
1:42:23popularized by kepner trigo incorporated
1:42:26in the seventies it is a powerful tool
1:42:29that helps define the problem and gather
1:42:31required information
1:42:33an example of a problem statement of
1:42:35paper cup leaks is given on the screen
1:42:39the six sigma team has to answer what is
1:42:42the problem what isn't the problem where
1:42:44is it where isn't it when is it when
1:42:48isn't it the problem to what extent is
1:42:50it a problem and to what extent isn't it
1:42:53a problem
1:42:54the information is then used to fill the
1:42:56question areas in the is and is not
1:42:59issue template in the analysis phase if
1:43:02a cause cannot describe the is and the
1:43:05is not data then it's not likely the
1:43:07main cause
1:43:09in the next screen we will list the
1:43:10criteria for the project objectives
1:43:15the project objectives must meet the
1:43:18smarts criteria
1:43:19smarts is an acronym of the
1:43:21characteristics desired in project
1:43:23objectives
1:43:25specific
1:43:26measurable attainable
1:43:28relevant
1:43:29time based and stretch
1:43:32the project deliverables should be
1:43:34specific
1:43:35example
1:43:36hospitals maintain records of all
1:43:38patients
1:43:39often a few forms are rejected or missed
1:43:43due to errors in recording the id
1:43:45numbers
1:43:46in this case setting the objective as
1:43:49reduce form rejection is very vague
1:43:52instead reduce patient id errors in
1:43:55recording lab results is specific and
1:43:58effectively targets solving the problem
1:44:01the project objectives should be
1:44:03quantifiable
1:44:05example setting the objective as fewer
1:44:08form rejections is very vague instead
1:44:11reduce patient id errors by 30 percent
1:44:14sets a specific goal
1:44:16the project objectives should be
1:44:18achievable and practical
1:44:21the project objectives should be
1:44:23relevant to the problem
1:44:25the project objectives must specify a
1:44:28time frame within which they should be
1:44:30delivered
1:44:31the project objectives must not be
1:44:33easily achievable
1:44:35example
1:44:36most problems and errors can be reduced
1:44:39by creating awareness
1:44:41hence the objective must stretch beyond
1:44:43the easily attainable state
1:44:46in the next screen we will understand
1:44:49project documentation
1:44:52project documentation refers to creating
1:44:55documents to provide details about the
1:44:58project
1:44:59such documents are used to gain a better
1:45:01understanding of the project prevent and
1:45:04resolve conflict among stakeholders and
1:45:06share plans and status for the project
1:45:10documentation of a project is critical
1:45:12throughout the project
1:45:14some of the benefits achieved through
1:45:16project documentation are mentioned
1:45:18below
1:45:19documentation serves as written proof
1:45:22for execution of the project
1:45:24it helps teams achieve a common
1:45:26understanding of the requirements and
1:45:28the status of the project
1:45:30it removes personal bias as there is a
1:45:33documented history of discussions and
1:45:36decisions made for the project
1:45:39depending on the nature of the project
1:45:41each project produces a number of
1:45:43different documents
1:45:45some of these documents are the project
1:45:47charter project plan and its subsidiary
1:45:50plans
1:45:52other examples of project documentation
1:45:55include project status reports including
1:45:57key milestones report risk items and
1:46:00pending action items
1:46:02the frequency of these reports is
1:46:04determined by the need and complexity of
1:46:06the project
1:46:08these reports are sent to all
1:46:09stakeholders to keep them abreast of the
1:46:12status of the project
1:46:13another example of project documentation
1:46:16is the final project report
1:46:19this report is prepared at the end of
1:46:20the project and includes a summary of
1:46:23the complete project
1:46:25project storyboard inputs generated from
1:46:28statistical tools
1:46:29outputs from spreadsheets checklists and
1:46:32other miscellaneous documents are also
1:46:34classified as project documents
1:46:37in the next screen we will understand
1:46:40project charter
1:46:45we will list the project charter
1:46:47sections in this screen
1:46:49the major sections of a project charter
1:46:52are project name and description
1:46:54business requirements
1:46:56name of the project manager
1:46:58project purpose or justification
1:47:00including roi
1:47:02stakeholder and stakeholder requirements
1:47:05broad timelines major deliverables
1:47:08constraints and assumptions and the
1:47:10budget summary of the charter
1:47:12in the next screen we will understand
1:47:14the project plan
1:47:18a project plan is the final approved
1:47:20document which is used to manage and
1:47:23control the various processes within the
1:47:25project and ensure its seamless
1:47:28execution
1:47:30the project manager uses the project
1:47:32charter as an input to create a detailed
1:47:35project plan
1:47:36a project plan comprises various
1:47:39sections
1:47:40prominent among them being the project
1:47:42management approach the scope statement
1:47:44the work breakdown structure the cost
1:47:47estimates scheduling defining
1:47:49performance baselines
1:47:51marking major milestones to be achieved
1:47:54and the key members and required staff
1:47:56personnel for the project
1:47:59it also includes the various open and
1:48:01pending decisions related to the project
1:48:04and the key risks involved
1:48:06additionally it also contains references
1:48:08to other subsidiary plans for managing
1:48:11risk scope schedule etc
1:48:14in the next screen we will learn about
1:48:16project scope
1:48:19we will look at different techniques
1:48:21used for interpreting the project scope
1:48:23in this screen
1:48:25project scope can be interpreted from
1:48:27the problem statement and project
1:48:29charter using various tools like the
1:48:31burrito chart and the cypoc map
1:48:34the principle behind the burrito chart
1:48:37or the 80 20 principle as we know it is
1:48:40a vital few trivial many
1:48:43the burrito chart helps the teams to
1:48:45trim the scope of the project by
1:48:48identifying the causes which have a
1:48:50major impact on the outcome of the
1:48:52project
1:48:53the cypac map is a high level process
1:48:56map which helps all team members in
1:48:59understanding the process functions in
1:49:01terms of addressing questions like who
1:49:04are the suppliers
1:49:05what are the inputs they provide
1:49:08what are the outputs that can be
1:49:10obtained and who are the customers
1:49:13as discussed earlier cypoc stands for
1:49:17suppliers inputs process
1:49:20outputs and customers
1:49:22in the subsequent screen we will learn
1:49:24about process
1:49:26maps scipok is a macro level map that
1:49:30provides an overview of the business
1:49:32process
1:49:34where a process map is a micro level
1:49:36flow chart that provides an in-depth
1:49:39detail of a process
1:49:41the process map covers details at all
1:49:43levels and provides a walk through the
1:49:45current process experience
1:49:48the cypac map is used as a basis while
1:49:51drawing a process map
1:49:53a level one process map provides
1:49:55in-depth information
1:49:57but the final process map drills further
1:50:00into detail
1:50:01in the following screen we will
1:50:03understand the project metrics
1:50:07let us discuss consequential metrics in
1:50:09this screen
1:50:11consequential metrics measure any
1:50:13negative consequences
1:50:15these can be business metrics process
1:50:18metrics or both
1:50:20they measure the negative effects of
1:50:22improving the primary or key metrics
1:50:26they are used to measure the indemnity
1:50:28triggered by any damage in the project
1:50:31the inconsistent use of consequential
1:50:34metrics can lead to loss of opportunity
1:50:37and rework after a project ends
1:50:40consequential metrics help to understand
1:50:42the cause and effect relationship
1:50:45between the primary and the secondary
1:50:47metrics and the impact it has on the
1:50:50organization
1:50:52let us take a look at an example for
1:50:54consequential metrics in the next screen
1:50:59we will discuss the best practices in
1:51:01this screen
1:51:02the following are some of the best
1:51:04practices of consequential metrics
1:51:07setting consequential metrics during the
1:51:10measure phase and monitoring these
1:51:12metrics after finalizing the project
1:51:14will help to analyze whether the link
1:51:16between previous primary and secondary
1:51:19metrics has been established
1:51:22also linking consequential metrics with
1:51:24primary metrics and finally linking them
1:51:27with secondary metrics provides clarity
1:51:30on the impact of these metrics
1:51:33assessing and evaluating the cause and
1:51:35effect relationship between these
1:51:37metrics is helpful to the organization
1:51:40as a whole
1:51:41in the next screen we will list some
1:51:43project planning tools
1:51:46the project manager uses various tools
1:51:49to plan and control a project
1:51:52one of the tools which he uses is the
1:51:54burrito chart
1:51:56other prominent tools include the
1:51:57network diagram the critical path method
1:52:01also called cpm
1:52:03the program evaluation and review
1:52:05technique which is also known as pert
1:52:08gantt charts and the work breakdown
1:52:11structure also known as wbs
1:52:15in the next screen we will discuss
1:52:17burrito chart
1:52:20pareto chart is a histogram ordered by
1:52:22the frequency of occurrence of events
1:52:25it is also known as the 80 20 rule or
1:52:29vital few trivial many
1:52:31it helps project teams to focus on the
1:52:33issues which cause the highest number of
1:52:35defects or complaints
1:52:38to explain further the given chart plots
1:52:40all the causes for defects in a product
1:52:43or service the values are represented in
1:52:46descending order by bars
1:52:48and the cumulative total is represented
1:52:50by the line
1:52:52burrito chart emphasizes that 80 percent
1:52:54of the effects come from 20 percent of
1:52:57the causes
1:52:58thus a burrito chart narrows the scope
1:53:00of the project or problem solving by
1:53:03identifying the major causes affecting
1:53:05quality
1:53:06pareto charts are useful only when
1:53:08required data is available
1:53:11if data is not available then other
1:53:13tools such as brainstorming and
1:53:16multi-voting should be used to find the
1:53:18root cause of any problem
1:53:20in the following screen we will continue
1:53:22to discuss burrito chart with an example
1:53:27a hotel receives plenty of complaints
1:53:29from its customers and the hotel manager
1:53:32wishes to identify the key areas of
1:53:34complaints
1:53:35complaints were received in the
1:53:37following areas cleaning check-in pool
1:53:40timings minibar room service and
1:53:43miscellaneous
1:53:45cleaning and check-in can be noted as
1:53:47areas of concern with 35 and 19
1:53:50complaints respectively
1:53:53is calculated for each cause of
1:53:55complaint and the cumulative is derived
1:53:58burrito chart is plotted using this data
1:54:01in the next screen we will discuss
1:54:03network diagrams
1:54:07network diagrams are one of the tools
1:54:09used by the project manager for project
1:54:12planning
1:54:13they are also sometimes referred to as
1:54:16arrow diagrams because they use arrows
1:54:18to connect activities and represent
1:54:20precedence and interdependencies between
1:54:23activities of a project
1:54:25there are some assumptions that need to
1:54:27be made while forming the network
1:54:29diagram
1:54:30the first assumption is that before a
1:54:32new activity begins all pending
1:54:35activities have been completed
1:54:37the second assumption is that all arrows
1:54:40indicate logical precedence
1:54:42this means that the direction of the
1:54:44arrow represents the sequence that
1:54:46activities need to follow
1:54:48the last assumption is that a network
1:54:51diagram must start from a single event
1:54:53and end with a single event
1:54:55there cannot be multiple start and end
1:54:58points to the network diagram in the
1:55:00next screen let us discuss some terms
1:55:02related to network diagram
1:55:06for the network diagram to calculate the
1:55:08total duration of the project the
1:55:10project manager needs to define four
1:55:13dates for each task
1:55:15the first two dates relate to the date
1:55:17by when the task can be started
1:55:19the first date is early start this is
1:55:22the earliest date by when the task can
1:55:24start
1:55:26the second date is late start this is
1:55:28the last date by when the task should
1:55:31start the second two dates relate to the
1:55:33dates when the task should be complete
1:55:36early finish is the earliest date by
1:55:38when the task can be completed
1:55:41late finish is the last date by when the
1:55:44task should be completed
1:55:46the duration of the task is calculated
1:55:49as the difference between the early
1:55:51start and early finish of the task
1:55:54the difference between the early start
1:55:56and late start of the task is called the
1:55:58slack time available for the task
1:56:01slack can also be calculated as the
1:56:03difference between the early finish and
1:56:06late finish dates of the task
1:56:08slack time or float time for a task is
1:56:11the amount of time the task can be
1:56:13delayed before it causes a delay in the
1:56:16overall project timeline
1:56:19in the next screen we will discuss
1:56:21critical path method
1:56:24critical path method also known as cpm
1:56:28is an important tool used by project
1:56:30managers to monitor the progress of the
1:56:33project and to ensure that the project
1:56:35is on schedule
1:56:37the critical path for a project is the
1:56:40longest sequence of tasks on the network
1:56:42diagram
1:56:44the critical path in the given network
1:56:45diagram is highlighted in orange
1:56:48critical path is characterized by zero
1:56:51slack for all tasks on the sequence
1:56:54this means that the smallest delay in
1:56:56any of the tasks on the critical path
1:56:59will cause a delay in the overall
1:57:01timeline of the project
1:57:03this makes it very important for the
1:57:05project manager to closely monitor the
1:57:08tasks on the critical path and ensure
1:57:10that the tasks go smoothly
1:57:13if needed the project manager can divert
1:57:15resources from other tasks that are not
1:57:18on the critical path to task on the
1:57:20critical path to ensure that the project
1:57:22is not delayed
1:57:24when a project manager removes resources
1:57:27from such tasks he needs to ensure that
1:57:30the task does not become a critical path
1:57:32task because of the reduced number of
1:57:34resources
1:57:36during the execution of the project
1:57:38the critical path can easily shift
1:57:40because of multiple factors and hence
1:57:43needs to be constantly monitored by the
1:57:45project manager
1:57:47a complex project can also have multiple
1:57:49critical paths in the next screen we
1:57:52will discuss project evaluation and
1:57:54review technique
1:57:57we will understand the concept of risk
1:58:00in this screen
1:58:02risk is an uncertain event or
1:58:04consequence probable of occurring during
1:58:07a project
1:58:08the main objectives of any project are
1:58:11time cost quality and scope
1:58:14risk affects at least one of the four
1:58:16project objectives
1:58:18it is important to understand that risk
1:58:20can be both positive as well as negative
1:58:24a positive risk enhances the success of
1:58:26the project whereas a negative risk is a
1:58:29threat to a project's success
1:58:32some of the terms used in risk analysis
1:58:34and management are risk probability
1:58:37issue and risk consequences
1:58:40the likelihood that a risk will occur is
1:58:42called risk probability
1:58:44to assess any risk is to assess the
1:58:46probability and impact of the risk
1:58:50issue is the occurrence of a risk
1:58:53risk consequences are the effects on
1:58:55project objectives if there is an
1:58:58occurrence of a risk or issue
1:59:01in the subsequent screen we will
1:59:03understand the process of risk analysis
1:59:06and management
1:59:09we will list and understand some of the
1:59:11elements of risk analysis in this screen
1:59:15qualitative method
1:59:17qualitative methods like interview
1:59:19checklists and brainstorming are used to
1:59:21identify potential risks
1:59:24quantitative method
1:59:26quantitative methods are databased and a
1:59:29computer is required to calculate and
1:59:31analyze
1:59:33these methods are used to evaluate the
1:59:35cost time and probabilistic combination
1:59:38of individual uncertainties
1:59:41feasibility
1:59:42feasibility is the study of the project
1:59:45risk
1:59:46this is usually carried out in the
1:59:47beginning of the project when the
1:59:49project is most flexible and risks can
1:59:52be reduced at a relatively low cost
1:59:55it helps in deciding different
1:59:57implementing options for the projects
2:00:00potential impact
2:00:02once the potential risks are identified
2:00:05the impact of these on the project is
2:00:07determined
2:00:08using this data possible solutions for
2:00:11the risks are identified
2:00:14rpn
2:00:15rpn of a failure is the product of its
2:00:18probability of occurrence severity and
2:00:21detectability
2:00:22a failure is prioritized based on its
2:00:25rpn value
2:00:27a high rpn indicates high risk
2:00:30rpn assists to prioritize risks
2:00:34avoiding risk
2:00:36when potential risks are identified
2:00:38their impact in terms of cost time
2:00:41resources and objective perspective is
2:00:44calculated
2:00:45if the impact is huge then avoiding the
2:00:48risk is the best option
2:00:51mitigating risk
2:00:52mitigating is the second option when
2:00:54dealing with risks
2:00:56the loss that arises from mitigating a
2:00:58risk is much less than the loss that
2:01:01arises from the temporary avoiding a
2:01:03risk
2:01:05accepting the risk
2:01:07if a risk cannot be avoided or mitigated
2:01:10then it has to be accepted
2:01:12the risk will be accepted if it doesn't
2:01:14greatly impact the cost time and product
2:01:16objective
2:01:18in the following screen we will discuss
2:01:20benefits of risk analysis
2:01:24benefits of risk analysis are as follows
2:01:27once the risk has been identified it can
2:01:30be either mitigated transferred or
2:01:33accepted
2:01:34when risk is identified in a task slack
2:01:37time is provided as a buffer identifying
2:01:40risks also helps in setting up an actual
2:01:42timeline for a project
2:01:45slack time for an activity in a project
2:01:47could be the result of a risk identified
2:01:50proactively
2:01:51identifying risks helps in setting
2:01:54realistic expectations from the project
2:01:57by communicating the risk probability
2:01:59and consequence to stakeholders
2:02:02risk analysis also helps to identify and
2:02:05plan contingency activities if the risk
2:02:08becomes an issue
2:02:10the project team is then well prepared
2:02:12to work on the issue thereby reducing
2:02:15the impact of the risk
2:02:17in the following screen we will take a
2:02:19look at the risk assessment matrix
2:02:23the potential risks of a project are
2:02:25assessed using the risk assessment
2:02:27matrix
2:02:29it covers potential risk areas like
2:02:31project scope team personnel material
2:02:35facility and equipment and communication
2:02:38each of these areas is assessed in terms
2:02:40of risk of loss of money productivity
2:02:43resources and customer confidence
2:02:46in the subsequent screen we will discuss
2:02:49project closure
2:02:51by definition a project has a beginning
2:02:54and an end
2:02:56but without a formal closure process
2:02:58project teams can fail to recognize the
2:03:01end and then the project can drag on
2:03:04sometimes at great expense
2:03:06every project requires closure
2:03:09for larger complex projects it's a good
2:03:11idea to close each major project phase
2:03:14for example design code and test
2:03:17individually
2:03:19project closure ensures that outcomes
2:03:22match the stated goals of the project
2:03:24customers and stakeholders are happy
2:03:26with the results
2:03:28critical knowledge is captured
2:03:30the team feels a sense of completion
2:03:33and project resources are released for
2:03:35new projects
2:03:37in the next screen we will list the
2:03:39goals of a project closure report
2:03:43the project closure report is created to
2:03:46accomplish the following goals
2:03:48review and validate the success of the
2:03:50project
2:03:52confirm outstanding issues limitations
2:03:55and recommendations
2:03:57outline tasks and activities
2:03:59accomplished to complete the activity
2:04:02highlight the best practices for future
2:04:04projects
2:04:06provide the project report or summary
2:04:09provide a project background overview
2:04:13summarize the planned activities of a
2:04:15project
2:04:16evaluate project performance
2:04:19provide a synopsis of the process
2:04:23generate discussions and recommendations
2:04:26generate project closure recommendations
2:04:30in the following screen we will list and
2:04:33understand project closure activities
2:04:37during project closure the project
2:04:40manager needs to take care of the
2:04:41following activities
2:04:43finalize the project documents
2:04:46much of a project's documentation is
2:04:49created during the life of the project
2:04:52document collection and update
2:04:53procedures are well established during
2:04:55the project lifeline
2:04:58capture the project knowledge
2:05:00project documents are helpful for future
2:05:03projects in troubleshooting the product
2:05:05or in a future audit
2:05:08set up a project library
2:05:10ideally the project library is set up at
2:05:13the beginning of the project
2:05:15and team members add documents as they
2:05:17produce them
2:05:19document the project learnings
2:05:22project learnings can be captured
2:05:23through team meetings meetings with
2:05:26stakeholders and sponsor
2:05:28and through feedbacks from consultants
2:05:30and vendors
2:05:32provide knowledge transfer
2:05:34the project manager needs to provide a
2:05:37summary of the project results to team
2:05:39members
2:05:40either as a presentation at a meeting or
2:05:43as a formal document
2:05:45consultants should not be relieved from
2:05:47their position until they have
2:05:48transferred all the important product
2:05:50maintenance knowledge to the team
2:05:53get a final sign off
2:05:55schedule a meeting with the project
2:05:57sponsor and key stakeholders to get
2:05:59their final sign off on the project
2:06:02close the project office
2:06:04if the project team used a project
2:06:06management office or a dedicated work
2:06:09area
2:06:09arrangements need to be made to return
2:06:11that space for general use
2:06:14recognize and reward
2:06:17the project manager has the best
2:06:19understanding of which of the team
2:06:20members have worked the best have
2:06:23transformed themselves with new skills
2:06:25and who might be ready for a new level
2:06:27of responsibility
2:06:29the project manager needs to report to
2:06:31the team superiors what each team member
2:06:34has brought to the project
2:06:36celebrate
2:06:38after completion of every project the
2:06:40team needs and deserves a celebration
2:06:43a team dinner a team outing gift
2:06:46certificates or other rewards are minor
2:06:48costs that generate a large return in
2:06:51terms of morale and job satisfaction
2:06:54make a public announcement
2:06:57an announcement to the organization is a
2:06:59good way to highlight the success of the
2:07:01project and its benefits to the company
2:07:04conclude
2:07:06formal project closure ensures that the
2:07:08team has met its objectives satisfied
2:07:11the customer
2:07:12captured important knowledge and been
2:07:14rewarded for their efforts
2:07:17let us proceed to the next topic of this
2:07:19lesson in the following screen
2:07:22in this topic we will discuss management
2:07:25and planning tools
2:07:26let us start with the discussion on
2:07:28affinity diagram in the next screen
2:07:32the affinity diagram method is employed
2:07:35by an individual or team to solve
2:07:37unfamiliar problems it is an effective
2:07:40medium where the consensus of the group
2:07:43is necessary
2:07:44the given affinity diagram is based on
2:07:47an organization where the employees are
2:07:49not satisfied
2:07:51to begin with each member writes down
2:07:53ideas and opinions on sticky notes
2:07:56each note can have a single idea
2:07:59the points which have surfaced during
2:08:01the brainstorming session are that the
2:08:03workers are unkind
2:08:05pay is low
2:08:06and it is difficult to survive on the
2:08:08pay structure working hours are too long
2:08:11etc
2:08:12in the next step all the sticky notes
2:08:14are pasted on a table or wall
2:08:17the sticky papers are arranged according
2:08:19to categories or thought patterns
2:08:22members happen to arrange their ideas
2:08:24based on affinity
2:08:26in case a particular idea is good to go
2:08:28into more than one category it is
2:08:31duplicated and added to several
2:08:32categories
2:08:34after the arrangement is done
2:08:36each category is named with a header
2:08:39card
2:08:40the header card captures the central
2:08:42idea of all the cards in that category
2:08:45and draws a boundary around them
2:08:48poor compensation combines ideas like
2:08:50low pay long working hours and
2:08:53complaints about wages
2:08:55poor work environment encompasses issues
2:08:57like poor lighting
2:08:59uncomfortable rooms and stuffy air
2:09:02similarly poor relationships prevail in
2:09:05the workspace as the workers are unkind
2:09:08and there is mutual dislike
2:09:11lack of motivation is due to repetitive
2:09:13work and no work related challenges
2:09:16you can see in the diagram on that slide
2:09:19that once all the ideas are grouped to
2:09:21the respective header cards
2:09:23a diagram is drawn and borders are
2:09:25placed around the group of ideas
2:09:28thus affinity diagram helps in grouping
2:09:31ideas with the common theme
2:09:33in the next screen we will discuss the
2:09:35interrelationship diagram
2:09:40during problem solving the
2:09:42inter-relationship diagram technique
2:09:45helps in identifying the relationship
2:09:47between problems and ideas in complex
2:09:50situations
2:09:52if the problem is really complex it may
2:09:54not be easy to determine the exact
2:09:56relationship between ideas
2:09:59given inter-relationship diagram is the
2:10:02result of a team brainstorming session
2:10:05which identified 10 major issues
2:10:07involved in developing an organization's
2:10:09quality plan
2:10:11initially the problem is defined and all
2:10:14the members put down their ideas on
2:10:16sticky notes
2:10:18each note contains only one idea
2:10:21all the sticky notes are put on a table
2:10:23for a random display
2:10:25in the next step the causes or areas of
2:10:28concern are identified
2:10:31and a cause effect arrangement of cards
2:10:33is constructed by drawing an arrow
2:10:35between the causes and effects of the
2:10:37cause
2:10:39this is done until all the ideas on the
2:10:41sticky notes are accounted for and made
2:10:43a part of the interrelationship diagram
2:10:47take a large sheet of paper and
2:10:49replicate the cause effect arrangement
2:10:51on it
2:10:52as depicted in the image a large number
2:10:55of outgoing arrows indicate the root
2:10:57cause
2:10:58whereas a higher number of incoming
2:11:01arrows indicates an outcome
2:11:03there are as many as six arrows
2:11:05originating from lack of quality
2:11:07strategy
2:11:09this leads us to understand that it is a
2:11:11root cause
2:11:13on the other hand there are three arrows
2:11:15ending with the idea lack of tqm
2:11:18commitment by managers
2:11:20making it an outcome
2:11:22in the next screen we will understand
2:11:25the tree diagram
2:11:28the tree diagram is a systematic
2:11:31approach to outline all the details
2:11:32needed to complete a given objective
2:11:35in other words it is a method used to
2:11:38identify the tasks and methods needed to
2:11:40solve a problem and reach a predefined
2:11:43goal
2:11:44it is mostly used while developing
2:11:46actions to execute a solution while
2:11:48analyzing processes in detail during the
2:11:51evaluation of implementation issues for
2:11:54several potential solutions
2:11:56and also as a communication tool to
2:11:59explain the details of a process to
2:12:01others
2:12:02the given tree diagram shows the plan of
2:12:05a coffee shop trying to set standards
2:12:07for the coffee it delivers
2:12:10first the objective is noted on a note
2:12:12card and placed on the far left side of
2:12:14the board
2:12:15the basic goal of the coffee shop is to
2:12:18provide a delightful cappuccino
2:12:20experience
2:12:21in the next step the coffee shop needs
2:12:24to determine the means required to
2:12:25achieve the goal and furnish three
2:12:27different solutions
2:12:29in other words the answers to the how or
2:12:32why questions of the objectives
2:12:34in this case the cappuccino needs to be
2:12:37at a comfortable temperature
2:12:39and it should have strong and pleasing
2:12:41coffee aroma with the right amount of
2:12:43sweetness
2:12:45in the next step the three issues
2:12:47mentioned in the second stage are
2:12:49addressed at length
2:12:50each issue is answered
2:12:53by maintaining the espresso and steamed
2:12:55milk temperature the cappuccino can be
2:12:58served at a palatable temperature
2:13:00strong flavored cappuccino can be
2:13:02prepared using a good amount of finely
2:13:04ground coffee beans
2:13:06and a good quality sweetener used in the
2:13:08right amount makes a great cappuccino
2:13:12thus the tree diagram can be used to
2:13:14achieve a goal or define a process
2:13:18in the following screen we will discuss
2:13:20prioritization matrices
2:13:24let us learn about matrix diagram in
2:13:26this screen
2:13:28matrix diagrams show the relationship
2:13:31between objectives and methods results
2:13:34and causes
2:13:35tasks and people etc
2:13:38their objective is to provide
2:13:39information about the relationship
2:13:42they provide importance of task and
2:13:44method elements of the subject
2:13:47they also help determine the strength of
2:13:49relationships between a grid of rows and
2:13:51columns
2:13:52they help in organizing a large amount
2:13:54of inter-process related activities
2:13:58let us discuss various types of matrices
2:14:01in the next screen
2:14:04let us learn about a process decision
2:14:06program chart in this screen
2:14:09process decision program chart or the
2:14:12pdpc method
2:14:14is used to chart the course of events
2:14:17from the beginning of a process till the
2:14:19goal
2:14:20while emphasizing the ability to
2:14:22identify the failure of important issues
2:14:25on activity plans
2:14:27the pdpc helps create appropriate
2:14:29contingency plans to limit the number of
2:14:32risks involved
2:14:34the pdpc is used before implementing a
2:14:37plan
2:14:38especially when the plan is large and
2:14:40complex if the plan must be completed on
2:14:43schedule or if the price of failure is
2:14:46quite high
2:14:48the given process decision program chart
2:14:51shows the process which can help in
2:14:53securing a contract
2:14:55the process starts when the seller
2:14:57receives an order request from a
2:14:59potential buyer
2:15:01this can lead to fixing an appointment
2:15:03with the buyer confirming the
2:15:05appointment date and meeting the buyer
2:15:08if the date is not fixed then buyers
2:15:10should be contacted till the meeting is
2:15:12confirmed
2:15:13without a meeting there is a risk of
2:15:15losing the order
2:15:17considering an optimistic scenario where
2:15:20a meeting is fixed with the buyer
2:15:22the seller describes the price of the
2:15:24product or service
2:15:26if the price is competitive the order is
2:15:28secured
2:15:30if the price is not competitive the
2:15:32seller may have to repeat the bid until
2:15:34the buyer agrees and the order is secure
2:15:38however the buyer may not agree to a
2:15:40revised bid either
2:15:42in which case the seller might lose the
2:15:44bid
2:15:46in such a scenario
2:15:47the seller can justify the pricing and
2:15:50pursue the buyer to agree to the bid
2:15:53it might work and the seller might
2:15:55secure the order
2:15:57in the next screen we will discuss the
2:15:59activity network diagram
2:16:03an activity network diagram is used to
2:16:06show the time required for solving a
2:16:08problem and to identify items that can
2:16:10be done in parallel
2:16:12it is used in scheduling and monitoring
2:16:14tasks within a complex project or
2:16:17process with interrelated tasks and
2:16:19resources moreover it is also used when
2:16:22you know the steps of the project or
2:16:24process their sequence and the time
2:16:26taken by each of the steps involved
2:16:29the original japanese name for this tool
2:16:32is arrow diagram
2:16:34the given activity network diagram shows
2:16:36a house construction plan and identifies
2:16:39the factors involved separately
2:16:42like the amount of time for each
2:16:43operation in one situation the
2:16:46relationship of work without time for
2:16:48each operation and an operation by
2:16:50itself
2:16:52the number of days is denoted by d
2:16:55so the time taken for an activity like
2:16:57foundation to scaffolding takes around
2:17:00five days plus four days which is nine
2:17:03days in total
2:17:05the line joining electrical work and
2:17:07interior walls is dotted
2:17:10this shows relation between them but
2:17:12without any timeline
2:17:15basically it means that electrical work
2:17:17has to be done before interior walls
2:17:20but the time is either not important or
2:17:23not available
2:17:24let us proceed to the next topic of this
2:17:27lesson in the following screen
2:17:31in this topic we will introduce business
2:17:33results for projects
2:17:35let us start with the discussion on
2:17:37defect per unit
2:17:40we will learn about throughput yield in
2:17:42this screen
2:17:43throughput yield or tpy
2:17:46is the number of acceptable pieces at
2:17:48the end of a process divided by the
2:17:51number of starting pieces excluding
2:17:53scrap and rework
2:17:55throughput yield is used to measure a
2:17:57single process only
2:17:59if the dpu is known tpy can be easily
2:18:03calculated as e to the power of the
2:18:06negative of dpu
2:18:08here e is the mathematical constant and
2:18:11has a value of 2.7183
2:18:14the expression can also be stated as dpu
2:18:18equals the negative of natural logarithm
2:18:20e of tpy
2:18:23in the next screen we will discuss
2:18:25rolled throughput yield
2:18:29rolled throughput yield or rty
2:18:32is the probability of the entire process
2:18:34producing zero defects
2:18:37rty is the true measure of process
2:18:39efficiency and is considered across
2:18:42multiple processes
2:18:44it is important as a metric when a
2:18:46process has excessive rework
2:18:49tdpu is total defects per unit
2:18:53and is defined for a set of processes
2:18:56when the total defects per unit is known
2:18:59rolled throughput yield is calculated
2:19:01using the expression e to the power of
2:19:04negative of tdpu
2:19:07the expression can also be written as
2:19:09tdpu
2:19:11is equal to negative of natural
2:19:13logarithm of rty
2:19:15when the defectives are known roll
2:19:18through foot yield can be calculated as
2:19:20the product of each process's first pass
2:19:23yield or fpy
2:19:25first pass yield is the number of
2:19:27products which pass without any rework
2:19:30over the total number of units
2:19:32first pass yield is calculated as total
2:19:35number of quality products over total
2:19:37number of units
2:19:39total number of quality products is
2:19:41total number of units minus total number
2:19:44of defective units
2:19:46in the following screen we will
2:19:47understand fby and rty with an example
2:19:55we will discuss process capability in
2:19:58this screen
2:19:59process capability or cp
2:20:02is defined as the inherent variability
2:20:06of a characteristic of a process or a
2:20:08product
2:20:09in other words it might also mean how
2:20:12well a process meets customer
2:20:15cp is an indicator of capability of a
2:20:18process
2:20:19and is expressed as difference of usl
2:20:22and lsl divided by product of six sigma
2:20:27usl stands for upper specification limit
2:20:31lsl is lower specification limit and
2:20:34sigma is the standard deviation of a
2:20:37process
2:20:38the difference between usl and lsl is
2:20:42also called the specification with or
2:20:45tolerance
2:20:46in the following screen we will discuss
2:20:49process capability indices
2:20:53process capability indices or cpk
2:20:56was developed to objectively measure the
2:20:59degree to which a process meets or does
2:21:01not meet customer requirements
2:21:04it was developed to account for the
2:21:05position of mean with respect to usl and
2:21:09lsl
2:21:11to calculate cpk
2:21:13the first step is to determine if the
2:21:15process mean is closer to the lsl or the
2:21:19usl
2:21:20if the process mean is closer to lsl
2:21:23cpkl is determined
2:21:26cpkl is mean minus lsl divided by
2:21:30product of 3 and sigma
2:21:33if the process mean is closer to usl
2:21:36cpku is calculated
2:21:39cpku is usl minus mean divided by
2:21:43product of three and sigma
2:21:46here mean is the process average and
2:21:49sigma represents the standard deviation
2:21:52if the process mean is equidistant
2:21:54either of the specification limit can be
2:21:57chosen
2:21:58cpk takes up the value of cpku and cpkl
2:22:03depending on whichever is the lower
2:22:05value
2:22:06in the next screen we will understand
2:22:08process capability indices with an
2:22:11example
2:22:13in this screen we will discuss cpk and
2:22:16cp interpretations
2:22:18a cp value of less than 1 indicates the
2:22:21process is not capable
2:22:24even if cp is greater than one to
2:22:27ascertain if the process really is not
2:22:29capable check the cpk value
2:22:33a cpk value of less than one indicates
2:22:36that the process is definitely not
2:22:38capable but might be if cp is greater
2:22:41than one and the process mean is at or
2:22:44near the midpoint of the tolerance range
2:22:47the cpk value will always be less than
2:22:50cp
2:22:51especially as long as the process mean
2:22:54is not at the center of the process
2:22:56tolerance range
2:22:58non-centering can happen when the
2:22:59process has not understood the customer
2:23:02expectations clearly or the process is
2:23:05complete as soon as the output reaches a
2:23:08specific limit
2:23:09for example a shirt size of 40 has a
2:23:13target chest diameter of 40 inches
2:23:16but the process consistently delivers
2:23:18shirts with a mean of 41 inches as the
2:23:21chest diameter
2:23:23a machine stops removing material as
2:23:25soon as the measured dimension is within
2:23:27specified limit
2:23:29let us proceed to the next topic of this
2:23:32lesson in the following screen
2:23:36in this topic we will discuss team
2:23:38dynamics and performance
2:23:40let us start with the discussion on team
2:23:43stages
2:23:45there are five typical stages in the
2:23:47team building process
2:23:49each team passes through these stages as
2:23:51they start and proceed through the
2:23:53project
2:23:54the five stages in the team building
2:23:57process are as follows
2:23:59forming storming norming performing and
2:24:02adjourning
2:24:04in the next screen we will discuss the
2:24:06first stage forming
2:24:10the first stage in the team building
2:24:12process is called the forming stage
2:24:14in this stage the team comes together
2:24:16and begins to formulate roles and
2:24:18responsibilities
2:24:20the team leader is identified and he or
2:24:22she starts directing the team and
2:24:24assigning responsibilities to other team
2:24:27members
2:24:28most team members at this stage are
2:24:31generally enthusiastic and motivated by
2:24:33desire to be accepted within the team
2:24:36the leader employs a directive style of
2:24:38management which includes delegating
2:24:41responsibility within the team providing
2:24:44a structure to the team and determining
2:24:46processes needed for the smooth
2:24:48functioning of the team
2:24:50toward the end of this phase the team
2:24:52should achieve a commitment to the
2:24:54project and an acceptance of a common
2:24:56purpose
2:24:58in the next screen we will discuss the
2:25:00second stage storming
2:25:03the second phase in the team building
2:25:05process is called the storming stage
2:25:08as suggested by the name itself in this
2:25:11stage conflicts start to arise within
2:25:14the team
2:25:15team members often struggle over
2:25:17responsibilities and control within the
2:25:20project
2:25:21it is the responsibility of the team
2:25:23leader to coach and conciliate the team
2:25:26the leader employs a coaching style of
2:25:28management which is reflected through
2:25:30facilitating change managing conflict
2:25:33and mediating understanding between
2:25:36different parties
2:25:37towards the end of this phase team
2:25:40members need to learn to voice
2:25:41disagreement openly and constructively
2:25:44while staying focused on common
2:25:46objectives and areas of agreement
2:25:49in the next screen we will discuss the
2:25:51third stage norming
2:25:54the third stage in the team building
2:25:56process is called the norming stage
2:26:00in this stage people get along and the
2:26:02team develops a unified commitment
2:26:03toward the project goal
2:26:05the team leader promotes the team and
2:26:07participates in the team activities
2:26:10team members look to the leader to
2:26:12clarify their understanding as some
2:26:15leadership roles begin to shift within
2:26:17the lower rungs of the group
2:26:19the leader employs a participatory style
2:26:21of management through facilitating
2:26:23change working to build consensus and
2:26:26overseeing quality control
2:26:28toward the end of this phase team
2:26:30members need to accept individual
2:26:32responsibilities and workout agreements
2:26:35about team procedures
2:26:37in the next screen we will discuss the
2:26:39fourth stage performing
2:26:43the next stage in the team building
2:26:45process is called the performing stage
2:26:48this is the most productive stage for
2:26:50the project team
2:26:52in this stage team members manage
2:26:54complex tasks and work toward the common
2:26:57goals of the project
2:26:58the leader employs a supervisory style
2:27:01of management by overseeing progress
2:27:04rewarding achievement and supervising
2:27:06process
2:27:07the team leader leads the project on
2:27:09more or less an automated mode
2:27:12when the project has completed
2:27:14successfully or when the end is in sight
2:27:17the team moves into the final stage
2:27:20in the next screen we will discuss the
2:27:23fifth stage adjourning
2:27:27the last stage of team building is
2:27:29called the adjourning stage
2:27:31in this stage the project is winding
2:27:33down and the goals are within reach
2:27:36the team members are dealing with their
2:27:38impending separation from the team
2:27:41the team leader provides feedback to the
2:27:43team the leader employs a supportive
2:27:45style of management by giving feedback
2:27:48celebrating accomplishments and
2:27:50providing closure
2:27:52the team leader needs to adopt a
2:27:54different style of leadership at every
2:27:56stage
2:27:57it is therefore important for a leader
2:27:59to understand these stages and identify
2:28:02the current stage that a team is in
2:28:05the success of the team depends on how
2:28:08well the leader can guide them through
2:28:10these phases in the next screen we will
2:28:13learn about negative dynamics
2:28:17team members can exhibit negative
2:28:19behavior in more than one way during the
2:28:21project life cycle
2:28:23this behavior has a negative effect on
2:28:26the dynamics of the team
2:28:28the first kind of negative participants
2:28:31fall in the category of overbearing
2:28:33participants
2:28:34these participants use their influence
2:28:37or expertise to take on a position of
2:28:40authority
2:28:41discounting contributions from other
2:28:43team members
2:28:44to cope with such participants team
2:28:47leaders must establish ground rules for
2:28:49participation and reinforce that the
2:28:52group has the right to explore any area
2:28:54pertinent to team goals and objectives
2:28:58another kind of negative participant is
2:29:01often referred to as the dominant
2:29:03participant
2:29:05these participants take up an excessive
2:29:07amount of group time by talking too much
2:29:10focusing on trivial concerns and
2:29:12otherwise preventing participation by
2:29:15others
2:29:16team leaders need to be able to control
2:29:18dominant participants without inhibiting
2:29:21their energy or enthusiasm
2:29:24some other participants are reluctant
2:29:26participants
2:29:28who feel intimidated and are not happy
2:29:30with the team process
2:29:32owing to their reluctance they miss
2:29:34opportunities to bring up data that is
2:29:37valuable to the project
2:29:39this can often lead to hostility within
2:29:41the team
2:29:43one way to deal with reluctant
2:29:45participants
2:29:46is to respond positively and with
2:29:49encouragement to any contribution from
2:29:51the team member
2:29:53teamwork is more than a natural
2:29:55consequence of working together
2:29:57team management is more than building a
2:29:59relationship with individual team
2:30:01members
2:30:02all teams face group challenges that
2:30:05need group based diagnosis and problem
2:30:07solving to ensure that negative
2:30:09participants are able to contribute and
2:30:12perform as part of the team
2:30:15in the next screen we will learn about
2:30:17group challenges
2:30:21six sigma team and their
2:30:22responsibilities are described here
2:30:25various roles assist the smooth
2:30:27execution of a six sigma project
2:30:31these roles are required to support the
2:30:33project by providing the information and
2:30:36resources that are needed to execute the
2:30:38project
2:30:39the first important member of the six
2:30:42sigma team is the executive sponsor
2:30:45sponsors are the source or conduit for
2:30:48project resources
2:30:50and they are usually the recipients of
2:30:52the benefits the project will produce
2:30:55the sponsor is responsible for setting
2:30:57the direction and priorities for the
2:30:59organization
2:31:01the sponsor may be a functional manager
2:31:03or an external customer
2:31:06the next important role is that of the
2:31:08process owners
2:31:10they work with the black belts to
2:31:12improve their respective processes
2:31:14they provide functional expertise about
2:31:17the process to the project
2:31:20usually this role is played by the
2:31:22functional managers in charge of
2:31:24specific processes
2:31:26the next role in the project is that of
2:31:28the champions
2:31:30they are typically upper level managers
2:31:32who control and allocate resources to
2:31:35promote process improvements
2:31:37they ensure the organization is
2:31:39providing necessary resources to the
2:31:41project
2:31:42and the project is fitting into the
2:31:44strategic plans of the organization
2:31:48the first role related to the execution
2:31:50of the project is the role of the master
2:31:52black belt
2:31:54this role acts as a consultant to team
2:31:56leaders and offers expertise in the use
2:31:59of six sigma tools and methodologies
2:32:02master black belts are experts in six
2:32:05sigma statistical tools and are
2:32:08qualified to teach high level six sigma
2:32:10methodologies and applications
2:32:13each master black belt will have
2:32:15multiple black belts under him
2:32:18black belts are the leaders of
2:32:19individual six sigma projects
2:32:22they lead project teams and conduct the
2:32:25detailed analysis required in six sigma
2:32:28methodologies
2:32:29black belts act as instructors and
2:32:32mentors for green belts
2:32:34and educate them in six sigma tools and
2:32:37methods they also protect the interests
2:32:40of the project by coordinating with
2:32:42functional managers
2:32:44green belts are trained in six sigma but
2:32:47typically lead project teams working in
2:32:49their own areas of expertise
2:32:52they are focused on the basic six sigma
2:32:54tools for acceleration of projects
2:32:57green belts work on projects on a
2:32:59part-time basis dividing time between
2:33:01project and functional responsibilities
2:33:05an executive is the person who manages
2:33:08and leads the team to ensure smooth
2:33:10working of tasks and has the power to
2:33:13execute decisions
2:33:15a coach takes on a number of roles
2:33:18he or she is the person who trains
2:33:20mentors teaches and guides the team when
2:33:23required
2:33:24coach also motivates and builds
2:33:26confidence of the members
2:33:28a facilitator is a guide for the team or
2:33:31group also known as a discussion leader
2:33:35facilitators help the group or team to
2:33:37understand their common objective and
2:33:40plan their activities
2:33:42a sponsor is a person who supports the
2:33:44event or the project by providing all
2:33:47the required resources
2:33:49a team member is an individual who
2:33:51belongs to a particular project team
2:33:54a team member contributes to the
2:33:56performance of the team and actively
2:33:58participates for fulfillment of the
2:34:00project objectives
2:34:03the progress achievements and the
2:34:05details of the project have to be
2:34:07effectively communicated to the team
2:34:09management customers and stakeholders
2:34:13we will learn about modes of
2:34:14communication in the next screen
2:34:18let us understand communication within
2:34:20the team in this screen
2:34:23the purpose of communication within the
2:34:25team and the modes of communication used
2:34:28are as follows
2:34:30meetings and emails are suitable to
2:34:32communicate the roles and
2:34:33responsibilities of the team members
2:34:37meetings memos and emails are used by
2:34:39the team to understand the project
2:34:41status
2:34:43workshops and meetings are conducted to
2:34:45identify the outstanding tasks risks and
2:34:48their corrective actions
2:34:51team meetings assist decision making
2:34:54meetings and emails ensure coordination
2:34:56and efficient work the next screen will
2:34:59focus on communication with stakeholders
2:35:04the purpose of communication with
2:35:06stakeholders and the modes of
2:35:08communication used are as follows
2:35:11meeting emails and events are suitable
2:35:14to convey project objectives and goals
2:35:17to stakeholders
2:35:19meetings emails and newsletters assist
2:35:22stakeholders in understanding project
2:35:24status workshops meetings and events
2:35:28help stakeholders to identify the
2:35:30adverse effects of a situation
2:35:33meetings with stakeholders assist
2:35:35decision making process
2:35:38in the next screen we will discuss the
2:35:40communication technique
2:35:42communication techniques can be grouped
2:35:45in various ways
2:35:47the first grouping of communication
2:35:49techniques is based on the direction in
2:35:51which communication flows vertical
2:35:54communication consists of two subtypes
2:35:57namely downward flow of communication
2:35:59and upward flow of communication
2:36:03in the downward flow of communication
2:36:06the managers must pass information and
2:36:08give orders and directives to the lower
2:36:10levels in the organization
2:36:13on the contrary
2:36:14upward communication consists of
2:36:16information relayed from the bottom or
2:36:19grassroot levels to the higher levels of
2:36:22the company
2:36:23horizontal communication refers to the
2:36:26sharing of information across the same
2:36:28levels of the organization
2:36:31this can be in the form of formal and
2:36:33informal communication
2:36:36formal communications are official
2:36:38company sanctioned methods of
2:36:39communicating to the employees
2:36:42the grapevine rumor mill etc are some of
2:36:45the means of informal communication in
2:36:48the organization
2:36:50the second grouping of communication
2:36:52techniques is based on the usage of
2:36:54words
2:36:56verbal communication includes use of
2:36:58words for communication
2:37:00via telephone face to face etc
2:37:04non-verbal communication conveys
2:37:06messages without the use of words
2:37:08through body language facial expressions
2:37:11etc
2:37:13the last grouping of communication
2:37:15techniques
2:37:16is based on participation of the people
2:37:18involved in communication
2:37:21one way communication happens when
2:37:23information is relayed from the sender
2:37:25to the receiver without the expectation
2:37:28of a response like memos and
2:37:31announcements
2:37:32two-way communication is a method in
2:37:34which both parties are involved in the
2:37:36exchange of information
2:37:39team tools are a part of the team
2:37:42dynamics and performance
2:37:44the various team tools that are used are
2:37:46brainstorming
2:37:48nominal group technique and multi-voting
2:37:51if getting your learning started is half
2:37:53the battle what if you could do that for
2:37:55free
2:37:56visit skillup by simply learn click on
2:37:58the link in the description to know more
2:38:01this lesson will cover the details of
2:38:03the measure phase
2:38:04the key objective of the measure phase
2:38:07is to gather as much information as
2:38:09possible on the current processes
2:38:12this involves three key tasks that is
2:38:15creating a detailed process map
2:38:17gathering baseline data
2:38:20and summarizing and analyzing the data
2:38:23let us understand process modeling in
2:38:25the following screen
2:38:28process modeling refers to the
2:38:29visualization of a proposed system
2:38:32layout or other change in the process
2:38:35process modeling and simulation can
2:38:38determine the effectiveness or
2:38:39ineffectiveness of a new design or
2:38:42process
2:38:43they can be done using process mapping
2:38:46and flow charts
2:38:47we will learn about these in the
2:38:49forthcoming screens
2:38:51let us understand process mapping in
2:38:54this screen
2:38:55process mapping refers to a workflow
2:38:58diagram which gives a clear
2:39:00understanding of the process or a series
2:39:02of parallel processes
2:39:04it is also known as process charting or
2:39:07flow charting
2:39:09process mapping can be done either in
2:39:11the measure phase or the analyze phase
2:39:15the features of process mapping are as
2:39:17follows
2:39:19process mapping is usually the first
2:39:21step in process improvement
2:39:24process mapping gives a wider
2:39:26perspective of the problems and
2:39:28opportunities for process improvement
2:39:32it is a systematic way of recording all
2:39:34the activities performed
2:39:37process mapping can be done by using any
2:39:40of the methods like flow charts written
2:39:42procedures and detailed work
2:39:45instructions
2:39:46let us learn about flowchart in the
2:39:49following screen
2:39:51a flowchart is a pictorial
2:39:53representation of all the steps of a
2:39:55process in consecutive order
2:39:58it is used to plan a project
2:40:00document processes and communicate the
2:40:03process methodology with others
2:40:06there are many symbols used in a
2:40:08flowchart and the common symbols are
2:40:11shown in the given table
2:40:13it is recommended you take a look at the
2:40:15symbols and their description for better
2:40:17understanding
2:40:19click the button to view an example of a
2:40:21flowchart
2:40:24the given flowchart shows the processes
2:40:27involved in software development
2:40:29the flowchart starts with the start box
2:40:32which connects to the design box
2:40:34in a software project a software design
2:40:37is followed by coding which is then
2:40:39followed by testing
2:40:41in the next step there is a check for
2:40:43errors
2:40:44in case of errors it is evaluated for
2:40:47the error type
2:40:48if it is a design error it goes back to
2:40:51the beginning of the design stage
2:40:53if it is not a design error it is then
2:40:56routed to the beginning of the coding
2:40:58stage
2:40:59on the contrary if there are no errors
2:41:02the flowchart ends
2:41:06let us learn about written procedures in
2:41:08this screen
2:41:10a written procedure is a step-by-step
2:41:12guide to direct the reader through a
2:41:14task
2:41:15it is used when the process of a routine
2:41:18task is lengthy and complex
2:41:20and it is essential for everyone to
2:41:22strictly follow the rules
2:41:24written procedures can also be used when
2:41:27you want to know what is going on during
2:41:29product or process development phases
2:41:32there are a number of benefits of
2:41:34documenting procedures
2:41:36writing procedures help you avoid
2:41:38mistakes and ensure consistency
2:41:41they streamline the process and help
2:41:44your employees take relevant decisions
2:41:46and save a lot of time
2:41:49written procedures help in improving
2:41:51quality
2:41:52they are simple to understand
2:41:54as they tend to describe the processes
2:41:57at a general level
2:41:59in the next screen we will discuss how
2:42:01work instructions are helpful in
2:42:03understanding the process in detail
2:42:07work instructions define how one or more
2:42:09activities involved in a procedure
2:42:12should be written in a detailed manner
2:42:14with the aid of technology or other
2:42:16resources like flowcharts
2:42:18they provide step-by-step details for a
2:42:20sequence of activities organized in a
2:42:23logical format so that an employee can
2:42:26follow it easily and independently
2:42:29for example in the internal audit
2:42:31procedure how to fill out the audit
2:42:33results report comes under work
2:42:35instructions
2:42:37selection of the three process mapping
2:42:39tools is based on the amount of detail
2:42:42involved in the process
2:42:44for a less detailed process you can
2:42:46select flowchart and for a detailed
2:42:48process with lots of instructions you
2:42:51can select work instructions
2:42:54click the button to view an example of
2:42:56work instructions
2:42:59this example shows the work instructions
2:43:02for shipping electronic instruments
2:43:05the company name is nutri worldwide inc
2:43:07the instructions are written by brianna
2:43:09scott and approved by andrew murphy it
2:43:12is a one-page instruction
2:43:15the work instructions are documented for
2:43:17the shipping of electronic instruments
2:43:19by the shipping department
2:43:21the scope of the project states that it
2:43:23is applicable to shipping operations
2:43:26the procedure is divided into three
2:43:28broad steps
2:43:30as a first step the order for the
2:43:32shipment must be prepared
2:43:34in this step the shipping person
2:43:36receives an order number from the sales
2:43:38department through an automatic order
2:43:40system
2:43:41the quantity of the instrument and its
2:43:43card number are looked up from the
2:43:45system file and the packaging is done as
2:43:48per the instructions on the card
2:43:51the next step is packaging
2:43:53special packing instructions must be
2:43:56checked
2:43:57the instruments are then marked as per
2:43:59the instructions on the card and packed
2:44:01in a special or standard container as
2:44:03per the requirement
2:44:05the order number is written in the
2:44:07shipping system and the packing list and
2:44:09shipping documentation are obtained
2:44:12finally the quantity of instruments and
2:44:15the documents is checked
2:44:18let us understand process input and
2:44:20output variables in this screen
2:44:24any improvement of a process has a few
2:44:26prerequisites
2:44:28to improve a process the key process
2:44:31output variables kpov
2:44:34and key process input variables kpiv
2:44:38should first be measured
2:44:40metrics for key process variables
2:44:42include percent defective operation cost
2:44:46elapsed time
2:44:48backlog quantity and documentation
2:44:51errors
2:44:52critical variables are best identified
2:44:55by the process owners
2:44:57process owners know and understand each
2:45:00step of a process and are in a better
2:45:02position to identify the critical
2:45:04variables initially
2:45:06once identified the relationship between
2:45:09the variables is depicted using tools
2:45:11such as cypoc and cause and effect
2:45:14matrix
2:45:15the process input variables results are
2:45:18compared to determine which input
2:45:20variables have the greatest effect on
2:45:22the output variables
2:45:24let us proceed to the next topic of this
2:45:27lesson in the following screen
2:45:30in this topic we will discuss
2:45:32probability and statistics in detail
2:45:35let us learn about probability in the
2:45:37following screen
2:45:38probability refers to the chance of
2:45:41something occurring or happening
2:45:43an outcome is the result of a single
2:45:45trial of an experiment
2:45:48suppose there are n possible outcomes
2:45:50that are equally likely
2:45:52the probability that a specific type of
2:45:54event or outcome say f can occur is the
2:45:58number of specific outcomes divided by
2:46:01the total possible outcomes
2:46:03click the button to view an example of
2:46:06probability
2:46:09in the event of tossing a coin what is
2:46:11the probability of the occurrence of
2:46:14heads
2:46:15a single trial of tossing a coin has two
2:46:18outcomes heads and tails
2:46:21hence the probability of heads occurring
2:46:23is one divided by two the total number
2:46:26of outcomes
2:46:28let us look at some basic properties of
2:46:30probability in this screen
2:46:33there are three basic properties of
2:46:35probability
2:46:36click each property to know more
2:46:40property 1 states that the probability
2:46:42of an event is always between 0 and 1
2:46:45both inclusive
2:46:48according to property 2 the probability
2:46:51of an event that cannot occur is zero
2:46:54in other words an event that cannot
2:46:56occur is called an impossible event
2:47:01property 3 states that the probability
2:47:03of an event that must occur is one
2:47:06in other words an event that must occur
2:47:09is called a certain event
2:47:11if e is an event then the probability of
2:47:15its occurrence is given by p of e
2:47:18it is also read as the probability of
2:47:21event e
2:47:24in this screen let us look at some
2:47:26common terms used in probability along
2:47:28with an example
2:47:30the commonly used terms in probability
2:47:32are sample space
2:47:34venn diagram and event
2:47:37sample space is the collection of all
2:47:39possible outcomes for a given experiment
2:47:42in the coin example discussed earlier
2:47:44the sample space consists of one
2:47:46instance each of heads and tails
2:47:48if two coins are tossed the sample space
2:47:51would be four in total a venn diagram
2:47:54shows all hypothetically possible
2:47:56logical relations between a finite
2:47:58collection of sets
2:48:00an event is a collection of outcomes for
2:48:02an experiment which is any subset of the
2:48:05sample space
2:48:07click the button to view an example of
2:48:09probability
2:48:12what is the probability of getting a 3
2:48:14followed by two when a dice is thrown
2:48:17twice
2:48:19when the dice is thrown twice the first
2:48:21throw can have any number from one to
2:48:23six
2:48:25similarly the second throw can also have
2:48:27any number from one to six so the total
2:48:30sample space is six times six
2:48:34that is thirty six
2:48:36the event in this case is three followed
2:48:38by two
2:48:40this can happen in only one way
2:48:42so the probability in the question is
2:48:45one divided by thirty-six
2:48:49let us discuss the basic concepts of
2:48:51probability in this screen
2:48:54some basic concepts of probability are
2:48:57independent event
2:48:58dependent event
2:49:00mutually exclusive and mutually
2:49:03inclusive events
2:49:05click each concept to know more
2:49:09when the probability of occurrence of an
2:49:11event does not affect the probability of
2:49:14occurrence of another event
2:49:16the two events are said to be
2:49:18independent
2:49:19suppose you rolled a dice and flipped a
2:49:22coin at the same time
2:49:24the probability of getting any number on
2:49:26the dice in no way influences the
2:49:28probability of getting heads or tails on
2:49:31the coin
2:49:33when the probability of one event
2:49:35occurring influences the likelihood of
2:49:38the other event the events are said to
2:49:40be dependent
2:49:43events are said to be mutually exclusive
2:49:46if the occurrence of any one of them
2:49:48prevents the occurrence of all the
2:49:50others in other words only one event can
2:49:53occur at a time
2:49:55consider an example of flipping a coin
2:49:58when you flip a coin you will either get
2:50:00heads or tails but not both
2:50:03you can add the probabilities of these
2:50:05two events to prove they are mutually
2:50:08exclusive
2:50:10any two events wherein one event cannot
2:50:12occur without the other are said to be
2:50:15mutually inclusive events
2:50:19in this screen let us learn about the
2:50:21multiplication rules also known as and
2:50:24rules
2:50:26the multiplication rules or and rules
2:50:29depend on the event dependency
2:50:32for independent events that is if two
2:50:35events are independent of each other the
2:50:37special multiplication rule applies
2:50:41for mutually independent events the
2:50:43special multiplication rule is as
2:50:46follows
2:50:47if the events a b c and so on are
2:50:51independent of each other
2:50:53then the probability of a and b and c
2:50:56and so on is equal to the product of
2:50:58their individual probabilities
2:51:01click the button to view an example of
2:51:04this rule
2:51:06suppose there are three events which are
2:51:08independent of each other such as the
2:51:11event of flipping a coin and getting
2:51:13heads
2:51:14drawing a card and getting an ace and
2:51:16throwing a dice and getting a one
2:51:19what is the probability of occurrence of
2:51:21all these events
2:51:23the answer is the probability of a and b
2:51:27and c
2:51:28is equal to the product of their
2:51:30individual probabilities
2:51:32which is half multiplied by
2:51:34one-thirteenth multiplied by one-sixth
2:51:37the result is
2:51:390.0064 which is 0.64
2:51:44hence there is 0.64 probability of all
2:51:49of the events occurring
2:51:51we will continue the discussion on
2:51:53multiplication rules in this screen
2:51:56the multiplication rule for
2:51:58non-independent or conditional events
2:52:01which is also the general multiplication
2:52:03rule is as follows
2:52:06if a and b are two events then the
2:52:08probability of a and b is equal to the
2:52:11product of probability of a and the
2:52:13probability of b given a alternatively
2:52:17we can say that for any two events their
2:52:20joint probability is equal to the
2:52:22probability that one of these events
2:52:24occurs multiplied with the conditional
2:52:27probability of the other event given the
2:52:29first event
2:52:31click the button to view an example of
2:52:33this rule
2:52:35a bag contains six golden coins and four
2:52:39silver coins
2:52:40two coins are drawn without replacement
2:52:42from the bag
2:52:44what is the probability that both of the
2:52:46coins are silver
2:52:48let a be the event that the first coin
2:52:51is silver and b be the event that the
2:52:53second coin is silver there are ten
2:52:57coins in the bag
2:52:58four of which are silver
2:53:00therefore p of a equals four divided by
2:53:03ten
2:53:05after the first selection there are nine
2:53:07coins in the bag
2:53:09three of which are silver
2:53:11therefore p of b given a equals three
2:53:14divided by nine
2:53:16therefore based on the rule of
2:53:18multiplication
2:53:19probability of a intersection b
2:53:22equals four divided by ten multiplied by
2:53:26three divided by nine
2:53:28the answer is twelve divided by ninety
2:53:31which is zero point one three three four
2:53:34hence there is thirteen percent
2:53:35probability that both the coins are
2:53:38silver
2:53:41in this screen we will look at the
2:53:42definitions and formula of permutation
2:53:45and combination
2:53:47permutation is the total number of ways
2:53:50in which a set group or number of things
2:53:53can be arranged
2:53:54the order matters to a great extent in
2:53:57permutation
2:53:58the manner in which the objects or
2:54:00numbers are arranged will be considered
2:54:02in permutation
2:54:04the formula for permutation is
2:54:07npr equals p of n and r equals n
2:54:10factorial divided by n minus r factorial
2:54:14where n is the number of objects and r
2:54:17is the number of objects taken at a time
2:54:20the unordered arrangement of set group
2:54:22or number of things is known as
2:54:24combination
2:54:26the order does not matter in combination
2:54:29the formula for combination is ncr
2:54:32equals c of n and r equals n factorial
2:54:35divided by r factorial multiplied by n
2:54:38minus r factorial where n is the number
2:54:42of objects and r is the number of
2:54:44objects taken at a time
2:54:47let us look at an example for
2:54:48calculating permutation and combination
2:54:51in the following screen
2:54:54from a group of ten employees a company
2:54:56has to select four for a particular
2:54:58project
2:55:00in how many ways can the selection
2:55:02happen given the following conditions
2:55:05when the arrangements of employees needs
2:55:07to be different
2:55:08when the arrangement of employees need
2:55:10not be different
2:55:12click the button to know the answer
2:55:15in the given example the values of n and
2:55:18r are ten and four respectively
2:55:21let us consider the first condition
2:55:24from a group of ten employees four
2:55:27employees need to be selected
2:55:29the arrangement needs to be different
2:55:32using the permutation formula
2:55:34n p r equals p of n and r equals n
2:55:38factorial divided by n minus r factorial
2:55:4210 p four equals p of ten and four
2:55:45equals ten factorial divided by ten
2:55:48minus four factorial equals five
2:55:50thousand forty
2:55:52therefore the four employees can be
2:55:54selected in five thousand forty ways
2:55:58let us now consider the second condition
2:56:01from a group of ten employees four
2:56:04employees need to be selected
2:56:06the arrangement of employees need not be
2:56:09different
2:56:10using the combination formula
2:56:12n c r equals c of n and r equals n
2:56:16factorial divided by r factorial
2:56:19multiplied by n minus r factorial
2:56:2310 c 4 equals c of ten and four equals
2:56:26ten factorial divided by four factorial
2:56:29multiplied by ten minus four factorial
2:56:32equals two hundred and ten
2:56:35therefore the four employees can be
2:56:37selected from a group of 10 employees in
2:56:40210 different ways
2:56:43let us understand the two types of
2:56:46statistics in this screen
2:56:48statistics refers to the science of
2:56:50collection analysis interpretation and
2:56:54presentation of data
2:56:56in six sigma statistical methods and
2:56:58principles are used to measure and
2:57:00analyze the process performance and
2:57:02improvements
2:57:04there are two major types of statistics
2:57:07descriptive statistics and inferential
2:57:09statistics
2:57:11descriptive statistics is also known as
2:57:14enumerative statistics and inferential
2:57:17statistics is also known as analytical
2:57:20statistics
2:57:21descriptive statistics include
2:57:23organizing summarizing and presenting
2:57:25the data in a meaningful way
2:57:28whereas inferential statistics includes
2:57:31making inferences and drawing
2:57:33conclusions from the data
2:57:35descriptive statistics describes what's
2:57:38going on in the data the main objective
2:57:40of inferential statistics is to make
2:57:43inferences from the data to more general
2:57:45conditions
2:57:46histograms pie charts box plots
2:57:50frequency distributions and measures of
2:57:52central tendency mean median and mode
2:57:56are all examples of descriptive
2:57:58statistics
2:57:59on the other hand examples of
2:58:01inferential statistics are hypothesis
2:58:03testing scatter diagrams etc
2:58:08the main objective of statistical
2:58:10inference is to draw conclusions on
2:58:13population characteristics based on the
2:58:15information available in the sample
2:58:18collecting data from a population is not
2:58:20always easy especially if the size of
2:58:23the population is big
2:58:25the easier way is to collect a sample
2:58:28from the population and from the sample
2:58:30statistic collected make an assessment
2:58:33about the population parameter
2:58:35click the button to see an example of
2:58:38statistical inference
2:58:41the management team of a cricket council
2:58:43wants to know if the team's performance
2:58:45has improved after recruiting a new
2:58:47coach
2:58:48the management conducts a test to prove
2:58:50this statistically
2:58:52let us consider y a and y b
2:58:55where y a stands for efficiency of coach
2:58:59a
2:58:59and yb stands for efficiency of coach b
2:59:03to conduct the test the basic assumption
2:59:06is coach a and coach b are both
2:59:09effective
2:59:10this basic assumption is known as null
2:59:12hypothesis
2:59:14here let us assume the status quo is
2:59:17null hypothesis hence null hypothesis h
2:59:20o can be given by y a equals yb
2:59:24the management team also challenges
2:59:26their basic assumption by assuming the
2:59:29coaches are not equally effective
2:59:32this is their alternate hypothesis
2:59:35the alternate hypothesis states that the
2:59:37efficiencies of the two coaches differ
2:59:40if the null hypothesis is proven wrong
2:59:42the alternate hypothesis must be right
2:59:46hence alternate hypothesis h1 can be
2:59:49given y a is not equal to yb
2:59:53these hypothesis statements are used in
2:59:56a hypothesis test which will be
2:59:58discussed in the later part of the
3:00:00course
3:00:02in this screen we will learn about the
3:00:04types of errors
3:00:06when collecting data from a population
3:00:08as a sample
3:00:09and forming a conclusion on the
3:00:11population based on the sample you run
3:00:13into the risk of committing errors
3:00:16there are two possible errors that can
3:00:18happen
3:00:19type 1 error and type 2 error
3:00:22the type 1 error occurs when the null
3:00:24hypothesis is rejected when it is in
3:00:27fact true
3:00:28type 1 error is also known as producer's
3:00:31risk
3:00:32the chance of committing a type 1 error
3:00:35is known as alpha
3:00:37alpha or significance level is the
3:00:40chance of committing a type one error
3:00:42and is typically chosen to be five
3:00:44percent
3:00:46this means the maximum amount of risk
3:00:48you have for committing a type one error
3:00:50is five percent
3:00:52let us consider the previous example
3:00:55arriving at a conclusion that coach b is
3:00:58better than coach a when in fact they
3:01:00are at the same level is a type one
3:01:03error
3:01:04the risk you have of committing this
3:01:06error is 5
3:01:08which means there is a 5 chance your
3:01:10experiment can give wrong results the
3:01:13type 2 error occurs when the null
3:01:15hypothesis is accepted when it is in
3:01:18fact false also when you reject the
3:01:22alternate hypothesis when it is actually
3:01:24true
3:01:25you commit a type 2 error
3:01:28type 2 error is also referred to as
3:01:30consumer's risk
3:01:32in comparing the two coaches the coaches
3:01:34were actually different in their
3:01:36efficiencies but the conclusion was that
3:01:38they are the same
3:01:40the chance of committing a type 2 error
3:01:42is known as beta the maximum chance of
3:01:46committing a type 2 error is 20
3:01:49in the next screen we will learn about
3:01:51central limit theorem
3:01:55central limit theorem clt states that
3:01:57for a sample size greater than 30 the
3:02:00sample mean is very close to the
3:02:02population mean
3:02:03in simple words the sample mean
3:02:05approaches the normal distribution
3:02:07for example if you have sample 1 and its
3:02:10mean is mean 1 sample 2 and its mean is
3:02:14mean 2 and so on
3:02:16take the means of mean one mean two etc
3:02:19and you will find that it is the same as
3:02:21the population mean
3:02:23population mean is the average of the
3:02:25sample means
3:02:26in such cases the standard error of mean
3:02:30also known as sem that represents the
3:02:32variability between the sample means is
3:02:35very less
3:02:37the sem is often used to represent the
3:02:39standard deviation of the sample
3:02:42the formula for sem is population
3:02:45standard deviation divided by the square
3:02:47root of the sample size
3:02:50selecting a sample size also depends on
3:02:52the concept called power also known as
3:02:55power of the test
3:02:57we will cover this concept in detail in
3:03:00the later part of the course
3:03:02let us look at the graphical
3:03:03representation of the central limit
3:03:06theorem in the following screen
3:03:09the plot of the three numbers 2 three
3:03:11and four looks as shown in the graph
3:03:14it is interesting to note that the total
3:03:16number of times each digit is chosen is
3:03:18six
3:03:19when the plot of the sample mean of nine
3:03:22samples of size two each is drawn
3:03:24it looks like the red line which is
3:03:26plotted in the figure
3:03:28the x-axis shows numbers of the mean
3:03:30which are two two point five three and
3:03:33four
3:03:34on the y-axis the frequency is plotted
3:03:37the point at which arrows from number
3:03:39two and three converge is the mean of
3:03:41two and three
3:03:43similarly the point at which arrows from
3:03:45two and four converge is the mean of the
3:03:47numbers two and four
3:03:49let us discuss the concluding points of
3:03:51the central limit theorem in the next
3:03:53screen
3:03:55the central limit theorem concludes that
3:03:58the sampling distributions are helpful
3:04:00in dealing with non-normal data
3:04:02if you take the sample data points from
3:04:04a population and plot the distribution
3:04:07of the means of the sample you get the
3:04:09sampling distribution of the means
3:04:12the mean of the sampling distribution
3:04:14also known as the mean of means will be
3:04:16equal to the population mean
3:04:19also the sampling distribution
3:04:21approaches normality as the sample size
3:04:23increases
3:04:25note that clt enables you to draw
3:04:28inferences from the sample statistics
3:04:30about the population parameters
3:04:32this is irrespective of the distribution
3:04:35of the population
3:04:37clt also becomes the basis for
3:04:39calculating confidence interval for
3:04:41hypothesis tests as it allows the use of
3:04:44a standard normal table
3:04:47let us proceed to the next topic of this
3:04:49lesson in the following screen
3:04:52in this topic we will cover the concept
3:04:54of statistical distributions
3:04:56let us start with discrete probability
3:04:58distribution in the following screen
3:05:01discrete probability distribution is
3:05:04characterized by the probability mass
3:05:06function
3:05:07it is important to be familiar with
3:05:09discrete distributions while dealing
3:05:11with discrete data
3:05:13some of the examples of discrete
3:05:15probability distribution are binomial
3:05:18distribution poisson distribution
3:05:20negative binomial distribution geometric
3:05:24distribution and hyper geometric
3:05:26distribution
3:05:28we will focus only on the two most
3:05:30useful discrete distributions
3:05:33binomial distribution and poisson
3:05:35distribution
3:05:37like most probability distributions
3:05:39these distributions also help in
3:05:42predicting the sample behavior that has
3:05:44been observed in a population
3:05:47let us learn about binomial distribution
3:05:49in the following screen
3:05:52binomial distribution is a probability
3:05:55distribution for discrete data
3:05:58named after the swiss mathematician
3:06:00jacob bernoulli it is an application of
3:06:03popular knowledge to predict the sample
3:06:05behavior
3:06:06binomial distribution also describes the
3:06:09discrete data as a result of a
3:06:12particular process like the tossing of a
3:06:14coin for a fixed number of times and the
3:06:17success or failure in an interview
3:06:20a process is known as bernoulli's
3:06:22process when the process output has only
3:06:25two possible values like defective or
3:06:28okay
3:06:29pass or fail and yes or no
3:06:33binomial distribution is used to deal
3:06:35with defective items defect is any
3:06:38non-compliance with a specification
3:06:41defective is a product or service with
3:06:43at least one defect
3:06:45binomial distribution is most suitable
3:06:48when the sample size is less than 30 and
3:06:50less than 10 percent of the population
3:06:53it is the percentage of non-defective
3:06:55items provided the probability of
3:06:58creating a defective item remains the
3:07:00same over a period
3:07:02let us look at the equation
3:07:04the probability of exactly r successes
3:07:08out of a sample size of n
3:07:10is denoted by p of r
3:07:12which is equal to ncr whole multiplied
3:07:16by p to the power of r and one minus p
3:07:19whole to the power of n minus r
3:07:22in the equation b is the probability of
3:07:25success
3:07:26r is the number of successes desired and
3:07:29n is the sample size
3:07:32to continue discussing the binomial
3:07:34distribution
3:07:35let us look at some of its key
3:07:37calculations in the following screen
3:07:41the mean of a binomial distribution is
3:07:44denoted by mio and is given by n
3:07:47multiplied by p
3:07:49the standard deviation of a binomial
3:07:51distribution
3:07:52is denoted by sigma which is equal to n
3:07:55multiplied by p multiplied by 1 minus p
3:08:00the method of calculating factorials say
3:08:02factorial of 5 is the product of five
3:08:06four three two and one
3:08:08which is equal to one hundred twenty
3:08:11similarly factorial of four is the
3:08:14product of four three two and one
3:08:17which is equal to twenty four let us
3:08:20look at an example of calculating
3:08:22binomial distribution in the next screen
3:08:26suppose you wish to know the probability
3:08:29of getting heads five times in eight
3:08:31coin tosses
3:08:33you can use the binomial equation for
3:08:35the same
3:08:36click the answer button to see how this
3:08:38is done
3:08:40the tossing of a coin has only two
3:08:42outcomes heads and tails
3:08:45it means that the probability of each
3:08:47outcome is 0.5 and it remains fixed over
3:08:50a period of time
3:08:52additionally the outcomes are
3:08:53statistically independent
3:08:56in this case the probability of success
3:08:58denoted by p is 0.5 the number of
3:09:02successes desired is denoted by r which
3:09:04is five and the sample size is denoted
3:09:07by n which is eight
3:09:09therefore the probability of five heads
3:09:12is equal to factorial of eight c r which
3:09:15is eight divided by factorial of five
3:09:18and factorial of eight minus five whole
3:09:21multiplied by zero point five to the
3:09:23power of five multiplied by one minus
3:09:26zero point five whole to the power of
3:09:28eight minus five this calculation gives
3:09:31a result of zero point two one eight
3:09:34seven
3:09:35which is equal to twenty one point eight
3:09:37seven percent
3:09:39let us learn about poisson distribution
3:09:41in this screen
3:09:42poisson distribution is named after
3:09:44simeone denis on and is also used for
3:09:47discrete data
3:09:49poisson distribution is an application
3:09:52of the population knowledge to predict
3:09:54the sample behavior
3:09:55it is generally used for describing the
3:09:58probability distribution of an event
3:10:00with respect to time or space
3:10:03some of the characteristics of poisson
3:10:05distribution are as follows
3:10:07poisson distribution describes the
3:10:09discrete data resulting from a process
3:10:11like the number of calls received by a
3:10:13call center agent or the number of
3:10:15accidents at a signal
3:10:18unlike binomial distribution which deals
3:10:20with binary discrete data possum
3:10:22distribution deals with integers which
3:10:24can take any value
3:10:26poisson distribution is suitable for
3:10:28analyzing situations wherein the number
3:10:31of trials similar to the sample size in
3:10:33binomial distribution is large and tends
3:10:37towards infinity
3:10:38additionally it is used in situations
3:10:41where the probability of success in each
3:10:43trial is very small almost tending
3:10:46towards zero
3:10:47this is the reason why poisson
3:10:49distribution is applicable for
3:10:51predicting the occurrence of rare events
3:10:53like plane crashes car accidents etc and
3:10:57is therefore widely used in the
3:10:58insurance sector poisson distribution
3:11:01can be used for predicting the number of
3:11:03defects as well given a low defect
3:11:05occurrence rate
3:11:07let us look at the formula for
3:11:09calculating poisson distribution in the
3:11:11next screen
3:11:13the poisson distribution for a
3:11:15probability of exactly x occurrences is
3:11:18given by p of x equals to lambda to the
3:11:21power of x multiplied with log e to the
3:11:23power of minus lambda whole divided by
3:11:26factorial of x
3:11:28in this equation lambda is the mean
3:11:31number of occurrences during the
3:11:32interval
3:11:33x is the number of occurrences desired
3:11:36and e is the base of natural logarithm
3:11:39which is equal to 2.71828
3:11:42the mean of the poisson distribution is
3:11:45given by the lambda and the standard
3:11:47deviation of a poisson distribution is
3:11:49given by sigma which is the square root
3:11:52of lambda
3:11:53let us look at an example to calculate
3:11:56poisson distribution in the next screen
3:11:59the past records of a road junction
3:12:01which is accident prone show a mean
3:12:04number of five accidents per week at
3:12:06this junction
3:12:07assume that the number of accidents
3:12:09follows up with song distribution and
3:12:12calculate the probability of any number
3:12:14of accidents happening in a week
3:12:16click the button to know the answer
3:12:20given the situation you know that the
3:12:22value of lambda or mean is five
3:12:25so p of zero that is the probability of
3:12:27zero accidents per week is calculated as
3:12:30five to the power of zero multiplied by
3:12:33e to the power of minus five whole
3:12:35divided by factorial of zero
3:12:38the answer is zero point zero zero six
3:12:42applying the same formula the
3:12:44probability of one accident per week is
3:12:460.03
3:12:48the probability of more than two
3:12:50accidents per week is one minus the sum
3:12:53of probabilities of zero one and two
3:12:56accidents
3:12:57which is 0.884
3:13:00in other words the probability is 88.4
3:13:03percent
3:13:05let us learn about normal distribution
3:13:07in this screen
3:13:09the normal or gaussian distribution is a
3:13:11continuous probability distribution
3:13:14the normal distribution is represented
3:13:16as n and depends on two factors
3:13:20me u which stands for mean and sigma
3:13:23which gives the standard deviation of
3:13:24the data points
3:13:26normal distribution normally has a
3:13:28higher frequency of values around the
3:13:31mean and lesser occurrences away from it
3:13:34it is often used as a first
3:13:36approximation to describe real valued
3:13:38random variables that tend to cluster
3:13:40around a single mean value
3:13:43the distribution is bell shaped and
3:13:45symmetrical the total area under the
3:13:47normal curve is one which is p of x
3:13:51various types of data such as body
3:13:53weight height the output of a
3:13:55manufacturing device etc follow the
3:13:57normal distribution
3:13:59additionally normal distribution is
3:14:01continuous and symmetrical with the
3:14:03tails asymptotic to the x-axis which
3:14:06means they touch the x-axis at infinity
3:14:10let us continue to discuss normal
3:14:12distribution in the following screen
3:14:15in a normal distribution to standardize
3:14:17comparisons of dispersion or the
3:14:19different measurement units like inches
3:14:21meters grams etc
3:14:23a standard z variable is used the uses
3:14:26of z value are as follows
3:14:29while the value of z or the number of
3:14:31standard deviations is unique for each
3:14:34probability within the normal
3:14:36distribution
3:14:37it helps in finding probabilities of
3:14:39data points anywhere within the
3:14:41distribution
3:14:42it is dimensionless as well that is it
3:14:45has no units such as millimeters liters
3:14:48columns etc
3:14:50there are different formulas to arrive
3:14:52at the normal distribution
3:14:53we will focus on one commonly used
3:14:56formula for calculating normal
3:14:57distribution which is
3:15:00z equals y minus mu whole divided by
3:15:03sigma
3:15:04here z is the number of standard
3:15:06deviations between y and the mean
3:15:09denoted by mu
3:15:11y is the value of the data point in
3:15:13concern
3:15:14me u is mean of the population or data
3:15:16points and sigma is the standard
3:15:19deviation of the population or data
3:15:21points
3:15:22let us look at an example for
3:15:24calculating normal distribution in the
3:15:26following screen
3:15:28suppose the time taken to resolve
3:15:31customer problems follows a normal
3:15:32distribution with a mean of 250 hours
3:15:36and standard deviation of 23 hours
3:15:39find the probability of a problem
3:15:41resolution taking more than 300 hours
3:15:44click the button to know the answer
3:15:47in this case y is equal to 300 me u
3:15:50equals 250 and sigma equals 23.
3:15:54applying the normal distribution formula
3:15:57z is equal to 300 minus 250 whole
3:16:00divided by 23
3:16:02the result is 2.17
3:16:05when you look at the normal distribution
3:16:07table the z value of 2.17 covers an area
3:16:12of
3:16:130.98499 under itself this means the
3:16:16probability of a problem taking 0 to 300
3:16:19hours to be resolved is 98.5 percent and
3:16:23therefore the chances of a problem
3:16:25resolution taking more than 300 hours is
3:16:28point five percent
3:16:31let us understand the usage of z table
3:16:33in this screen
3:16:35the graphical representation of z table
3:16:38usage is given here
3:16:40the probability of areas under the curve
3:16:42is one
3:16:43for the actual value one can identify
3:16:46the z score by using the z table
3:16:49as shown this probability is the area
3:16:52under the curve to the left of point
3:16:54plus a to zero
3:16:57using the actual data when you calculate
3:16:59mean and standard deviation and the
3:17:02values are 25 and 5 respectively it is
3:17:05the normal distribution
3:17:07if the same data is standardized to a
3:17:10mean value of zero and standard
3:17:12deviation value of one
3:17:14it is the standard normal distribution
3:17:17in the next screen we will take a look
3:17:19at the z table
3:17:22the z table gives the probability that z
3:17:25is between zero and a positive number
3:17:28there are different forms of normal
3:17:30distribution z tables followed globally
3:17:33the most common form of z table with
3:17:36positive z scores is shown here
3:17:39the value of a called the percentage
3:17:41point is given along the borders of the
3:17:44table in bold and is to two decimal
3:17:47places
3:17:48the values in the main table are the
3:17:50probabilities that z is between 0 and
3:17:53plus a
3:17:55note that the values running down the
3:17:57table are to one decimal place
3:18:00the numbers along the column change only
3:18:02for the second decimal place
3:18:04let us look at some examples and how to
3:18:07use a z table in the following screen
3:18:11let us find the value of p of z less
3:18:14than zero
3:18:15the table is not needed to find the
3:18:17answer once we know that the variable z
3:18:20takes a value less than or equal to zero
3:18:24first the area under the curve is one
3:18:26and second the curve is symmetrical
3:18:29about z equals zero
3:18:31hence there is 0.5 or 50 percent above
3:18:35chance of z equals zero and zero point
3:18:37five or fifty percent below chance of z
3:18:41equals zero
3:18:43let us find the value of p of z greater
3:18:46than one point one two
3:18:48in this case the chance of z is greater
3:18:51than a number in this case 1.12
3:18:54you can find this by using the following
3:18:57fact
3:18:58the opposite or complement of an event
3:19:00of a is the event of not a
3:19:03that is the opposite or complement of
3:19:05event a occurring is the event a not
3:19:08occurring
3:19:10its probability is given by p of not
3:19:12equal a equals one minus p of a
3:19:16in other words p of z greater than 1.12
3:19:19is 1 minus the opposite
3:19:21which is p of z lesser than 1.12
3:19:25using the table p of z less than 1.12
3:19:29equals 0.5 plus p of 0 less than z less
3:19:34than 1.12 equals 0.5 plus
3:19:380.3686
3:19:40which is
3:19:410.8686
3:19:43hence the answer is p of z greater than
3:19:461.12 equals 1 minus
3:19:490.8686 which is
3:19:520.1314
3:19:54note the answer is less than 0.5
3:19:58let us find the value of p of z lies
3:20:01between 0 and 1.12
3:20:04in this case where z falls within an
3:20:06interval the probability can be read
3:20:08straight off the table
3:20:10p of z lies between 0 and 1.12 equals
3:20:140.3686
3:20:17we will learn about chi-square
3:20:18distribution in this screen
3:20:21chi-square distribution is also known as
3:20:24chi squared or chi-square
3:20:26distribution chi squared with k minus 1
3:20:29degrees of freedom is the distribution
3:20:32of a sum of the squares of k independent
3:20:35standard normal random variables
3:20:38the chi-square distribution is one of
3:20:40the most widely used probability
3:20:42distributions in inferential statistics
3:20:46it is also known as hypothesis testing
3:20:49and the distribution is used in
3:20:50hypothesis tests
3:20:52when used in hypothesis tests it only
3:20:55needs one sample for the test to be
3:20:57conducted
3:20:58conventionally degree of freedom is k
3:21:01minus one
3:21:02where k is the sample size
3:21:05for example if w x y and z are four
3:21:08random variables with standard normal
3:21:11distributions
3:21:12then the random variable f
3:21:15which is the sum of w square x square y
3:21:18square and z square has a chi-square
3:21:21distribution
3:21:22the degrees of the freedom of the
3:21:24distribution df equals the number of
3:21:26normally distributed variables used
3:21:29in this case df is equal to four
3:21:33let us look at the formula to calculate
3:21:35chi-square distribution in the following
3:21:37screen
3:21:39chi-square calculated or sigma or the
3:21:43chi-square index equals f of o minus f
3:21:47of e whole square divided by f of e
3:21:50here f of o stands for an observed
3:21:53frequency and f of e stands for an
3:21:56expected frequency determined through a
3:21:58contingency table
3:22:00let us understand t distribution in the
3:22:03next screen
3:22:04the t distribution method is the most
3:22:07appropriate method to be used in the
3:22:09following situations
3:22:11when you have a sample size of less than
3:22:1330 when the population standard
3:22:16deviation is not known
3:22:18when the population is approximately
3:22:20normal
3:22:21unlike the normal distribution a t
3:22:24distribution is lower at the mean and
3:22:26higher at the tails as seen in the image
3:22:30t distribution is used for hypothesis
3:22:32testing also as seen in the image the t
3:22:36distribution is symmetrical in shape but
3:22:39flatter than the normal distribution
3:22:42as the sample size increases the t
3:22:45distribution approaches normality
3:22:47for every possible sample size or
3:22:50degrees of freedom there is a different
3:22:53t distribution
3:22:54let us learn about f distribution in the
3:22:57following screen
3:23:00the f distribution is a ratio of two
3:23:03chi-square distributions
3:23:05a specific f distribution is denoted by
3:23:08the ratio of the degrees of freedom for
3:23:11the numerator chi-square
3:23:13and the degrees of freedom for the
3:23:15denominator chi-square
3:23:17the f-test is performed to calculate and
3:23:20observe if the standard deviations or
3:23:23variances of two processes are
3:23:25significantly different
3:23:27the project teams are usually concerned
3:23:30about reducing the process variance
3:23:33as per the formula f calculated equals
3:23:36s1 square divided by s2 square where s1
3:23:40and s2 are the standard deviations of
3:23:43the two samples
3:23:45if the f calculated is one it implies
3:23:48there is no difference in the variance
3:23:51if s1 is greater than s2
3:23:54then the numerator must be greater than
3:23:57the denominator
3:23:58in other words df1 equals n1 minus 1 and
3:24:03df2 equals n2 minus 1.
3:24:07from the f distribution table you can
3:24:10easily find out the critical f
3:24:12distribution at alpha and the degrees of
3:24:14freedom of the samples of two different
3:24:17processes df1 and df2
3:24:20let us proceed to the next topic of this
3:24:23lesson in the following screen
3:24:26in this topic we will discuss collecting
3:24:28and summarizing data in detail
3:24:31let us learn about types of data in the
3:24:33following screen
3:24:35data is objective information which
3:24:38everyone can agree on it is a collection
3:24:41of facts from which conclusions may be
3:24:43drawn
3:24:44the two types of data are attribute data
3:24:47and variable data
3:24:49click each type to know more
3:24:52discrete data is data that can be
3:24:54counted and only includes numbers such
3:24:57as two forty or thousand fifty
3:25:00attribute data is commonly called pass
3:25:03fail or good bad data
3:25:06attribute or discrete data cannot be
3:25:09broken down into a smaller unit
3:25:11meaningfully
3:25:12it answers questions such as how many
3:25:15how often or what type
3:25:17some examples of attribute data are
3:25:20number of defective products percentage
3:25:22of defective products
3:25:25frequency at which a machine is repaired
3:25:27or the type of award received
3:25:32any data that can be measured on a
3:25:34continuous scale is continuous or
3:25:37variable data
3:25:38this type of data answers questions such
3:25:41as how long what volume or how far
3:25:45examples of continuous data include
3:25:48height weight
3:25:49time taken to complete a task
3:25:52temperature and so on
3:25:55let us understand the importance of
3:25:56selecting the data type in this screen
3:26:00deciding the data type facilitates
3:26:02analysis and interpretation
3:26:05therefore the first step in the measure
3:26:07phase is to determine what type of data
3:26:10should be collected
3:26:11this can be done by considering the
3:26:13following
3:26:15the first consideration is to identify
3:26:17what is already known
3:26:19for this the values already identified
3:26:21for the process are listed
3:26:24these include critical to quality
3:26:26parameters or ctq's
3:26:28key process output variables or
3:26:31kpovs
3:26:33and the key process input variable or
3:26:36kpivs
3:26:38next to understand how to proceed with
3:26:41the data gathered it is necessary to
3:26:43determine the data type that fits the
3:26:45metrics for the key variables identified
3:26:49the question now arises
3:26:51why should the data type be identified
3:26:54this is important as it enables the
3:26:57right set of data to be collected
3:27:00analyzed and used to draw inferences
3:27:03it is not advisable to convert one type
3:27:06of data into another
3:27:08converting attribute data to variable
3:27:10data is difficult and requires
3:27:13assumptions to be made about the process
3:27:16it may also require additional data
3:27:18gathering including retesting units
3:27:21let us look at measurement scales in the
3:27:23following screen
3:27:26there are four measurement scales
3:27:28arranged in the table in increasing
3:27:30order of their statistical desirability
3:27:33in the nominal scale the data consists
3:27:36of only names or categories and there is
3:27:38no possibility of ordering
3:27:41an example of this type of measurement
3:27:43can be a bag of colored balls which
3:27:46contains ten green balls five black
3:27:49balls eight yellow balls and nine white
3:27:52balls
3:27:53this is the least informative of all
3:27:55scales
3:27:56the most appropriate measure of central
3:27:58tendency for this scale is mode
3:28:01in the ordinal or ranking scale data is
3:28:04arranged in order and values can be
3:28:06compared with each other
3:28:08an example of this scale can be the
3:28:10ratings given to different restaurants
3:28:13three for a five for b
3:28:15two for c and four for d
3:28:18the central tendency for this scale is
3:28:21median or mode the interval scale is
3:28:24used for ranking items in step order
3:28:26along a scale of equidistant points
3:28:29for example the temperatures of three
3:28:32metal rods are 100 degrees 200 degrees
3:28:35and 600 degrees fahrenheit respectively
3:28:39note that three times 200 degrees is not
3:28:42the same as 600 degrees as a temperature
3:28:44measurement
3:28:46the central tendency here is mean median
3:28:49or mode
3:28:50mean is used if the data does not have
3:28:52any outliers
3:28:54the ratio scale represents variable data
3:28:57and is measured against a known standard
3:28:59or increment
3:29:01however this scale also has an absolute
3:29:04zero that is no numbers exist below zero
3:29:08an example of the ratio scale are
3:29:10physical measures where height weight
3:29:12and electric charge represent ratio
3:29:15scale data
3:29:16note that negative length is not
3:29:18possible
3:29:19again here you would use mean median or
3:29:22mode as the central tendency measure
3:29:26in the next screen we will learn about
3:29:28assuring data accuracy
3:29:31to ensure data is accurate sampling
3:29:34techniques are used
3:29:36sampling is the process act or technique
3:29:39of selecting an appropriate test group
3:29:41or sample from a larger population
3:29:44it is preferable to survey 100 people to
3:29:47surveying 10 000 people
3:29:50sampling saves the time money and effort
3:29:52involved in collecting data
3:29:54the three types of sampling techniques
3:29:56described here are random sampling
3:30:00sequential sampling and stratified
3:30:02sampling
3:30:03click each type to know more
3:30:07random sampling is the technique where a
3:30:09group of subjects or a sample for study
3:30:12is selected from a larger group or
3:30:14population at random
3:30:18sequential sampling is similar to
3:30:20multiple sampling plans except that it
3:30:22can in theory continue indefinitely
3:30:26in other words it is a non-probability
3:30:28sampling technique wherein the
3:30:30researcher picks a single subject or a
3:30:33group of subjects in a given time
3:30:35interval conducts the study analyzes the
3:30:38results and then picks another group of
3:30:41subjects if needed and so on
3:30:44in stratified sampling the idea is to
3:30:47take samples from subgroups of a
3:30:49population
3:30:51this technique gives an accurate
3:30:52estimate of the population parameter
3:30:56in this screen we will compare simple
3:30:58random sampling with stratified sampling
3:31:02simple random sampling is easy to do
3:31:05while stratified sampling takes a lot of
3:31:07time
3:31:08the possibility of simple random
3:31:11sampling giving erroneous results is
3:31:13very high
3:31:15while stratified sampling minimizes the
3:31:17chances of error
3:31:19simple random sampling doesn't have the
3:31:22power to show possible causes of
3:31:24variation while stratified sampling if
3:31:27done correctly will show assignable
3:31:29causes
3:31:31in the next screen we will look at the
3:31:33check sheet method of collecting data
3:31:36the process of collecting data is
3:31:38expensive wrongly collected data leading
3:31:41to wrong analysis and inferences results
3:31:44in resources being wasted
3:31:47a check sheet is a structured form
3:31:49prepared to collect and analyze data
3:31:52it is a generic tool that is relatively
3:31:54simple to use and can be adopted for a
3:31:57variety of purposes
3:31:59check sheets are used when the data can
3:32:01be observed and collected repeatedly by
3:32:04the same person or at the same location
3:32:07they are also used while collecting data
3:32:10from a production process
3:32:12a common example is calculating the
3:32:14number of absentees in a company
3:32:17the table shows absentee data collected
3:32:20for a week
3:32:21we will discuss data coding and its
3:32:24advantages in the following screen
3:32:27data coding is a process of converting
3:32:30and condensing raw data into categories
3:32:33and sets so that the data can be used
3:32:35for further analysis
3:32:37the benefits of data coding are listed
3:32:39here
3:32:41data coding simplifies the large
3:32:43quantity of data that is collected from
3:32:45sources
3:32:46the large amount of data makes analysis
3:32:49and drawing conclusions difficult
3:32:51it leads to chaos and ambiguity
3:32:54data coding simplifies the data by
3:32:56coding it into variables and then
3:32:59categorizing these variables
3:33:02raw data cannot be easily entered into
3:33:04computers for analysis
3:33:06data coding is used to convert raw data
3:33:09into process data that can be easily fed
3:33:12into computing systems for calculation
3:33:15and analysis
3:33:17coding of data makes it easy to analyze
3:33:20the data
3:33:21converted data can either be analyzed
3:33:23directly or fed into computers
3:33:26the analyst can easily draw conclusions
3:33:28when all the data is categorized and
3:33:31computerized
3:33:33data coding also enables organized
3:33:35representation of data
3:33:37division of data into categories helps
3:33:40organize large chunks of information
3:33:43thus making analysis and interpretation
3:33:45easier data coding also ensures that
3:33:48data repetition does not occur and
3:33:51duplicate entries are eliminated so that
3:33:53the final result is not affected in the
3:33:56following screen we will discuss
3:33:58measures of central tendency of the
3:34:00descriptive statistics in detail
3:34:04a measure of central tendency is a
3:34:06single value that indicates the central
3:34:09point in a set of data and helps in
3:34:11identifying data trends the three most
3:34:14commonly used measures of the central
3:34:17tendency are mean median and mode
3:34:21click each measure to know more
3:34:24mean is the most common measure of
3:34:26central tendency it is the sum of all
3:34:29the data values divided by the number of
3:34:32data points
3:34:33also called arithmetic mean or average
3:34:36it is the most widely used measure of
3:34:38central tendency
3:34:41also known as positional mean median is
3:34:44the number present in the middle of the
3:34:46data set when the numbers are arranged
3:34:48in ascending or descending order
3:34:52if the data set has an even number of
3:34:54entries then the median is the mean of
3:34:56the two middle numbers
3:34:59median can also be calculated by the
3:35:01formula n plus one divided by two
3:35:04where n is the number of entries
3:35:08mode also known as frequency mean is the
3:35:11value that occurs most frequently in a
3:35:14set of data
3:35:15data sets that have more than one mode
3:35:18are known as bimodal data
3:35:22let us look at an example for
3:35:24determining mean median and mode in this
3:35:27screen
3:35:28the data set has the numbers one two
3:35:31three four five five six seven and eight
3:35:36click the button to know the answer
3:35:39as previously defined
3:35:41mean is the sum of all the data items
3:35:44divided by the number of items
3:35:46therefore the mean is equal to 41
3:35:49divided by 9 which is equal to four
3:35:52point five six
3:35:54the number in the middle of the data set
3:35:56is five
3:35:57therefore the median is five
3:36:00mode is the most frequently occurring
3:36:02number which is again five
3:36:05in this screen we will understand the
3:36:07effect of outliers on the data set
3:36:10let us consider a minor change to the
3:36:13data set a new number 100 is added to
3:36:16the data set
3:36:18on using the same formula to calculate
3:36:20mean
3:36:21the new mean is
3:36:2315.11
3:36:24ideally 50 percent of values should lie
3:36:27on either side of the mean value
3:36:30however in this example it can be seen
3:36:33that almost 90 percent of values lie
3:36:35below the mean value of 15.11
3:36:39and only one value above the mean
3:36:42the data point 100 is called an outlier
3:36:46an outlier is an extreme value in the
3:36:48data set that skews the mean value to
3:36:51one side of the data set
3:36:53note that the median remains unchanged
3:36:55at five therefore mean is not an
3:36:58appropriate measure of central tendency
3:37:01if the data has outliers
3:37:03median is preferred in this case
3:37:06in the next screen we will look at
3:37:07measures of dispersion of the
3:37:09descriptive statistics
3:37:12apart from central tendency another
3:37:15important parameter to describe a data
3:37:17set is spread or dispersion
3:37:20contrary to the measures of central
3:37:22tendency such as mean median and mode
3:37:26measures of dispersion express the
3:37:28spread of values
3:37:30higher the variation of data points
3:37:32higher the spread of the data
3:37:35the three main measures of dispersion
3:37:37are range
3:37:38variance and standard deviation
3:37:41we will discuss each of these in the
3:37:44upcoming screens
3:37:46let us start with the first measure of
3:37:48dispersion range
3:37:50the range of a particular set of data is
3:37:52defined as the difference between the
3:37:54largest and smallest values of the data
3:37:58in the example the largest value of the
3:38:01data is nine and the smallest value is
3:38:04one
3:38:05therefore the range is nine minus one
3:38:08eight
3:38:09in calculating range all the data points
3:38:11are not needed and only the maximum and
3:38:14minimum values are required
3:38:17let us understand the next measure of
3:38:19dispersion variance in the following
3:38:22screen
3:38:25the variance denoted as sigma square or
3:38:28s square
3:38:29is defined as the average of squared
3:38:32mean differences
3:38:33and shows the variation in a data set
3:38:37to calculate the variance for a sample
3:38:39data set of 10 numbers type the numbers
3:38:42in an excel sheet calculate the variance
3:38:45using the formula equals v-a-r-p or
3:38:48v-a-r-s
3:38:51the v-a-r-p formula gives the population
3:38:54variance which is 7.24 for this example
3:38:58the v a r s formula gives the sample
3:39:01variance 8.04
3:39:05population variance is calculated when
3:39:07the data set is for the entire
3:39:09population and sample variance is
3:39:11calculated when data is available only
3:39:14for a sample of the population
3:39:16population variance is preferred over
3:39:19sample variance as the latter is only an
3:39:22estimate
3:39:23sample variance allows for a broader
3:39:25range of possible answers for the true
3:39:28mean of the population
3:39:29that is the confidence levels are higher
3:39:32in sample variance
3:39:34note that variance is a measure of
3:39:36variation and cannot be considered as
3:39:38the variation in a data set
3:39:41in the following screen we will
3:39:43understand the next measure of
3:39:44dispersion
3:39:46standard deviation
3:39:48standard deviation denoted by sigma or s
3:39:52is given by the square root of variance
3:39:55the statistical notation of this is
3:39:57given on screen
3:39:59standard deviation is the most important
3:40:01measure of dispersion
3:40:03standard deviation is always relative to
3:40:06the mean
3:40:07for the same data set the population
3:40:10standard deviation is 2.69
3:40:13and sample standard deviation is 2.83
3:40:17as in variance calculation if the data
3:40:20set is measured for every unit in a
3:40:22population
3:40:23the population standard deviation and
3:40:26sample standard deviation can be
3:40:28calculated in excel using the formula
3:40:31given on the screen
3:40:33the steps to manually calculate the
3:40:35standard deviation are
3:40:38first calculate the mean
3:40:40then calculate the difference between
3:40:43each data point and the mean and square
3:40:45that answer
3:40:47next calculate the sum of the squares
3:40:50next divide the sum of the squares by n
3:40:53or n minus 1 to find the variance
3:40:57lastly find the square root of variance
3:40:59which gives the standard deviation
3:41:02in the next screen we will look at
3:41:04frequency distribution of the
3:41:05descriptive statistics
3:41:08frequency distribution is a method of
3:41:10grouping data into mutually exclusive
3:41:13categories showing the number of
3:41:14observations in each class
3:41:17an example is presented to demonstrate
3:41:20frequency distribution
3:41:22a survey was conducted among the
3:41:24residents of a particular area to
3:41:26collect data on cars owned by each home
3:41:30a total of 20 homes were surveyed
3:41:33to create a frequency table for the
3:41:35results collected in the survey
3:41:37the first step is to divide the results
3:41:39into intervals and count the number of
3:41:42results in each interval
3:41:44for instance in this example
3:41:47the intervals would be the number of
3:41:48households with no car
3:41:51one car
3:41:52two cars and so on
3:41:55next a table is created with separate
3:41:57columns for the intervals the tallied
3:42:00results for each interval and the number
3:42:02of occurrences or frequency of results
3:42:05in each interval
3:42:07each result for a given interval is
3:42:09recorded with a tally mark in the second
3:42:11column the tally marks for each interval
3:42:14are added and the sum is entered in the
3:42:16frequency column
3:42:18the frequency table allows viewing
3:42:20distribution of data across a set of
3:42:22values at a glance in the following
3:42:24screen we will look at cumulative
3:42:27frequency distribution
3:42:29a cumulative frequency distribution
3:42:32table is similar to the frequency
3:42:34distribution table only more detailed
3:42:37there are additional columns for
3:42:39cumulative frequency percentage and
3:42:42cumulative percentage
3:42:44in the cumulative frequency column the
3:42:46cumulative frequency or the previous row
3:42:49or rows is added to the current row
3:42:52the percentage is calculated by dividing
3:42:54the frequency by the total number of
3:42:56results and multiplying by one hundred
3:42:59the cumulative percentage is calculated
3:43:02similar to the cumulative frequency
3:43:06let us look at an example for cumulative
3:43:08frequency distribution
3:43:11the ages of all the participants in a
3:43:13chess tournament are recorded
3:43:16the lowest age is 37 and the highest is
3:43:1991.
3:43:20keeping intervals of 10
3:43:23the lowest interval starts with the
3:43:24lower limit as 35
3:43:27and the upper limit as 44.
3:43:30similar intervals are created until an
3:43:32upper limit of ninety four
3:43:34in the frequency column the number of
3:43:37times a result appears in a particular
3:43:40interval is recorded in the cumulative
3:43:42frequency column the cumulative
3:43:44frequency of the previous row is added
3:43:46to the frequency of the current row
3:43:49for the first row the cumulative
3:43:51frequency is the same as the frequency
3:43:54in the second row the cumulative
3:43:56frequency is one plus two which is three
3:43:59and so on
3:44:01in the percentage column the percentage
3:44:03of the frequency is listed by dividing
3:44:06the frequency by the total number of
3:44:08results which is 10 and multiplying the
3:44:11value by 100
3:44:13for instance in the first row the
3:44:15frequency is one and the number of
3:44:18results is ten
3:44:20therefore the percentage is ten the
3:44:22final column is the cumulative
3:44:24percentage column in this column the
3:44:27cumulative frequency is divided by the
3:44:29total number of results which is ten and
3:44:32the value is multiplied by hundred
3:44:35note that the last number in this column
3:44:37should be equal to one hundred in this
3:44:39example the cumulative frequency is one
3:44:42and the total number of results is ten
3:44:45therefore the cumulative percentage of
3:44:47the first row is ten
3:44:49let us look at the stem and leaf plots
3:44:52which is one of the graphical methods of
3:44:54understanding distribution
3:44:56graphical methods are extremely useful
3:44:59tools to understand how data is
3:45:01distributed
3:45:02sometimes merely by looking at the data
3:45:04distribution errors in a process can be
3:45:07identified
3:45:09the stem and leaf method is a convenient
3:45:11method of manually plotting data sets
3:45:14it is used for presenting data in a
3:45:16graphical format to assist visualizing
3:45:19the shape of a given distribution
3:45:22in the example on the screen the
3:45:24temperatures in fahrenheit for the month
3:45:26of may are given
3:45:28to collate this information in a stem
3:45:30and leaf plot
3:45:31all the tens digits are entered in the
3:45:33stem column and all the units digits
3:45:36against each tens digit are entered in
3:45:38the leaf column to start with the lowest
3:45:41value is considered
3:45:42in this case the lowest temperature is
3:45:4551.
3:45:46in the first row five is entered in the
3:45:48stem column and zero in the leaf column
3:45:52the next lowest temperature is 58.
3:45:54eight is entered in the leaf column
3:45:56corresponding to five in the stem
3:45:59the next number is fifty-nine
3:46:01all the temperatures falling in the
3:46:03fifties are similarly entered
3:46:05in the next row the same process is
3:46:07repeated for temperatures in the sixties
3:46:10this is continued till all the
3:46:12temperature values are entered in the
3:46:14table
3:46:16let us understand another graphical
3:46:18method in the next screen
3:46:20box and whisker plots
3:46:23a box and whisker graph based on medians
3:46:26or quartiles is used to display a data
3:46:29set in a way that allows viewing the
3:46:30distribution of the data points easily
3:46:33consider the following example the
3:46:36lengths of 13 fish caught in a lake were
3:46:39measured and recorded
3:46:40the data set is given on the screen
3:46:43the first step to draw a box and whisker
3:46:45plot is therefore to arrange the numbers
3:46:48in increasing order next find the median
3:46:51as there is an odd number of data
3:46:54entries the median is the number in the
3:46:56middle of the data set which in this
3:46:58case is 12.
3:47:00the next step is to find the lower
3:47:02median or quartile
3:47:04this is the median of the lower six
3:47:06numbers
3:47:07the middle of these numbers is halfway
3:47:09between 8 and 9 which would be 8.5
3:47:13similarly the upper median or quartile
3:47:16is located for the upper six numbers to
3:47:18the right of the median the upper median
3:47:21is halfway between the two values
3:47:23fourteen and fourteen
3:47:25therefore the upper median is fourteen
3:47:29let us now understand how the box and
3:47:31whisker chart is drawn using the values
3:47:34of the median and upper and lower
3:47:36quartiles
3:47:37the next step is a numbered line is
3:47:39drawn extending far enough to include
3:47:42all the data points
3:47:44then a vertical line is drawn from the
3:47:46median point 12.
3:47:48the lower and upper quartiles 8.5 and 14
3:47:52respectively are marked with vertical
3:47:54lines and these are joined with the
3:47:56median line to form two boxes as shown
3:47:59on the screen
3:48:00next two whiskers are extended from
3:48:03either ends of the boxes as shown to the
3:48:05smallest and largest numbers in the data
3:48:07set
3:48:085 and 20 respectively
3:48:10the box and whiskers graph is now
3:48:13complete the following inferences can be
3:48:16drawn from the box and whisker plot
3:48:18the lengths of the fish range from five
3:48:20to twenty the range is therefore fifteen
3:48:24the quartiles split the data into four
3:48:27equal parts
3:48:28in other words one quarter of the data
3:48:30numbers is less than 8.5
3:48:33one quarter between 8.5 and 12
3:48:36next quarter of the data numbers are
3:48:38between 12 and 14
3:48:40and another quarter has data numbers
3:48:43greater than 14.
3:48:46in this screen we will learn about
3:48:48another graphical method scatter
3:48:50diagrams
3:48:52a scatter diagram or scatter plot is a
3:48:55tool used to analyze the relationship or
3:48:58correlation between two sets of
3:49:00variables x and y
3:49:02with x as the independent variable and y
3:49:05as the dependent variable
3:49:07a scatter diagram is also useful when
3:49:10cause effect relationships have to be
3:49:12examined or root causes have to be
3:49:15identified
3:49:16there are five different types of
3:49:18correlation that can be used in a
3:49:20scatter diagram
3:49:22let us learn about them in the next
3:49:23screen
3:49:26the five types of correlation are
3:49:28perfect positive correlation moderate
3:49:31positive correlation
3:49:32no relation or no correlation
3:49:35moderate negative correlation
3:49:37and perfect negative correlation
3:49:40click each type to learn more
3:49:43in perfect positive correlation the
3:49:46value of dependent variable y increases
3:49:49proportionally with any increase in the
3:49:51value of independent variable x
3:49:54this is said to be one is to one
3:49:57that is any change in one variable
3:49:59results in an equal amount of change in
3:50:02the other
3:50:03the following example is presented to
3:50:05demonstrate perfect positive correlation
3:50:08the consumption of milk is found to
3:50:11increase proportionally with an increase
3:50:13in the consumption of coffee
3:50:15the data is presented in the table on
3:50:17the screen
3:50:19the scatter diagram for the data is also
3:50:21shown
3:50:22it can be observed from the graph that
3:50:24as x increases y also increases
3:50:28proportionally
3:50:29hence the points are linear
3:50:33in this type of correlation as the value
3:50:36of the x variable increases the value of
3:50:39y also increases but not in the same
3:50:42proportion
3:50:44to demonstrate this the following
3:50:46example is presented the increase in
3:50:49savings for increase in salary is shown
3:50:51in the table
3:50:53as you can notice in the scatter diagram
3:50:56the points are not linear
3:50:58although the value of y increases with
3:51:01increase in the value of x the increase
3:51:04is not proportional
3:51:07when a change in one variable has no
3:51:10impact on the other
3:51:11there is no relation or correlation
3:51:14between them
3:51:15let us consider the following example
3:51:18to study the relation between the number
3:51:20of fresh graduates in the city and the
3:51:23job openings available
3:51:25data for both was collected over a few
3:51:27months and tabulated as shown
3:51:30the scatter diagram for the same is also
3:51:32displayed
3:51:34it can be observed that the data points
3:51:36are scattered and there is no trend
3:51:38emerging from the graph
3:51:40therefore there is no correlation
3:51:42between the number of fresh graduates
3:51:45and the number of job openings in the
3:51:47city
3:51:50in moderate negative correlation an
3:51:52increase in one variable results in a
3:51:55decrease in the other variable however
3:51:58this change is not proportional to the
3:52:00change in the first variable
3:52:02to demonstrate modern negative
3:52:04correlation
3:52:05the prices of different products are
3:52:07listed along with the number of units
3:52:10sold for each product the data is shown
3:52:12in the table
3:52:14from the scatter diagram shown
3:52:16it can be observed that higher the price
3:52:19of a product
3:52:20lesser are the number of units of that
3:52:22product sold
3:52:24however the decrease in the number of
3:52:26units with increasing price is not
3:52:28proportional
3:52:32in perfect negative correlation an
3:52:34increase in one variable results in a
3:52:37proportional decrease of the other
3:52:39variable
3:52:40this is also an example of one is to one
3:52:43correlation
3:52:45as an example the effect of an increase
3:52:47in the project time extension on the
3:52:49success of project is considered
3:52:52the data is shown in the table the
3:52:55scatter diagram for the data shows a
3:52:57proportional decrease in the probability
3:53:00of the project's success with each
3:53:02extension of the project time
3:53:04hence the points are linear
3:53:07perfect correlations are rare in the
3:53:09real world
3:53:10when encountered they should be
3:53:12investigated and verified
3:53:16in this screen we will look at another
3:53:18graphical method histograms
3:53:21histograms are similar to bar graphs
3:53:23except that the data in histograms is
3:53:26grouped into intervals
3:53:28they are used to represent category wise
3:53:30data graphically
3:53:32a histogram is best suited for
3:53:33continuous data
3:53:35the following example illustrates how a
3:53:38histogram is used to represent data
3:53:41data on the number of hours spent by a
3:53:43group of 15 people on a special project
3:53:46in one week is collected
3:53:49this data is then divided into intervals
3:53:51of two and the frequency table for the
3:53:54data is created
3:53:56the histogram for the same data is also
3:53:58displayed
3:54:00looking at the histogram it can be
3:54:02observed at a glance that most of the
3:54:05team members spent between two to four
3:54:07hours on the project
3:54:09in the following screen we will look at
3:54:11the next graphical method
3:54:14normal probability plots
3:54:16normal probability plots are used to
3:54:19identify if a sample has been taken from
3:54:21a normal distributed population
3:54:24when sample data from a normal
3:54:25distributive population is represented
3:54:28as a normal probability plot it forms a
3:54:31straight line
3:54:33the following example is presented to
3:54:35illustrate normal probability plots
3:54:38a sampling of diameters from a drilling
3:54:40operation is done and the data is
3:54:42recorded
3:54:43the data set is given to create a normal
3:54:47probability plot the first step is to
3:54:49construct a cumulative frequency
3:54:51distribution table
3:54:53this is followed by calculating the mean
3:54:55rank probability by dividing the
3:54:57cumulative frequency by the number of
3:54:59samples plus one and multiplying the
3:55:02answer by one hundred
3:55:04the fully populated table for mean rank
3:55:07probability estimation is shown on the
3:55:09screen
3:55:10please take a look at the same
3:55:12in the next step a graph is plotted on
3:55:14log paper or with minitab using this
3:55:17data
3:55:18minitab is a statistical software used
3:55:21in six sigma
3:55:22minitab normal probability plot
3:55:24instructions are also given on the
3:55:26screen the completed graph is shown on
3:55:29the screen
3:55:30from the graph it can be seen that the
3:55:32random sample forms a straight line and
3:55:35therefore the data is taken from a
3:55:36normally distributed population
3:55:39let us proceed to the next topic of this
3:55:42lesson
3:55:43in this topic we will discuss
3:55:45measurement system analysis or msa in
3:55:48detail
3:55:49let us understand what msa is in the
3:55:51following screen throughout the dmac
3:55:53process the output of the measurement
3:55:55system ms is used for metrics analysis
3:55:59and control efforts
3:56:00an error prone measurement system will
3:56:02only lead to incorrect data
3:56:05incorrect data leads to incorrect
3:56:06conclusions
3:56:08it is important to set right the ms
3:56:10before collecting the data
3:56:13measurement system analysis or msa
3:56:16is a technique that identifies
3:56:18measurement error or variation and
3:56:21sources of that error in order to reduce
3:56:23the variation
3:56:25it evaluates the measuring system to
3:56:27ensure the integrity of data used for
3:56:30analysis
3:56:31msa is therefore one of the first
3:56:33activities in the measure phase
3:56:36the measurement systems capability is
3:56:38calculated analyzed and interpreted
3:56:41using gauge repeatability and
3:56:43reproducibility to determine measurement
3:56:45correlation bias linearity percent
3:56:49agreement and precision or tolerance
3:56:52let us discuss the objectives of msa in
3:56:55the next screen
3:56:59a primary objective of msa is to obtain
3:57:02information about the type of
3:57:04measurement variation associated with
3:57:06the measurement system
3:57:08it is also used to establish criteria to
3:57:11accept and release new measuring
3:57:12equipment
3:57:14msa also compares measuring one method
3:57:17against another it helps to form a basis
3:57:20for evaluating a method which is
3:57:22suspected of being deficient
3:57:24the measurement system variations should
3:57:27be resolved to arrive at the correct
3:57:28baselines for the project objectives
3:57:31as baselines contain crucial data based
3:57:34on which decisions are taken it is
3:57:36extremely important that the measurement
3:57:38system be free of error as far as
3:57:40possible
3:57:42let us look at measurement analysis in
3:57:44detail in the next screen
3:57:48in measurement analysis the observed
3:57:50value is equal to the sum of the true
3:57:52value and the measurement error
3:57:55the measurement error can be a negative
3:57:57or a positive value
3:57:59measurement error refers to the net
3:58:01effect of all sources of measurement
3:58:03variability that cause an observed value
3:58:06to deviate from the true value
3:58:09true variability is the sum of the
3:58:11process variability and the measurement
3:58:14variability process variability and
3:58:17measurement variability must be
3:58:19evaluated and improved together
3:58:22measurement variability should be
3:58:24addressed before looking at process
3:58:26variability
3:58:28if process variability is corrected
3:58:30before resolving measurement variability
3:58:33then any improvements to the process can
3:58:36not be trusted to have taken place owing
3:58:38to a faulty measurement system
3:58:41in the following screen we will identify
3:58:44the types of measurement errors
3:58:47the two types of measurement errors are
3:58:50measurement system bias and measurement
3:58:52system variation
3:58:54click each type to know more
3:58:56measurement system bias involves
3:58:58calibration study
3:59:00in the calibration study the total mean
3:59:02is given by the sum of the process mean
3:59:04and the measurement mean
3:59:06the statistical notation is shown on the
3:59:09screen
3:59:10measurement system variation involves
3:59:13gauge repeatability and reproducibility
3:59:16or grrr study
3:59:18in the grr study the total variance is
3:59:21calculated by adding the process
3:59:23variance with the measurement variation
3:59:25the statistical notation is shown on the
3:59:28screen
3:59:30in this screen we will discuss the
3:59:32sources of variation the chart on the
3:59:34screen lists the different sources of
3:59:36variation
3:59:37observed process variation is divided
3:59:40into two
3:59:41actual process variation and measurement
3:59:43variation
3:59:45actual process variation can be divided
3:59:48into long-term and short-term process
3:59:50variations
3:59:52in a gauge rr study process variation is
3:59:55often called part variation
3:59:58measurement variation can be divided
4:00:00into variations caused by operators and
4:00:03variations due to the gauge
4:00:05the variation due to operators is owing
4:00:08to reproducibility
4:00:10variation due to gauge is owing to
4:00:12repeatability
4:00:14both actual process variation and
4:00:16measurement variation have a common
4:00:18factor that is variation within a sample
4:00:22let us understand gauge repeatability
4:00:25and reproducibility or grr in the next
4:00:28screen
4:00:30gauge repeatability and reproducibility
4:00:32or grr is a statistical technique to
4:00:36assess if a gauge or gauging system will
4:00:38obtain the same reading each time a
4:00:40particular characteristic or parameter
4:00:42is measured
4:00:44gauge repeatability is the variation in
4:00:47measurement when one operator uses the
4:00:49same gauge to measure identical
4:00:51characteristics of the same part
4:00:53repeatedly
4:00:55gauge reproducibility is the variation
4:00:57in the average of measurements when
4:00:59different operators use the same gauge
4:01:02to measure identical characteristics of
4:01:04the same part
4:01:05the figures on the screen illustrate
4:01:07gauge repeatability and reproducibility
4:01:11in the next screen we will discuss the
4:01:13components of grr study
4:01:17the figure on the screen illustrates the
4:01:19difference between gauge repeatability
4:01:22and reproducibility
4:01:24the figure shows the repeatability and
4:01:26reproducibility for six different parts
4:01:29represented by the numbers one to six
4:01:31for two different trial readings by
4:01:33three different operators
4:01:35as can be observed a difference in
4:01:38reading for part one indicated by the
4:01:40color green by three different operators
4:01:43is known as reproducibility error a
4:01:46difference in reading of part 4
4:01:48indicated by red by the same operator in
4:01:51two different trials is known as the
4:01:53repeatability error
4:01:55in the following screen we will look at
4:01:57some guidelines for gauge repeatability
4:02:00and reproducibility studies
4:02:03the following should be kept in mind
4:02:05while carrying out gauge repeatability
4:02:07and reproducibility or grr studies
4:02:11grr studies should be performed over the
4:02:14range of expected observations
4:02:16care should be taken to use actual
4:02:18equipment for gr studies
4:02:21written procedures and approved
4:02:23practices should be followed as would
4:02:25have been in actual operations
4:02:27the measurement variability should be
4:02:29represented as is
4:02:31not the way it was designed to be
4:02:34after grr the measurement variability is
4:02:37separated into casual components sorted
4:02:40according to priority and then targeted
4:02:42for action
4:02:43in the following screen let us look at
4:02:46some more concepts associated with grr
4:02:51bias is the distance between the sample
4:02:53mean value and the sample true value
4:02:56it is also called accuracy
4:02:59bias is equal to mean minus reference
4:03:01value process variation is equal to six
4:03:04times the standard deviation
4:03:06the bias percentage is calculated as
4:03:09bias divided by the process variation
4:03:12the next term is linearity
4:03:14linearity refers to the consistency of
4:03:17bias over the range of the gauge
4:03:19linearity is given by the product of
4:03:21slope and process variation
4:03:24precision is the degree of repeatability
4:03:27or closeness of data
4:03:29smaller the dispersion in the data set
4:03:32better the precision
4:03:33the variation in the gauge is the sum of
4:03:36variation due to repeatability and the
4:03:38variation due to reproducibility
4:03:42in the following screen we will
4:03:43understand measurement resolution
4:03:47measurement resolution is the smallest
4:03:49detectable increment that an instrument
4:03:51will measure or display
4:03:53the number of increments in the
4:03:55measurement system should extend over
4:03:57the full range for a given parameter
4:04:00some examples of wrong gauges or
4:04:02incorrect measurement resolution are
4:04:04given here a truck weighing scale is
4:04:07used for measuring the weight of a
4:04:09t-pack
4:04:10a caliper capable of measuring
4:04:12differences of 0.1 millimeters
4:04:15is used to show compliance when the
4:04:17tolerance limits are plus or minus 0.07
4:04:20millimeters
4:04:21thus the measurement system that matches
4:04:24the range of the data should only be
4:04:26used
4:04:27an important prerequisite for grr
4:04:29studies is that the gauge has an
4:04:31acceptable resolution
4:04:34in the next screen we will look at
4:04:35examples for repeatability and
4:04:37reproducibility
4:04:40repeatability is also called equipment
4:04:43variation or ev
4:04:46it occurs when the same technician or
4:04:48operator repeatedly measures the same
4:04:50part or process under identical
4:04:53conditions with the same measurement
4:04:55system
4:04:56the following example illustrates this
4:04:58concept
4:04:59a 36 kilometer per hour pace mechanism
4:05:02is timed by a single operator over a
4:05:05distance of 100 meters on a stopwatch
4:05:08and three readings are taken
4:05:10trial one takes nine seconds trial two
4:05:13takes ten seconds and trial 3 takes 11
4:05:16seconds
4:05:18the process is measured with the same
4:05:20equipment in identical conditions by the
4:05:22same operator
4:05:23assuming no operator error the variation
4:05:26in the three readings is known as
4:05:28repeatability or equipment variation
4:05:32reproducibility is also called appraiser
4:05:35variation or av
4:05:38it occurs when different technicians or
4:05:40operators measure the same part or
4:05:42process under identical conditions using
4:05:44the same measurement system
4:05:47let us extend the example for
4:05:48repeatability to include data measured
4:05:51by two operators
4:05:52the readings are displayed on the slide
4:05:55the difference in the readings of both
4:05:57operators is called reproducibility or
4:06:00appraiser variation
4:06:02it is important to resolve equipment
4:06:04variation before appraiser variation
4:06:07if appraiser variation is resolved first
4:06:10the results will still not be identical
4:06:12due to variation in the equipment itself
4:06:16in this screen we will learn about data
4:06:19collection in grr
4:06:21there are some important considerations
4:06:23for data collection in grr studies there
4:06:26are usually three operators and around
4:06:28ten units to measure
4:06:30general sampling techniques must be used
4:06:33to represent the population and each
4:06:35unit must be measured two to three times
4:06:38by each operator it is important that
4:06:40the gauge be calibrated accurately it
4:06:43should also be ensured that the gauge
4:06:45has an acceptable resolution
4:06:48another practice is that the first
4:06:50operator measures all the units in
4:06:52random order
4:06:53then this order is maintained by all
4:06:55other operators
4:06:57all the trials must be repeated
4:06:59in the next screen we will discuss the
4:07:01anova method of analyzing grr studies
4:07:06the anova method is considered to be the
4:07:08best method for analyzing grr studies
4:07:11this is because of two reasons
4:07:14the first being anova not only separates
4:07:17equipment and operator variation but
4:07:19also provides insight on the combined
4:07:21effect of the two
4:07:23second anova uses standard deviation
4:07:26instead of range as a measure of
4:07:28variation and therefore gives a better
4:07:31estimate of the measurement system
4:07:33variation
4:07:34the one drawback of using anova is the
4:07:36considerations of time resources and
4:07:39cost
4:07:40in the next screen we will understand
4:07:42how msa can be interpreted
4:07:45two results are possible for an msa
4:07:49in the first case the reproducibility
4:07:51error is larger than the repeatability
4:07:54error
4:07:55this occurs when the operators are not
4:07:57trained and calibrations on the gauge
4:07:59dial are not clear
4:08:01the other possibility is that the
4:08:03repeatability error is larger than the
4:08:06reproducibility error
4:08:08this is clearly a maintenance issue and
4:08:10can be resolved by calibrating the
4:08:12equipment or performing maintenance on
4:08:14the equipment
4:08:16this indicates that the gauge needs
4:08:18redesign to be more rigid and the
4:08:20location needs to be improved
4:08:22it also occurs when there is ambiguity
4:08:25in sops
4:08:27msa is an experiment which seeks to
4:08:30identify the components of variation in
4:08:32the measurement
4:08:34in the following screen we will look at
4:08:36a template used for grr studies
4:08:40a sample gauge rr sheet is given on this
4:08:43screen
4:08:44the operators here are andrew murphy and
4:08:46lucy wang who are the appraisers in this
4:08:49study
4:08:50they have measured and rated the
4:08:52performance of three employees
4:08:54ibrahim glasoff brianna scott and jason
4:08:57schmidt
4:08:59this is a sample template for a gauge r
4:09:02r study
4:09:03the parts are shown across the top of
4:09:05the sheet
4:09:06in this case the measurement system is
4:09:08being evaluated using three parts the
4:09:11employees abraham glasoff brianna scott
4:09:15and jason schmidt
4:09:17the operators measure each part
4:09:19repeatedly
4:09:20from this data the average x and ranges
4:09:23are for each inspector and for each part
4:09:26are calculated
4:09:28the grand average for each inspector and
4:09:31each part is also calculated
4:09:34in this example a control limit ucl in
4:09:37the sheet was compared with the
4:09:39difference in averages of the two
4:09:41inspectors to identify if there is a
4:09:44significant difference in their
4:09:46measurements
4:09:47their difference is 0.111
4:09:51which is outside the ucl of 0.108
4:09:55given the r average of 0.042
4:09:59in the next screen we will look at the
4:10:01results page for this grr study
4:10:05the sheet on the screen displays the
4:10:07results for the data entered in the
4:10:08template in the previous screen
4:10:11please spend some time to go through the
4:10:12data for a better understanding of the
4:10:14concept
4:10:16in the following screen we will look at
4:10:18the interpretation to this results page
4:10:22the percentage grrr value is highlighted
4:10:25in the center right of the table in the
4:10:27previous screen
4:10:28there are three important observations
4:10:30to be made here about the gauge rr study
4:10:34first this study also shows the
4:10:36interaction between operators and parts
4:10:39if the percentage grr value is less than
4:10:4230 then the gauge is acceptable and the
4:10:45measurement system does not require any
4:10:47change
4:10:48if the value is greater than 30 then the
4:10:51gauge needs correction
4:10:53the equipment variation is checked and
4:10:55resolved first followed by the appraiser
4:10:58variation
4:10:59second if ev equals zero it means the ms
4:11:03is reliable the equipment is perfect and
4:11:06the variation in the gauge is
4:11:08contributed by different operators
4:11:10if the av is equal to zero the ms is
4:11:14precise
4:11:15third if ev is equal to zero and there
4:11:18is a v the operators have to be trained
4:11:21to ensure all operators follow identical
4:11:24steps during measurement and the av is
4:11:27minimal
4:11:28the interaction between operators and
4:11:30parts can also be studied under grr
4:11:33using part variation the trueness and
4:11:36precision cannot be determined in a grrr
4:11:39if only one gauge or measurement method
4:11:41is evaluated as it may have an inherent
4:11:44bias that would go undetected merely by
4:11:47varying operators and parts
4:11:50let us proceed to the next topic of this
4:11:52lesson in the following screen
4:11:55in this topic we will discuss process
4:11:57and performance capability in detail
4:12:00in the following screen we will look at
4:12:02the differences between natural process
4:12:05limits and specification limits
4:12:08natural process limits or control limits
4:12:12are derived from the process data and
4:12:14are the voice of the process
4:12:16the data consists of real-time values
4:12:19from past process performance
4:12:21therefore these values represent the
4:12:24actual process limits
4:12:26and indicate variation in the process
4:12:29the two control limits are upper control
4:12:32limit ucl and lower control limit lcl
4:12:37specification limits are provided by
4:12:39customers based on their requirements or
4:12:42the voice of the customer and cannot be
4:12:44changed by the organization
4:12:47these limits act as targets for the
4:12:49organization
4:12:51and processes are designed around the
4:12:53requirements
4:12:54the product or service has to meet
4:12:56customer requirements and has to be well
4:12:59within the specification limits
4:13:02if the product or service does not meet
4:13:04customer requirements it is considered
4:13:06as a defect
4:13:08therefore specification limits are the
4:13:10intended results or requirements from
4:13:13the product or service that are defined
4:13:15by the customer
4:13:17the two specification limits are upper
4:13:19specification limit or usl
4:13:22and lower specification limit or lsl
4:13:27the difference between the two is called
4:13:29tolerance
4:13:30an important point to note is that for a
4:13:33process
4:13:34if the control limits lie within the
4:13:36specification limits the process is said
4:13:38to be under control
4:13:41conversely if specification limits lie
4:13:44within the control limits the process
4:13:46will not meet customer requirements
4:13:49in the following screen we will look at
4:13:51process performance metrics and how they
4:13:54are calculated
4:13:56the two major metrics used to measure
4:13:58process performance are defects per unit
4:14:01or dpu
4:14:02and defects per million opportunities or
4:14:06dpmo dpu is calculated by dividing the
4:14:10number of defects by the total number of
4:14:12units
4:14:13dpmo is calculated by multiplying the
4:14:16defects per opportunity with 1 million
4:14:20in the following screen we will look at
4:14:22an example for calculating process
4:14:24performance
4:14:28in this example the quality control
4:14:30department checks the quality of
4:14:31finished goods by sampling a batch of 10
4:14:34items from the produced lot every hour
4:14:36the data is collected over 24 hours
4:14:40the table displays the data for the
4:14:41number of defectives for the sampling
4:14:43period
4:14:45if items are consistently found to be
4:14:47outside the control limits on any given
4:14:49day the production process is stopped
4:14:52for the next day
4:14:53let us now interpret the results of the
4:14:55sampling
4:14:57in this example as the sample size is
4:14:59constant dpu or p-bar is used to
4:15:03calculate the process capability
4:15:06the total number of defectives is 34 and
4:15:09the subgroup size is 10.
4:15:11the total number of units is 10
4:15:13multiplied by 24 which is 240.
4:15:17the defects per unit is 0.0124
4:15:21the defects per million opportunities is
4:15:24obtained by multiplying the defects per
4:15:26unit with 1 million which is 141
4:15:29666.66
4:15:32therefore by looking at the dpmo table
4:15:36it can be said that the process is
4:15:37currently working at 2.6 sigma or 86.4
4:15:42percent yield
4:15:44we will learn about process stability
4:15:46studies in this screen
4:15:48the activities carried out in the
4:15:50measure phase are msa collection of data
4:15:54statistical calculations and checking
4:15:56for accuracy and validity
4:15:59this is followed by a test for stability
4:16:01as changes cannot be made to an unstable
4:16:04process with a set of data believed to
4:16:06be accurate the process is checked for
4:16:08stability
4:16:10this is important because if a process
4:16:12is unstable no changes can be
4:16:15implemented
4:16:16why does a process become unstable
4:16:19a process can become unstable due to
4:16:21special causes of variation
4:16:24multiple special causes of variation
4:16:26lead to instability
4:16:28a single special cause leads to an out
4:16:30of control condition
4:16:32run charts in minitab can be used to
4:16:35check for process stability
4:16:37let us look at the steps to plot a run
4:16:39chart in minitab in the following screen
4:16:42to plot a run chart in minitab first
4:16:45enter the sample data collected to check
4:16:47for stability
4:16:49next click stat on the minitab window
4:16:51followed by quality tools
4:16:54next click run charts
4:16:56select the column and choose the
4:16:58subgroup size as two click ok
4:17:02the graph shown on the screen is
4:17:03interpreted by looking at the last four
4:17:05values
4:17:06if any of the p values is less than 0.05
4:17:11the presence of special causes of
4:17:13variation can be validated
4:17:15this means there is a good chance that
4:17:17the process will become unstable
4:17:20in the following screen we will look at
4:17:23process stability studies causes of
4:17:25variation
4:17:28variation can be due to two types of
4:17:30causes
4:17:31common causes of variation and special
4:17:34causes of variation
4:17:36click each type to learn more
4:17:39common causes of variation are the many
4:17:41sources of variation within a process
4:17:44which have a stable and repeatable
4:17:46distribution over a period
4:17:48they contribute to a state of
4:17:49statistical control where the output is
4:17:52predictable
4:17:54some other factors which do not always
4:17:56act on the process can also cause
4:17:58variation
4:17:59these are special causes of variation
4:18:02these are external to the process and
4:18:05are irregular in nature
4:18:07when present the process distribution
4:18:09changes and the process output is not
4:18:11stable over a period
4:18:13special causes may result in defects and
4:18:16need to be eliminated to bring the
4:18:18process under control
4:18:20run charts indicate the presence of
4:18:22special causes of variation in the
4:18:24process
4:18:25if special causes are detected the
4:18:28process has to be brought to a stop and
4:18:30a root cause analysis has to be carried
4:18:33out if the root cause analysis reveals
4:18:36the special cause to be undesirable
4:18:38corrective actions are taken to remove
4:18:40the special cause
4:18:43we will learn about verifying process
4:18:45stability and normality in this screen
4:18:48based on the type of variation of
4:18:50process exhibits it can be verified if
4:18:52the process is in control
4:18:55if there are special causes of variation
4:18:58the process output is not stable over
4:19:00time
4:19:01the process cannot be said to be in
4:19:03control
4:19:05conversely if there are only common
4:19:07causes of variation in a process the
4:19:10output forms a distribution that is
4:19:12stable and predictable over time
4:19:15a process being in control means the
4:19:17process does not have any special causes
4:19:20of variation
4:19:22once a process is understood to be
4:19:24stable the control chart data can be
4:19:26used to calculate the process capability
4:19:29indices
4:19:30in the following screen we will discuss
4:19:32process capability studies
4:19:35process capability is the actual
4:19:37variation in the process specification
4:19:40to carry out a process capability study
4:19:43first plan for data collection next
4:19:46collect the data finally plot and
4:19:48analyze the results
4:19:50obtaining the appropriate sampling plan
4:19:53for the process capability study depends
4:19:55on the purpose and whether there are any
4:19:57customer or standard requirements for
4:19:59the study
4:20:01for new processes or a project proposal
4:20:04the project capability can be estimated
4:20:06by a pilot run
4:20:08let us look at the objectives of process
4:20:10capability studies in the next screen
4:20:13the objectives of a process capability
4:20:15study are to establish a state of
4:20:17control over a manufacturing process
4:20:20and then maintain the state of control
4:20:22over a period of time
4:20:24on comparing the natural process limits
4:20:26or the control limits within the
4:20:28specification limits any of the
4:20:30following outcomes is possible
4:20:33first the process limits are found to
4:20:35fall between the specification limits
4:20:37this shows the process is running well
4:20:40and no action is required
4:20:42the second possibility is that the
4:20:44process spread and the specification
4:20:47spread are approximately the same
4:20:50in this case the process is centered by
4:20:52making an adjustment to the centering of
4:20:54the process
4:20:56this would bring the batch of products
4:20:57within the specifications
4:21:00the third possibility is that the
4:21:02process limits fall outside the
4:21:04specification limits
4:21:06in this case reduce the variability by
4:21:09partitioning the pieces of batches to
4:21:11locate and target the largest offender
4:21:14a design experiment can be used to
4:21:16identify the primary source of variation
4:21:19in the following screen we will learn
4:21:21about identifying characteristics in
4:21:23process capability
4:21:25process capability deals with the
4:21:27ability of the process to meet customer
4:21:30requirements
4:21:31therefore it is crucial that the
4:21:33characteristics selected for a process
4:21:35capability study indicates a key factor
4:21:38in the quality of the product or process
4:21:41also it should be possible to influence
4:21:43the value of the characteristic by
4:21:45adjusting the process
4:21:47the operating conditions that affect the
4:21:49characteristic should also be defined
4:21:51and controlled
4:21:53apart from these requirements other
4:21:55factors determining the characteristics
4:21:57to be measured are customer purchase
4:21:59order requirements or industry standards
4:22:02in the following screen we will look at
4:22:04identifying specifications or tolerances
4:22:07in process capability
4:22:10the process specification or tolerances
4:22:13are defined either by industry standards
4:22:15based on customer requirements or by the
4:22:18organization's engineering department in
4:22:20consultation with the customer
4:22:22a stability study followed by a
4:22:24comprehensive capability study also
4:22:27helps in identifying if the process mean
4:22:29meets the target or the customer mean
4:22:32the process capability study indicates
4:22:34whether the process is capable it is
4:22:37used to determine if the output
4:22:39consistently meets specifications and
4:22:41the probability of a defect or defective
4:22:45this information is used to evaluate and
4:22:47improve the process to meet the
4:22:49tolerance requirements
4:22:51in the following screen we will learn
4:22:53about process performance indices
4:22:57process performance is defined as a
4:22:59statistical measurement of the outcome
4:23:01of a process characteristic which may or
4:23:04may not have been demonstrated to be in
4:23:06a state of statistical control
4:23:08in other words it is an estimate of the
4:23:11process capability of a process during
4:23:13its initial setup before it has been
4:23:15brought into a state of statistical
4:23:17control
4:23:18it differs from the process capability
4:23:21in that for process performance a state
4:23:23of statistical control is not required
4:23:26the three basic process performance
4:23:28indices are process performance or pp
4:23:32process performance index or ppk
4:23:35and process capability index denoted as
4:23:39ppm or cpm click each index to know more
4:23:45pp stands for process performance it is
4:23:48computed by subtracting the lower
4:23:50specification limit from the upper
4:23:52specification limit the whole divided by
4:23:55natural process variation or six sigma
4:24:00ppk is the process performance index and
4:24:04a minimum of the values of the upper and
4:24:06lower process capability indices
4:24:09the upper and lower process capability
4:24:11indices are calculated as shown on the
4:24:14screen
4:24:15ppu or upper process capability index is
4:24:19given by the formula usl minus x divided
4:24:22by 3s
4:24:24ppl or lower process capability index is
4:24:28given by x minus lsl divided by 3s
4:24:32here x is process average better known
4:24:35as x bar and s is sample standard
4:24:38deviation
4:24:41cpm denotes the process capability index
4:24:45mean which accounts for the location of
4:24:47the process average relative to a target
4:24:50value
4:24:51it can be calculated as shown on the
4:24:53screen
4:24:54here myu stands for process average
4:24:58sigma symbol denotes the process
4:25:00standard deviation
4:25:01usl is the upper specification limit and
4:25:05lsl is the lower specification limit
4:25:08t is the target value which is typically
4:25:11the center of the tolerance
4:25:13x i is the sample reading and n is the
4:25:16number of sample readings
4:25:19we will look at the key terms in process
4:25:21capability in this screen
4:25:24zst or short term capability is the
4:25:27potential performance of the process in
4:25:29control at any given point of time
4:25:32it is based on the sample collected in
4:25:34the short term
4:25:36the long term performance is denoted by
4:25:38zlt
4:25:39it is the actual performance of the
4:25:41process over a given period of time
4:25:44subgroups are several small samples
4:25:46collected consecutively
4:25:48each sample forms a subgroup
4:25:51the subgroups are chosen so that the
4:25:52data points are likely to be identical
4:25:54within the subgroup but different
4:25:57between two subgroups
4:25:59the process shift is calculated by
4:26:01subtracting the long-term capability
4:26:03from the short-term capability
4:26:06the process shift also reflects how well
4:26:08a process is controlled
4:26:10it is usually a factor of 1.5
4:26:13let us look at short-term and long-term
4:26:15process capability in the next screen
4:26:19the concept of short-term and long-term
4:26:21process shift is explained graphically
4:26:23on this screen
4:26:25there are three different samples taken
4:26:26at time 1 time 2 and time 3. the smaller
4:26:30waveforms represent the short-term
4:26:32capability and they are joined with
4:26:35their means to show the shift in
4:26:36long-term performance
4:26:38the long-term performance curve is shown
4:26:40below with the target value marked in
4:26:43the center
4:26:44it is important to note that over a
4:26:46period of time or subgroups a typical
4:26:49process will shift by approximately 1.5
4:26:52times the standard deviation
4:26:54also long-term variation is more than
4:26:57short-term variation
4:26:59this difference is known as the sigma
4:27:01shift and is an indicator of the process
4:27:04control
4:27:05the reasons for a process shift include
4:27:08changes in operators raw material used
4:27:11wear and tear and time periods
4:27:14we will discuss the assumptions and
4:27:16conventions of process variations in the
4:27:18following screen
4:27:20long-term variation is always longer
4:27:22than short-term variation
4:27:24click each term to know more
4:27:27short-term variations are due to the
4:27:29common causes the variance is inherent
4:27:32in the process and known as the natural
4:27:34variation
4:27:36short-term variations show variation
4:27:38within subgroup and are therefore called
4:27:40within subgroup variation
4:27:43they are usually a small number of
4:27:44samples collected at short intervals
4:27:47in short-term variation the variation
4:27:49due to common causes are captured
4:27:51however common causes are difficult to
4:27:54identify and correct
4:27:56the process may have to be redesigned to
4:27:58remove common causes of variation
4:28:02long-term variations are due to common
4:28:05as well as special causes the added
4:28:07variation or abnormal variation is due
4:28:10to factors external to the usual process
4:28:13long-term variation is also known as the
4:28:16overall variation and is a sample
4:28:18standard deviation for all the samples
4:28:21put together
4:28:22long-term variation shows variations
4:28:24within the subgroup and between
4:28:26subgroups
4:28:28special causes increasing variation
4:28:30include changes in operators raw
4:28:33material and wear and tear
4:28:35the special causes need to be identified
4:28:37and corrected for process improvement
4:28:41this screen explains how the factors of
4:28:43stability capability spread and defect
4:28:46summary are used to interpret the
4:28:48process condition this table gives the
4:28:51process condition for different levels
4:28:53or types of variation with reference to
4:28:55common causes and special causes
4:28:58the table is read as follows
4:29:01in the first scenario the process has
4:29:03lesser common causes of variation or ccv
4:29:07and no special causes of variation or
4:29:10scv
4:29:11in this case the variability is less the
4:29:14capability is high the possibility of
4:29:17defects is less and the process is said
4:29:20to be capable and in control
4:29:23next if the process has lesser ccv and
4:29:26some scb are present then it has high
4:29:30variability low capability and a high
4:29:33possibility of defects
4:29:35the process is said to be out of control
4:29:38and incapable
4:29:40the third possibility is that the
4:29:42process has high ccv and no scv
4:29:47in this case the variability is moderate
4:29:49to high the capability is very low
4:29:53and possibility of defects is very high
4:29:57although the process is in control it is
4:29:59incapable
4:30:01finally at the other extreme is the
4:30:04situation where the process has high ccv
4:30:07and scv is also present
4:30:10here the process has high variability
4:30:13low capability high possibility of
4:30:16defects and is out of control and
4:30:19incapable
4:30:20this table is a quick reference to
4:30:22understand process conditions
4:30:25in the next screen we will compare the
4:30:27cpk and cp values
4:30:30when cpk and cp values are compared
4:30:33three outcomes are possible
4:30:35when cpk is lesser than cp it can be
4:30:38inferred that the mean is not centered
4:30:41when cpk is equal to cp the inference is
4:30:44that the process is accurate
4:30:47the process is considered capable if cpk
4:30:50is greater than one
4:30:51this will happen only if variations are
4:30:53in control
4:30:55cpk can never be greater than cp
4:30:58if this situation occurs the
4:31:00calculations have to be rechecked
4:31:03we will look at an example problem for
4:31:05calculating process variation in the
4:31:07following screen
4:31:09the table on this screen shows data for
4:31:11customer complaint resolution time over
4:31:14a period of three weeks
4:31:16each week's data forms a sub group
4:31:18for example the resolution time is 48
4:31:21hours for a particular case in week 1.
4:31:24in week 2 the case takes up 50 hours and
4:31:27in week 3 it takes about 49 hours
4:31:30the subgroup size is 10.
4:31:32let us understand how to calculate
4:31:34long-term and short-term standard
4:31:36deviations are calculated for this data
4:31:40the average for each week is calculated
4:31:42by dividing the total number of
4:31:44complaints resolved by the subgroup size
4:31:47a grand average is also calculated for
4:31:50all the three weeks
4:31:51the variations within subgroups and
4:31:54between subgroups for each week are
4:31:56calculated
4:31:57this is followed by calculating the
4:31:59total variations within and between
4:32:01subgroups
4:32:03overall variation is given by the sum of
4:32:05total variation within subgroups and
4:32:08total variation between subgroups
4:32:11finally the standard deviations for the
4:32:12short term and the long term are
4:32:14calculated using the formula given on
4:32:16the screen
4:32:18the results for the process variation
4:32:20calculations are as follows
4:32:23the grand average for all three weeks is
4:32:2547.5
4:32:27the total variation within subgroups is
4:32:301023.8
4:32:33the total variation between subgroups is
4:32:36161.67
4:32:39both these variations are added to give
4:32:41the overall variation of 1185.5
4:32:47the short-term standard deviation is 6.2
4:32:50and the long-term standard deviation is
4:32:536.4
4:32:55note that the overall variation can also
4:32:57be calculated with the usual sample
4:32:59variance formula
4:33:02let us discuss the effective mean shift
4:33:04on the process capability in this screen
4:33:07the table given here shows the defect
4:33:10level at different sigma multiple values
4:33:12and different mean shifts
4:33:14from the table it can be seen that when
4:33:16the mean is centered within the
4:33:17specification limits and the process
4:33:20capability is 1 that is plus or minus 3s
4:33:24fits within the specification limits
4:33:27the dpmo is 2700 percent and the
4:33:30probability of a good result is
4:33:3399.73 percent
4:33:35if the mean shifts by 1.5 sigma then a
4:33:38tail moves outside the specification
4:33:40limit to a greater extent
4:33:42now the dpmo increases to over sixty six
4:33:45thousand this is almost a twenty five
4:33:48hundred percent increase in defects
4:33:50if the process has a process capability
4:33:53of two that is plus or minus six s fits
4:33:56within the specification limits and the
4:33:58mean shifts by 1.5 sigma then the
4:34:01probability of a good result is
4:34:0499.99966 percent
4:34:06this is the same as a process with a
4:34:08capability of 1.5 that is plus or minus
4:34:124.5 s fitting within the specification
4:34:15limits and no shift in the mean
4:34:17the long-term and short-term capability
4:34:19table shows the variations in
4:34:21capabilities
4:34:23for the purposes of six sigma the
4:34:25assumption is that the long-term
4:34:27variability will have a 1.5 s difference
4:34:30from the short-term variability
4:34:32as seen in statistical process control
4:34:35this assumption can be challenged if
4:34:37control charts are used and these kinds
4:34:40of shifts are detected quickly
4:34:42in the chart it can be seen that the
4:34:44mean shift is negligible as the process
4:34:47capability increases
4:34:49therefore for a six sigma process the
4:34:51long-term variation does not have much
4:34:53effect
4:34:55in the next screen we will look at key
4:34:57concepts in process capability for
4:34:59attribute data
4:35:01the customary procedure for defining
4:35:03process capability for attribute data is
4:35:06to define the mean rate of
4:35:07non-conformity
4:35:09defects and defectives are examples of
4:35:11non-conformity
4:35:12defects per million opportunities or
4:35:14dpmo is the measure of process
4:35:17capability for attribute data
4:35:20for this the mean and the standard
4:35:22deviation for attribute data have to be
4:35:24defined
4:35:25for defectives p bar is used for
4:35:27checking process capability for both
4:35:29constant and variable sample sizes
4:35:32for defects c bar and u bar are used for
4:35:35constant and variable sample sizes
4:35:38respectively
4:35:39the p bar c bar and u bar are the
4:35:42equivalent of the standard deviation
4:35:44denoted by sigma for continuous data
4:35:47in this topic we will learn about the
4:35:49patterns of variation in detail
4:35:52let us start with the classes of
4:35:54distributions in the following screen
4:35:57when data obtained from the measurement
4:35:59phase is plotted on a chart it is
4:36:02observed that it exhibits a variety of
4:36:04distributions depending on the data type
4:36:07and its source
4:36:09these distribution patterns will help
4:36:11you understand the data better
4:36:14probability statistics and inferential
4:36:17statistics
4:36:18are the methods used to describe the
4:36:20parameters for the classes of
4:36:22distributions
4:36:23click each method to know more
4:36:26probability is based on the assumed
4:36:29model of distribution and it is used to
4:36:32find the chances of a certain outcome or
4:36:34event to occur
4:36:37statistics uses the measured data to
4:36:40determine a model to describe the data
4:36:42used
4:36:44inferential statistics describe the
4:36:46population parameters based on the
4:36:48sample data using a particular model
4:36:52in this screen we will discuss the types
4:36:55of distributions there are two types of
4:36:58distributions discrete distribution and
4:37:01continuous distribution
4:37:04discrete distribution includes binomial
4:37:07distribution and poisson distribution
4:37:10continuous distribution includes normal
4:37:12distribution chi square distribution
4:37:16t distribution and f distribution let us
4:37:19learn about discrete probability
4:37:21distribution in the following screen
4:37:24discrete probability distribution is
4:37:27characterized by the probability mass
4:37:29function
4:37:30it is important to be familiar with
4:37:32discrete distributions while dealing
4:37:34with discrete data some of the examples
4:37:37of discrete probability distribution are
4:37:40binomial distribution poisson
4:37:42distribution negative binomial
4:37:45distribution geometric distribution and
4:37:48hyper geometric distribution
4:37:51we will focus only on the two most
4:37:53useful discrete distributions binomial
4:37:56distribution and poisson distribution
4:38:00like most probability distributions
4:38:02these distributions also help in
4:38:04predicting the sample behavior that has
4:38:07been observed in a population let us
4:38:10learn about binomial distribution in the
4:38:13following screen
4:38:16binomial distribution is a probability
4:38:19distribution for discrete data
4:38:21named after the swiss mathematician
4:38:24jacob bernoulli it is an application of
4:38:26popular knowledge to predict the sample
4:38:29behavior
4:38:30binomial distribution also describes the
4:38:33discrete data as a result of a
4:38:35particular process like the tossing of a
4:38:38coin for a fixed number of times and the
4:38:41success or failure in an interview
4:38:44a process is known as bernoulli's
4:38:46process when the process output has only
4:38:49two possible values like defective or
4:38:52okay pass or fail and yes or no
4:38:56binomial distribution is used to deal
4:38:59with defective items defect is any
4:39:02non-compliance with a specification
4:39:04defective is a product or service with
4:39:07at least one defect
4:39:09binomial distribution is most suitable
4:39:11when the sample size is less than 30 and
4:39:14less than 10 percent of the population
4:39:17it is the percentage of non-defective
4:39:19items provided the probability of
4:39:21creating a defective item remains the
4:39:24same over a period
4:39:26let us look at the equation
4:39:28the probability of exactly r successes
4:39:31out of a sample size of n is denoted by
4:39:35p of r
4:39:36which is equal to ncr whole multiplied
4:39:40by p to the power of r and one minus p
4:39:43whole to the power of n minus r
4:39:46in the equation b is the probability of
4:39:49success
4:39:50r is the number of successes desired and
4:39:53n is the sample size
4:39:55to continue discussing the binomial
4:39:58distribution let us look at some of its
4:40:00key calculations in the following screen
4:40:03the mean of a binomial distribution is
4:40:06denoted by mio and is given by n
4:40:09multiplied by p
4:40:10the standard deviation of a binomial
4:40:13distribution is denoted by sigma which
4:40:16is equal to n multiplied by p multiplied
4:40:19by one minus p
4:40:21the method of calculating factorials say
4:40:24factorial of five
4:40:26is the product of five four three two
4:40:29and one
4:40:30which is equal to one hundred twenty
4:40:33similarly factorial of four is the
4:40:36product of four three two and one
4:40:39which is equal to twenty four
4:40:42let us look at an example of calculating
4:40:44binomial distribution in the next screen
4:40:49suppose you wish to know the probability
4:40:51of getting heads five times in eight
4:40:54coin tosses
4:40:55you can use the binomial equation for
4:40:57the same
4:40:58click the answer button to see how this
4:41:00is done
4:41:02the tossing of a coin has only two
4:41:05outcomes heads and tails
4:41:07it means that the probability of each
4:41:09outcome is 0.5 and it remains fixed over
4:41:13a period of time
4:41:14additionally the outcomes are
4:41:16statistically independent
4:41:18in this case the probability of success
4:41:21denoted by p is 0.5 the number of
4:41:24successes desired is denoted by r which
4:41:27is five and the sample size is denoted
4:41:30by n which is eight
4:41:32therefore the probability of five heads
4:41:35is equal to factorial of eight c r which
4:41:38is eight divided by factorial of five
4:41:41and factorial of eight minus five whole
4:41:44multiplied by zero point five to the
4:41:46power of five multiplied by one minus
4:41:48zero point five whole to the power of
4:41:51eight minus five
4:41:52this calculation gives a result of zero
4:41:55point two one eight seven
4:41:57which is equal to twenty one point eight
4:42:00seven percent
4:42:03let us learn about poisson distribution
4:42:05in this screen
4:42:07fuson distribution is named after simeon
4:42:09denis posson and is also used for
4:42:12discrete data
4:42:13poisson distribution is an application
4:42:16of the population knowledge to predict
4:42:18the sample behavior
4:42:20it is generally used for describing the
4:42:22probability distribution of an event
4:42:24with respect to time or space
4:42:27some of the characteristics of poisson
4:42:29distribution are as follows
4:42:31poisson distribution describes the
4:42:33discrete data resulting from a process
4:42:35like the number of calls received by a
4:42:37call center agent or the number of
4:42:40accidents at a signal
4:42:42unlike binomial distribution which deals
4:42:44with binary discrete data poisson
4:42:47distribution deals with integers which
4:42:49can take any value poisson distribution
4:42:52is suitable for analyzing situations
4:42:54wherein the number of trials similar to
4:42:56the sample size in binomial distribution
4:42:59is large and tends towards infinity
4:43:03additionally it is used in situations
4:43:05where the probability of success in each
4:43:08trial is very small almost tending
4:43:10towards zero
4:43:12this is the reason why poisson
4:43:13distribution is applicable for
4:43:15predicting the occurrence of rare events
4:43:17like plane crashes car accidents etc and
4:43:21is therefore widely used in the
4:43:23insurance sector poisson distribution
4:43:25can be used for predicting the number of
4:43:27defects as well given a low defect
4:43:30occurrence rate
4:43:31let us look at the formula for
4:43:33calculating poisson distribution in the
4:43:35next screen the poisson distribution for
4:43:37a probability of exactly x occurrences
4:43:41is given by p of x equals to lambda to
4:43:44the power of x multiplied with log e to
4:43:46the power of minus lambda whole divided
4:43:48by factorial of x
4:43:51in this equation lambda is the mean
4:43:53number of occurrences during the
4:43:55interval
4:43:56x is the number of occurrences desired
4:43:59and e is the base of natural logarithm
4:44:02which is equal to 2.71828
4:44:05the mean of the poisson distribution is
4:44:08given by the lambda and the standard
4:44:10deviation of a poisson distribution is
4:44:12given by sigma which is the square root
4:44:15of lambda
4:44:16let us look at an example to calculate
4:44:19poisson distribution in the next screen
4:44:22the past records of a road junction
4:44:24which is accident prone show a mean
4:44:27number of five accidents per week at
4:44:29this junction
4:44:30assume that the number of accidents
4:44:32follows a poisson distribution and
4:44:35calculate the probability of any number
4:44:37of accidents happening in a week
4:44:39click the button to know the answer
4:44:43given the situation you know that the
4:44:45value of lambda or mean is five
4:44:48so p of zero that is the probability of
4:44:50zero accidents per week is calculated as
4:44:53five to the power of 0 multiplied by e
4:44:56to the power of minus 5 whole divided by
4:44:59factorial of zero
4:45:01the answer is
4:45:030.006
4:45:05applying the same formula the
4:45:07probability of one accident per week is
4:45:090.03
4:45:11the probability of more than two
4:45:13accidents per week is one minus the sum
4:45:16of probabilities of zero one and two
4:45:19accidents
4:45:20which is 0.884
4:45:23in other words the probability is 88.4
4:45:26percent
4:45:29let us learn about continuous
4:45:31probability distribution in this screen
4:45:34continuous probability distribution is
4:45:37characterized by the probability density
4:45:40function
4:45:41a variable is said to be continuous if
4:45:44the range of possible values falls along
4:45:46a continuum
4:45:48for example loudness of cheering at a
4:45:50ball game weight of cookies in a package
4:45:54length of a pen
4:45:55or the time required to assemble a car
4:45:59continuous probability distributions
4:46:01help in predicting the sample behavior
4:46:04observed in a population
4:46:07let us learn about normal distribution
4:46:09in this screen
4:46:11the normal or gaussian distribution is a
4:46:13continuous probability distribution
4:46:16the normal distribution is represented
4:46:18as n and depends on two factors
4:46:21me u which stands for mean and sigma
4:46:24which gives the standard deviation of
4:46:26the data points
4:46:28normal distribution normally has a
4:46:30higher frequency of values around the
4:46:32mean and lesser occurrences away from it
4:46:36it is often used as a first
4:46:38approximation to describe real valued
4:46:40random variables that tend to cluster
4:46:42around a single mean value
4:46:45the distribution is bell shaped and
4:46:47symmetrical the total area under the
4:46:49normal curve is 1 which is p of x
4:46:53various types of data such as body
4:46:55weight height the output of a
4:46:56manufacturing device etc follow the
4:46:59normal distribution
4:47:01additionally normal distribution is
4:47:03continuous and symmetrical with the
4:47:05tails asymptotic to the x-axis which
4:47:08means they touch the x-axis at infinity
4:47:11let us continue to discuss normal
4:47:13distribution in the following screen
4:47:17in a normal distribution to standardize
4:47:19comparisons of dispersion or the
4:47:21different measurement units like inches
4:47:23meters grams etc
4:47:25a standard z variable is used the uses
4:47:28of z value are as follows
4:47:31while the value of z or the number of
4:47:34standard deviations is unique for each
4:47:36probability within the normal
4:47:38distribution
4:47:39it helps in finding probabilities of
4:47:41data points anywhere within the
4:47:43distribution
4:47:44it is dimensionless as well that is it
4:47:47has no units such as millimeters liters
4:47:50coulombs etc
4:47:52there are different formulas to arrive
4:47:54at the normal distribution
4:47:56we will focus on one commonly used
4:47:58formula for calculating normal
4:48:00distribution which is
4:48:02z equals y minus mu whole divided by
4:48:05sigma
4:48:06here z is the number of standard
4:48:08deviations between y and the mean
4:48:11denoted by miu
4:48:13why is the value of the data point in
4:48:15concern
4:48:16me u is mean of the population or data
4:48:18points and sigma is the standard
4:48:21deviation of the population or data
4:48:23points
4:48:24let us look at an example for
4:48:26calculating normal distribution in the
4:48:28following screen
4:48:31suppose the time taken to resolve
4:48:33customer problems follows a normal
4:48:35distribution with a mean of 250 hours
4:48:38and standard deviation of twenty three
4:48:40hours
4:48:41find the probability of a problem
4:48:43resolution taking more than three
4:48:45hundred hours
4:48:46click the button to know the answer
4:48:50in this case y is equal to three hundred
4:48:52me u equals 250 and sigma equals 23.
4:48:57applying the normal distribution formula
4:48:59z is equal to 300 minus 250 whole
4:49:03divided by 23. the result is 2.17
4:49:08when you look at the normal distribution
4:49:10table the z value of 2.17 covers an area
4:49:14of
4:49:150.98499 under itself
4:49:18this means the probability of a problem
4:49:21taking 0 to 300 hours to be resolved is
4:49:2498.5 percent and therefore the chances
4:49:27of a problem resolution taking more than
4:49:30300 hours is 1.5 percent
4:49:34let us understand the usage of z table
4:49:36in this screen
4:49:38the graphical representation of z table
4:49:40usage is given here
4:49:43the probability of areas under the curve
4:49:45is one
4:49:46for the actual value one can identify
4:49:49the z score by using the z table
4:49:52as shown this probability is the area
4:49:55under the curve to the left of point
4:49:57plus a to zero
4:50:00using the actual data when you calculate
4:50:02mean and standard deviation and the
4:50:05values are 25 and 5 respectively it is
4:50:08the normal distribution
4:50:10if the same data is standardized to a
4:50:13mean value of 0 and standard deviation
4:50:16value of 1
4:50:17it is the standard normal distribution
4:50:20in the next screen we will take a look
4:50:22at the z table
4:50:25the z table gives the probability that z
4:50:28is between zero and a positive number
4:50:31there are different forms of normal
4:50:33distribution z tables followed globally
4:50:36the most common form of z table with
4:50:38positive z scores is shown here
4:50:41the value of a called the percentage
4:50:44point is given along the borders of the
4:50:46table in bold and is to two decimal
4:50:49places
4:50:51the values in the main table are the
4:50:53probabilities that z is between zero and
4:50:56plus a
4:50:58note that the values running down the
4:50:59table are to one decimal place
4:51:02the numbers along the column change only
4:51:04for the second decimal place
4:51:07let us look at some examples on how to
4:51:09use a z table in the following screen
4:51:13let us find the value of p of z less
4:51:15than zero
4:51:17the table is not needed to find the
4:51:19answer once we know that the variable z
4:51:22takes a value less than or equal to zero
4:51:26first the area under the curve is one
4:51:28and second the curve is symmetrical
4:51:31about z equals zero
4:51:33hence there is zero point five or fifty
4:51:35percent above chance of z equals zero
4:51:38and zero point five or fifty percent
4:51:41below chance of z equals zero
4:51:46let us find the value of p of z greater
4:51:48than one point one two
4:51:51in this case the chance of z is greater
4:51:53than a number in this case 1.12
4:51:57you can find this by using the following
4:51:59fact
4:52:00the opposite or complement of an event
4:52:03of a is the event of not a
4:52:06that is the opposite or complement of
4:52:08event a occurring is the event a not
4:52:11occurring
4:52:12its probability is given by p of not a
4:52:15equals one minus p of a
4:52:18in other words p of z greater than one
4:52:20point one two is one minus the opposite
4:52:23which is p of z lesser than 1.12
4:52:28using the table p of z less than 1.12
4:52:31equals 0.5 plus p of 0 less than z less
4:52:36than 1.12 equals 0.5 plus
4:52:400.3686 which is
4:52:430.8686
4:52:45hence the answer is p of z greater than
4:52:481.12 equals 1 minus 0.86
4:52:53which is 0.1314
4:52:56note the answer is less than 0.5
4:53:00let us find the value of p of z lies
4:53:03between 0 and 1.12
4:53:06in this case where z falls within an
4:53:08interval the probability can be read
4:53:10straight off the table
4:53:12p of z lies between 0 and 1.12 equals
4:53:160.3686
4:53:19we will learn about chi-square
4:53:21distribution in this screen
4:53:23chi-square distribution is also known as
4:53:26chi squared or chi-square distribution
4:53:30chi squared with k-1 degrees of freedom
4:53:33is the distribution of a sum of the
4:53:35squares of k independent standard normal
4:53:38random variables
4:53:40the chi-squared distribution is one of
4:53:42the most widely used probability
4:53:44distributions in inferential statistics
4:53:47it is also known as hypothesis testing
4:53:51and the distribution is used in
4:53:52hypothesis tests
4:53:54when used in hypothesis tests it only
4:53:57needs one sample for the test to be
4:53:59conducted
4:54:01degree of freedom is k1 where k is the
4:54:05sample size
4:54:06click the button to view the chi-square
4:54:09distribution formula
4:54:12chi-square calculated or sigma or the
4:54:15chi-square index equals f of o minus f
4:54:18of e
4:54:19the whole square divided by f of e
4:54:22here f of o stands for an observed
4:54:25frequency and f of e stands for an
4:54:28expected frequency determined through a
4:54:31contingency table
4:54:33we will learn about the chi-square
4:54:34distribution in detail in the later part
4:54:37of this lesson
4:54:39let us proceed to the next screen to
4:54:41discuss t distribution
4:54:45the t distribution method is the most
4:54:47appropriate method to be used in the
4:54:49following situations
4:54:51when you have a sample size of less than
4:54:5430 when the population standard
4:54:56deviation is not known
4:54:58when the population is approximately
4:55:00normal
4:55:01unlike the normal distribution a t
4:55:04distribution is lower at the mean and
4:55:06higher at the tails as seen in the image
4:55:10t distribution is used for hypothesis
4:55:12testing
4:55:14also as seen in the image the t
4:55:16distribution is symmetrical in shape but
4:55:19flatter than the normal distribution
4:55:22as the sample size increases the t
4:55:25distribution approaches normality
4:55:27for every possible sample size or
4:55:30degrees of freedom there is a different
4:55:33t distribution
4:55:35let us learn about f distribution in the
4:55:37following screen
4:55:39the f distribution is a ratio of two
4:55:42chi-square distributions
4:55:44a specific f distribution is denoted by
4:55:47the ratio of the degrees of freedom for
4:55:50the numerator chi square
4:55:52and the degrees of freedom for the
4:55:54denominator chi-square
4:55:57the f-test is performed to calculate and
4:55:59observe if the standard deviations or
4:56:02variances of two processes are
4:56:04significantly different
4:56:07the project teams are usually concerned
4:56:09about reducing the process variance
4:56:12as per the formula f calculated equals
4:56:15s1 square divided by s2 square
4:56:19where s1 and s2 are the standard
4:56:21deviations of the two samples
4:56:24if the f calculated is 1
4:56:27it implies there is no difference in the
4:56:29variance
4:56:31if s 1 is greater than s 2
4:56:34then the numerator must be greater than
4:56:36the denominator
4:56:37in other words df1 equals n1 minus 1 and
4:56:42df2 equals n2 minus 1.
4:56:46from the f distribution table you can
4:56:49easily find out the critical f
4:56:51distribution at alpha and the degrees of
4:56:54freedom of the samples of two different
4:56:56processes df1 and df2
4:57:00let us proceed to the next topic of this
4:57:02lesson in the following screen
4:57:05in this topic we will discuss
4:57:07exploratory data analysis in detail
4:57:11let us learn about multivariate studies
4:57:13in the following screen
4:57:16multivariate studies or multi-variable
4:57:19studies are used to analyze variation in
4:57:21a process
4:57:23analyzing variation helps in
4:57:25investigating the stability of a process
4:57:29more stable the process less is the
4:57:31variation
4:57:32multivariate studies also help in
4:57:34identifying areas to be investigated
4:57:37finally they help in breaking down the
4:57:39variation into components to make the
4:57:42required improvements
4:57:44multivariate studies classify variation
4:57:46sources into three major types
4:57:49positional cyclical and temporal
4:57:52click each type to learn more
4:57:55positional variation occurs within a
4:57:58single piece or a product
4:58:00variation in pieces of a batch is also
4:58:03an example of positional variation
4:58:06in positional variation measurements at
4:58:08different locations of a piece would
4:58:10produce different values
4:58:13suppose a company is manufacturing a
4:58:15metal plate of thickness one inch and
4:58:18the plate thickness is different at many
4:58:20points
4:58:21it is an example of positional variation
4:58:24some of the other examples can be pallet
4:58:27stacking in a truck
4:58:28temperature gradient in an oven
4:58:31variation observed from cavity to cavity
4:58:33within a mold
4:58:35region of a country and line on invoice
4:58:40cyclical variation occurs when
4:58:43measurement differs from piece to piece
4:58:45or product to product but over a short
4:58:47period of time
4:58:49time is important because certain
4:58:51measurements may change if a product
4:58:53such as a hot metal sheet is measured
4:58:55after a long interval
4:58:57if the measurement at the same location
4:58:59in a piece varies with different pieces
4:59:02it is an example of cyclical variation
4:59:05other examples of cyclical variations
4:59:08are match to batch variation
4:59:10lot to lot variation and account
4:59:13activity week to week
4:59:16temporal variation occurs over a longer
4:59:19period of time
4:59:20such as machine wear and tear and
4:59:22changes in efficiency of an operator
4:59:24before and after lunch
4:59:27temporal variations may also be seasonal
4:59:30if the range of positional variation in
4:59:32a piece is more in winter than in summer
4:59:35it is an example of temporal variation
4:59:38the variation may occur because of
4:59:40unfavorable working conditions in winter
4:59:43process drift performance before and
4:59:45after breaks seasonal and shift based
4:59:47differences month-to-month closings and
4:59:50quarterly returns can be examples of
4:59:53temporal variation
4:59:56we will learn about creating a
4:59:57multivariate chart in this screen
5:00:00the outcome of multivariate studies is
5:00:03the multivariate chart
5:00:04it depicts the type of variation in the
5:00:07product and helps in identifying the
5:00:09root cause
5:00:11there are five major steps involved in
5:00:13creating a multivariate chart
5:00:16they are select process and
5:00:17characteristics
5:00:19decide sample size
5:00:21create a tabulation sheet
5:00:24plot the chart and link the observed
5:00:27values
5:00:28click each step to learn more
5:00:32the first step is to select the process
5:00:34and the relevant characteristics to be
5:00:36investigated
5:00:38for example selecting the process where
5:00:40the plate of one inch thickness is being
5:00:42manufactured
5:00:44in this process four equipment numbered
5:00:46one to four produce the one-inch plates
5:00:49the characteristic to be measured is the
5:00:51thickness of the plate ranging from 0.95
5:00:55inch to 1.05 inches
5:00:58any plate thickness outside this range
5:01:00is a defect
5:01:04the second step is to decide the sample
5:01:06size and the frequency of data
5:01:08collection
5:01:10in this example the sample size is five
5:01:13pieces per equipment and the frequency
5:01:15of collecting data is every two hours
5:01:18starting from eight in the morning to
5:01:20two in the afternoon
5:01:22then the tabulation sheet is created
5:01:24where the data will be recorded
5:01:27so one should measure the thickness of
5:01:28the plate being produced by the four
5:01:30equipment at a data collection frequency
5:01:33of two hours
5:01:36the third step is to create a tabulation
5:01:39sheet
5:01:40in this example the tabulation sheet
5:01:42with data records contains the columns
5:01:45with time equipment number and thickness
5:01:48as headers
5:01:51the fourth step is to plot the chart
5:01:54in this example a chart can be plotted
5:01:57with time on the x-axis and plate
5:02:00thickness on the y-axis
5:02:04the last step is to link the observed
5:02:06values
5:02:07in this example the observed values can
5:02:09be linked by appropriate lines
5:02:12we will continue to learn about creating
5:02:15a multivariate chart in this screen
5:02:18the path to create a multivariate chart
5:02:20in minitab is by selecting stat then
5:02:23quality tools followed by multivariate
5:02:26chart
5:02:28the multivariate chart created from the
5:02:30data recorded is shown here
5:02:32the upper specification limit of 1.05
5:02:36inches and the lower specification limit
5:02:38of 0.95 inches has been marked by green
5:02:42lines
5:02:43data outside these lines are defects
5:02:47the blue dots show the positional
5:02:49variation
5:02:50the dots are the measurements of pieces
5:02:52in a batch of any single equipment
5:02:55the black lines join the mean of the
5:02:57data recorded from the equipment
5:03:00the mean of the data recorded from the
5:03:02products of equipment number three
5:03:05is much below the similar mean of other
5:03:07equipment
5:03:09this shows that equipment number three
5:03:11is producing more defects than the other
5:03:14equipment
5:03:15the red line is the mean of the data
5:03:17recorded at a particular time
5:03:20the red line rises toward the right
5:03:22which means the data points shift up
5:03:25after 12 pm
5:03:27this may be because of the change in
5:03:29operator efficiency after a lunch break
5:03:32at 12 pm
5:03:34multivariate chart helps us visually
5:03:36depict the variations and establish the
5:03:39root causes of variations
5:03:42in the next screen we will learn about
5:03:44simple linear correlation
5:03:47correlation means association between
5:03:49variables
5:03:51simple linear regression and multiple
5:03:53regression techniques are very important
5:03:56as they help in validating the
5:03:58association of x with y
5:04:01the coefficient correlation shows the
5:04:04strength of the relationship between y
5:04:06and x
5:04:08to associate y with a single x
5:04:11and statistically validate the
5:04:13relationship correlation is used in
5:04:16excel
5:04:17use equals corel open bracket close
5:04:20bracket function to calculate
5:04:22correlation coefficient
5:04:25the dependent variable y may depend on
5:04:28many independent variables x
5:04:31but correlation is used to find the
5:04:33behavior of y as one of the x's changes
5:04:37correlation helps us to predict the
5:04:39direction of movement in values in y
5:04:42when x changes
5:04:44statistical significance of this
5:04:46movement is denoted by correlation
5:04:49coefficient r
5:04:51it is also known as pearson's
5:04:52coefficient of correlation
5:04:55in any correlation the value of the
5:04:57correlation coefficient is always
5:04:59between minus one and plus one
5:05:02positive value of r denotes the
5:05:04direction of movement in both variables
5:05:06is the same
5:05:08as x increases y also increases and vice
5:05:12versa
5:05:13negative value of r denotes that the
5:05:16direction of movement in both variables
5:05:18is in inverse fashion
5:05:20as x increases y decreases and as x
5:05:24decreases y increases
5:05:27when the value of r is zero it means
5:05:30that there is no correlation between the
5:05:32two variables
5:05:34higher the absolute value of r stronger
5:05:37the correlation between y and x
5:05:40absolute value of a number is its value
5:05:43without the sign
5:05:45plus four has an absolute value of four
5:05:48and minus four
5:05:50again has an absolute value of four
5:05:53an r value of greater than plus 0.85
5:05:57or lesser than minus 0.85
5:06:00indicates a strong correlation
5:06:03hence r value of minus 0.95
5:06:06shows a stronger correlation than the r
5:06:08value of minus 0.74
5:06:12the next screen will elaborate on
5:06:14correlation with the help of an example
5:06:16and illustrations through scatter plots
5:06:20the four graphs on the screen are
5:06:22scatter plots displaying four different
5:06:24levels of correlation
5:06:26correlation measures the linear
5:06:28association between the dependent
5:06:30variable or output variable y and one
5:06:34independent or input variable x
5:06:37as can be deduced from the graphs
5:06:40a definite pattern emerges as the
5:06:42absolute value of correlation
5:06:43coefficient r increases
5:06:46it is easy to see a pattern in r value
5:06:49of 0.9 and above than to see a pattern
5:06:52in r value of 0.07
5:06:55it is difficult to find a pattern below
5:06:57correlation coefficient of 0.5 click the
5:07:01example button to know more
5:07:04to understand how correlation helps let
5:07:07us consider an example
5:07:09a correlation test was performed on the
5:07:11scores of a set of students from their
5:07:14first grade high school and under
5:07:16graduation
5:07:17the undergraduation score was the
5:07:19dependent variable and first grade score
5:07:22the independent variable
5:07:24the value of correlation coefficient r
5:07:27was calculated to be 0.29
5:07:30and the correlation between
5:07:31undergraduation scores and high school
5:07:34scores was 0.45
5:07:37this means the high school scores have
5:07:39higher correlation compared to the first
5:07:41grade scores
5:07:43this states the performance of students
5:07:45in high school is a better indicator of
5:07:47their performance in undergraduation
5:07:50than their performance in the first
5:07:51grade
5:07:53although the correlation exists as both
5:07:55the values of r are less than 0.85
5:07:59it will be difficult to draw a straight
5:08:01line on the scanner plot
5:08:04in this screen we will learn about
5:08:06regression although correlation gives
5:08:08the direction of movement of the
5:08:10dependent variable y as independent
5:08:13variable x changes
5:08:15it does not provide the extent of the
5:08:17movement of y as x changes
5:08:20this degree of movement can be
5:08:22calculated using regression
5:08:25if a high percentage of variability in y
5:08:28is explained by changes in x
5:08:31one can use the model to write a
5:08:32transfer equation y is equal to f x and
5:08:36use the same equation to predict future
5:08:39values of y given x and x given y
5:08:43the output of regression on y and x is a
5:08:46transfer function equation that can
5:08:48predict values of y for any other value
5:08:52of x
5:08:53transfer function is generally denoted
5:08:55by f and the equation is written as y
5:08:59equals to f of x
5:09:01y can be regressed on one or more x's
5:09:04simultaneously
5:09:06simple linear regression is for 1x and
5:09:09multiple linear regression is for more
5:09:12than one x
5:09:13the next screen will focus on key
5:09:15concepts of regression
5:09:19there are two key concepts of regression
5:09:22transfer function to control y and vital
5:09:25x click each concept to learn more
5:09:29the output of regression is a transfer
5:09:32function f of x
5:09:34although the transfer function f of x
5:09:37gives the degree of movement in y as x
5:09:40changes
5:09:41it is not the correct transfer function
5:09:43to control y as there may be a low level
5:09:46of correlation between the two
5:09:50the main thrust of regression is to
5:09:52discover whether a significant
5:09:54statistical relationship exists between
5:09:56y and a particular x that is by looking
5:10:00at p values
5:10:02based on regression one can infer the
5:10:04vital x and eliminate the unimportant
5:10:07x's
5:10:08the analyze phase helps in understanding
5:10:11if there is statistical relevance
5:10:13between y and x if the relevance is
5:10:16established using metrics from
5:10:18regression analysis
5:10:20one can move forward with the tests
5:10:22the simple linear regression or slr
5:10:26should be used as a statistical
5:10:28validation tool in the beginning of this
5:10:30phase
5:10:33in this screen we will understand the
5:10:35concept of simple linear regression
5:10:38a simple linear regression equation is a
5:10:41fitted linear equation and is
5:10:44represented by the equation shown here
5:10:47in this equation y is the dependent
5:10:50variable and x is the independent
5:10:52variable
5:10:53a is the intercept of the fitted line on
5:10:56the y axis which is equal to the value
5:10:59of y at x equal to zero
5:11:02b is the regression coefficient or the
5:11:06slope of the line and c is the error in
5:11:09the regression model which has a mean of
5:11:12zero
5:11:13the next screen will focus on the least
5:11:15squares method in simple linear
5:11:17regression
5:11:20with reference to the error mentioned
5:11:22earlier if correlation coefficient of y
5:11:25and x is not equal to one
5:11:27meaning the relation is not perfectly
5:11:30linear there could be several lines that
5:11:32could fit in the scatter plot
5:11:35notice the two graphs displayed
5:11:37for the same set of five points two
5:11:40different types of lines are drawn and
5:11:42both of them have errors
5:11:44error refers to the points on the
5:11:46scatter plot that do not fall on the
5:11:49straight line drawn the second graph
5:11:51shows the y-intercept
5:11:53statistical software like minitab fits
5:11:56the line which has the least value of
5:11:58error squared and added
5:12:01as is clear from the graph error is the
5:12:03distance of the point from the fitted
5:12:05line
5:12:06typically the data lies off the line
5:12:10in perfect linear relation
5:12:12all points would lie on the line and
5:12:14error would be zero
5:12:16the distance from the point to the line
5:12:18is the error distance used in the sse
5:12:21calculations
5:12:23let's understand slr with the help of an
5:12:25example in the next screen
5:12:28consider the following example
5:12:31suppose a farmer wishes to predict the
5:12:33relationship between the amount spent on
5:12:36fertilizers and the annual sales of his
5:12:39crops
5:12:40he collects the data shown here for the
5:12:42last few years and determines his
5:12:45expected revenue if he spends eight
5:12:47dollars annually on fertilizers
5:12:50he has targeted sales of thirty one
5:12:52dollars this year
5:12:54the steps to perform
5:12:56simple linear regression in ms excel are
5:12:59as follows
5:13:01copy the data table on an excel
5:13:03worksheet
5:13:04select all the data from b1 to c6
5:13:08this is assuming the years table appears
5:13:11in cells a1 to a6
5:13:14click insert and choose the plain
5:13:16scatter chart
5:13:18it is titled scatter with only markers
5:13:21the basic scatter chart will appear as
5:13:24shown on the screen
5:13:26right click on the data points in the
5:13:27scatter chart and choose the option add
5:13:30trendline
5:13:32then choose the option linear and select
5:13:35the boxes titled display r squared value
5:13:38and display equation
5:13:41a linear line will appear which is
5:13:43called the best fit line
5:13:45or the least squares line
5:13:48to use the data for regression analysis
5:13:51the interpretation of the scatter chart
5:13:53is as follows
5:13:55the r-square value or the coefficient of
5:13:58determination conveys if the model is
5:14:00good and can be used
5:14:02the r-square value here is 0.3797
5:14:08it means 38 percent of variability in y
5:14:12is explained by x
5:14:14the remaining 62 variation is
5:14:17unexplained or due to residual factors
5:14:21other factors like rain amount and
5:14:23variability sunshine temperatures seed
5:14:27type and seed quality could be tested
5:14:30the low value of r square statistically
5:14:33validates poor relationship between y
5:14:36and x
5:14:37thus the equation presented cannot be
5:14:39used for further analysis
5:14:42in a similar situation one should refer
5:14:45to the cause and effect matrix and study
5:14:47the relationship between y and a
5:14:50different x variable
5:14:53we will discuss multiple linear
5:14:55regression in this screen
5:14:57if a new variable x2 is added to the r
5:15:01square model the impact of x1 and x2 on
5:15:05y gets tested
5:15:07this is known as multiple linear
5:15:09regression
5:15:10the value of r square changes due to the
5:15:13introduction of the new variable
5:15:16the resulting value of r square which
5:15:19can be used in cases of multiple
5:15:21regression is known as r square adjusted
5:15:25the model can be used if r square
5:15:27adjusted value is greater than seventy
5:15:30percent
5:15:31we will look at the key concepts in the
5:15:33next screen
5:15:36the key concepts in multiple linear
5:15:38regression are as follows
5:15:41the residuals or the differences between
5:15:43the actual value and the predicted value
5:15:46give an indication of how good the model
5:15:49is
5:15:50if the errors or residuals are small and
5:15:53predictions use x's that are within the
5:15:55range of the collected data the
5:15:58predictions should be fine
5:16:00the sum of squares total can be
5:16:02calculated as follows
5:16:05sum of squares total or sst
5:16:08equals the sum of squares of regression
5:16:11or ssr plus sum of squares of error or
5:16:15sse
5:16:17to arrive at sum of squares of
5:16:19regression ssr
5:16:21use the formula ssr equals sum of
5:16:25squares total
5:16:26or sst minus sum of squares of error or
5:16:30sse
5:16:32since ssr is sse subtracted from sst
5:16:37value of sse should be less than sst
5:16:43r squared is sum of squares of
5:16:45regression or ssr divided by sum of
5:16:48squares total or sst
5:16:52calculating sst and sse helps in
5:16:56determining ssr and r square
5:17:00to get a sense of the error in the
5:17:02fitted model
5:17:03calculate the value of y for a given
5:17:06data using the fitted line equation
5:17:09to check for error take two observations
5:17:12of y at the same x
5:17:14the most important thing to remember in
5:17:17regression analysis
5:17:18is that the obtained fitted line
5:17:21equation cannot be used to predict y for
5:17:24values of x outside the data
5:17:27for example it would not be possible to
5:17:29predict the amount spent on fertilizers
5:17:32for a forecasted sales of 15 or sixty
5:17:36dollars
5:17:37both data points lie outside the data
5:17:39set on which regression analysis is
5:17:42performed
5:17:43if y is dependent on many x's then
5:17:47simple linear regression analysis can be
5:17:50used to prioritize x
5:17:52but it requires running separate
5:17:54regressions on y with each x
5:17:58if an x does not explain variation in y
5:18:01then it should not be explored any
5:18:03further
5:18:04these were the interpretations of the
5:18:06simple linear regression equation
5:18:09in the next screen we will learn that
5:18:11despite a relationship being established
5:18:14between two variables the change in one
5:18:17may not cause a change in the other
5:18:20let us discuss the difference between
5:18:22correlation and causation in the
5:18:24following screen
5:18:26a regression equation denotes only a
5:18:28relationship between the y and the x
5:18:32this does not mean that a change in one
5:18:34variable will cause a change in the
5:18:36other
5:18:37if number of schools and incidents of
5:18:39crime in a city rise together there may
5:18:42be a relationship but no causation
5:18:45the increase in both the factors could
5:18:47be due to a third factor that is
5:18:50population
5:18:51in other words both of them may be
5:18:53dependent variables to an independent
5:18:56variable
5:18:57consider the graphs shown on the screen
5:19:00the graphs on the left show the
5:19:02relations between number of sneezes and
5:19:05incidence of death with respect to
5:19:07population
5:19:09both have a positive correlation
5:19:11finding a positive correlation between
5:19:14incidents of deaths and number of
5:19:16sneezes does not mean we assume sneezing
5:19:19is the cause of somebody's death
5:19:21despite the correlation being very
5:19:23strong as depicted in the graph on the
5:19:26right
5:19:27let us proceed to the next topic of this
5:19:29lesson in the following screen
5:19:32in this topic we will discuss hypothesis
5:19:35testing in detail
5:19:37let us learn about statistical and
5:19:40practical significance of hypothesis
5:19:42test in the following screen
5:19:45the differences between a variable and
5:19:48its hypothesized value may be
5:19:50statistically significant
5:19:52but may not be practical or economically
5:19:55meaningful
5:19:57for example based on a hypothesis test
5:20:00nutra worldwide inc wants to implement a
5:20:03trading strategy which is proven to
5:20:05provide statistically significant
5:20:07returns
5:20:09however it does not guarantee trading on
5:20:11this strategy would result in
5:20:13economically meaningful positive returns
5:20:17when the logical reasons are examined
5:20:19before implementation the returns are
5:20:22economically significant
5:20:24the returns may not be significant when
5:20:26statistically proven strategy is
5:20:29implemented directly
5:20:31the returns may not be economically
5:20:33significant after accounting for taxes
5:20:36transaction costs and risks inherent in
5:20:39the strategy
5:20:40thus there should be a practical or
5:20:42economic significant study before
5:20:45implementing any statistically
5:20:46significant data
5:20:48the next screen will briefly focus on
5:20:50the conceptual differences between a
5:20:53null and an alternate hypothesis
5:20:57the conceptual differences between a
5:20:59null and an alternate hypothesis are as
5:21:02follows
5:21:04assume the specification of the current
5:21:06process is itself the null hypothesis
5:21:10null hypothesis denoted as the basic
5:21:13assumption for any activity or
5:21:15experiment
5:21:16is represented as h o
5:21:19no hypothesis cannot be proved it can
5:21:22only be rejected or disproved
5:21:25it is important to note that if null
5:21:27hypothesis is rejected
5:21:30alternative hypothesis must be right
5:21:33for example assuming that a movie is
5:21:35good one plans to watch it
5:21:37therefore the null hypothesis in this
5:21:40scenario will be
5:21:41movie is good
5:21:44alternative hypothesis or h a
5:21:46challenges the null hypothesis
5:21:49or is the converse of the null
5:21:51hypothesis
5:21:53in this scenario alternate hypothesis
5:21:55will be movie is not good
5:21:58in the following screen we will discuss
5:22:00type 1 and type 2 error
5:22:04rejecting a null hypothesis when it is
5:22:07true is called type 1 error
5:22:09it is also known as producer's risk
5:22:12for example the rejection of a product
5:22:15by the qa team when it is not defective
5:22:18will cause loss to the producer suppose
5:22:21when a movie is good it is reviewed to
5:22:24be not good this reflects type 1 error
5:22:28in this case the null hypothesis
5:22:30is rejected when it is actually true
5:22:33the two important points to be noted are
5:22:36significance level or alpha is the
5:22:39chance of committing a type 1 error
5:22:42the value of alpha is
5:22:440.05 or 5 percent
5:22:48accepting a null hypothesis when it is
5:22:50false is called type 2 error
5:22:54it is also known as consumer's risk
5:22:57for example the acceptance of a
5:22:59defective product by the quality analyst
5:23:01of an organization will cause loss to
5:23:04the consumer who buys it
5:23:07minimizing type 2 error requires
5:23:09acceptance criteria to be very strict
5:23:12suppose when a movie is not good it is
5:23:15reviewed to be good this reflects type
5:23:18two error
5:23:19in this case the alternate hypothesis is
5:23:22rejected when it was actually true the
5:23:26two important points to be noted are
5:23:28beta is the chance of committing a type
5:23:302 error
5:23:32the value of beta is 0.2 or 20 percent
5:23:36any experiment should have as less beta
5:23:38value as possible
5:23:40the next screen will cover the key
5:23:42points to remember about type 1 and type
5:23:452 error
5:23:48as you start dealing with the two types
5:23:50of errors keep the following points in
5:23:52mind the probability of making one type
5:23:56of error can be reduced when one is
5:23:58willing to accept a higher probability
5:24:01of making the other type of error
5:24:04suppose the management of a company
5:24:05producing pacemakers wants to ensure no
5:24:09defective pacemaker reaches the consumer
5:24:12so the quality assurance team makes
5:24:15stringent guidelines to inspect the
5:24:17pacemakers
5:24:18this would invariably decrease the beta
5:24:21error or type 2 error
5:24:23but this will also increase the chance
5:24:26that a non-defective pacemaker is
5:24:28declared defective by the quality
5:24:31assurance team
5:24:32thus alpha error or type 1 error
5:24:35increases
5:24:37if all null hypotheses are accepted to
5:24:40avoid rejecting true null hypothesis
5:24:43it will lead to type 2 error
5:24:45typically alpha is set at 0.05
5:24:49which means that the risk of committing
5:24:51a type 1 error is 1 out of 20
5:24:54experiments
5:24:55in case of any product the teams must
5:24:58decide what type of error should be less
5:25:01and set the value of alpha and beta
5:25:04accordingly
5:25:05in the next screen we will discuss the
5:25:07power of test
5:25:10the power of a hypothesis test or the
5:25:13power of test is the probability of
5:25:16correctly rejecting the null hypothesis
5:25:18when it is false
5:25:20power of a test is represented by 1
5:25:23minus beta which is also the type 2
5:25:26error
5:25:27the probability of not committing a type
5:25:292 error is called the power of a
5:25:31hypothesis test
5:25:33the power of a test helps in improving
5:25:36the advantage of hypothesis testing
5:25:39the higher the power of a test the
5:25:42better it is for purposes of hypothesis
5:25:44testing given a choice of tests the one
5:25:48with the highest power should be
5:25:50preferred
5:25:51the only way to decrease the probability
5:25:54of a type 2 error given the significance
5:25:57level or probability of type 1 error
5:26:00is to increase the sample size
5:26:03it is important to note that quality
5:26:05inspection is done on sample pieces and
5:26:08not on all the products
5:26:10so beta error is a function of the
5:26:13sample size
5:26:14if the sample size is not appropriate
5:26:17the defects in a product line could
5:26:19easily be missed out giving a wrong
5:26:22perception of the quality of the product
5:26:25this will increase the type ii error
5:26:28to decrease this error the quality
5:26:30assurance team has to increase the
5:26:32sample size
5:26:34in hypothesis testing
5:26:36alpha is called the significance level
5:26:39and one minus alpha is called the
5:26:41confidence level of the test
5:26:43in the next screen we will focus on the
5:26:46determinants of sample size for
5:26:48continuous data
5:26:50the sample size can be calculated by
5:26:53answering three simple questions how
5:26:56much variation is present in the
5:26:58population
5:26:59at what interval does the true
5:27:01population mean need to be estimated
5:27:04and how much representation error is
5:27:06allowed in the sample
5:27:08continuous data is data which can be
5:27:10measured
5:27:12the sample size for continuous data can
5:27:14be determined by the formula shown on
5:27:16the screen
5:27:18we will learn about the standard sample
5:27:20size formula for continuous data in the
5:27:22next screen
5:27:25representation error or alpha error is
5:27:28generally assumed to be 5
5:27:30or 0.05
5:27:33hence the expression of 1 minus alpha
5:27:36divided by 2 amounts to
5:27:380.975 or 97.5 percent
5:27:43looking up the value of z 97.5 from the
5:27:46z table gives the value 1.96
5:27:51the expression reduces to the one shown
5:27:53on screen
5:27:55when alpha is 5
5:27:57z is 1.96
5:27:59to detect a change that is half the
5:28:02standard deviation
5:28:03one needs to get at least 16 data points
5:28:06for the sample
5:28:08click the example tab to view an example
5:28:10of continuous data calculation using
5:28:13standard sample size formula
5:28:17the population standard deviation for
5:28:19the time to resolve customer problems is
5:28:2230 hours
5:28:23what should be the size of a sample that
5:28:26can estimate the average problem
5:28:28resolution time within plus or minus 5
5:28:31hours tolerance with 99 percent
5:28:33confidence
5:28:35to know with 99 confidence that the time
5:28:38to resolve a customer problem ranges
5:28:40between 25 and 35 hours
5:28:43the value of z for 99.5 must be two
5:28:47point five seven five
5:28:49a good result should fall outside the
5:28:51range of zero point five percent which
5:28:54is one in two hundred trials
5:28:56it is expected that one hundred ninety
5:28:58nine out of two hundred trials will
5:29:00confirm a proper conclusion
5:29:03the calculation gives a result of 238.70
5:29:08one cannot have point seven zero of a
5:29:10sample so one needs to round up to the
5:29:13nearest integer
5:29:15if there are 239 samples
5:29:18the significance level is greater than
5:29:200.01
5:29:22which indicates the confidence is less
5:29:24than 99
5:29:26using 239 reduces alpha and increases
5:29:30the confidence level
5:29:32the rounded up value 239 means the
5:29:35expectations are being met for the
5:29:37confidence level of the test
5:29:41we will learn about the standard sample
5:29:43size formula for discrete data in this
5:29:46screen
5:29:47like continuous data one can find out
5:29:49the sample size required while dealing
5:29:51with discrete population
5:29:53if the average population proportion
5:29:56non-defective is p then population
5:29:59standard deviation can be calculated by
5:30:01using the expression shown on the screen
5:30:04the expression for sample size is
5:30:06present
5:30:08it is important to note that in this
5:30:09expression the interval or tolerance is
5:30:12in percentage
5:30:14click the example tab to view an example
5:30:16of discrete data calculation using
5:30:19standard sample size formula
5:30:23the non-defective population proportion
5:30:25for pen manufacturing is eighty percent
5:30:28what should be the sample size to draw a
5:30:31sample that can estimate the proportion
5:30:33of compliant pans within plus or minus
5:30:36five percent with an alpha of five
5:30:38percent consider calculating the sample
5:30:41size for discrete data for which the
5:30:43population proportion non-defective is
5:30:46eighty percent and the tolerance limit
5:30:48is within plus or minus five percent
5:30:51substituting the values it is found the
5:30:54sample size should be 246. in this
5:30:57example to know if the population
5:30:59proportion for good pens is still within
5:31:0175 to 85 percent
5:31:04and to have 95 percent confidence that
5:31:07the sample will allow a good conclusion
5:31:10one needs to inspect
5:31:12245.86 pounds
5:31:140.86 of a pen cannot be inspected so the
5:31:18value is rounded up to maintain the
5:31:20confidence level
5:31:22inspecting 245 or fewer pens reduces the
5:31:26confidence level
5:31:28this means the z value would be lower
5:31:30than 1.96
5:31:32and alpha would be greater than 0.05
5:31:36suppose one is willing to accept a
5:31:38greater range in the estimate
5:31:40the proportion is within 20 percent of
5:31:42the past results and approximately
5:31:45within one standard deviation of the
5:31:47proportion
5:31:48delta changes to 0.20 and the number of
5:31:52needed samples is 15.4 is approximately
5:31:56equal to 16.
5:32:00this screen will focus on the hypothesis
5:32:02testing roadmap
5:32:04though the basic determinants of
5:32:06accepting or rejecting a hypothesis
5:32:08remain the same
5:32:09various tests are used depending on the
5:32:12type of data from the figure shown on
5:32:14the screen you can conclude the type of
5:32:16test to be performed based on the kind
5:32:18of data and values available
5:32:21for discrete data if mean and standard
5:32:23deviation are both known the z-test is
5:32:26used
5:32:27and if mean is known but standard
5:32:29deviation is unknown the t-test is used
5:32:33if the standard deviation is unknown and
5:32:36if the sample size is less than 30
5:32:39it is preferable to use the t-test if
5:32:42variance is known one should go for
5:32:44chi-squared test
5:32:46if mean and standard deviation are known
5:32:48for a set of continuous data it is
5:32:51recommended to go for the z-test
5:32:54for mean comparison of two with standard
5:32:56deviation unknown
5:32:57go for t-test and for mean comparison of
5:33:00many with standard deviation unknown go
5:33:03for f-test
5:33:05also if the variance is known for
5:33:07continuous data go for f-test the next
5:33:10few screens will discuss in detail the
5:33:12tests for mean variance and proportions
5:33:15let us understand hypothesis test for
5:33:18means theoretical through an example in
5:33:20the next screen
5:33:22the examples of hypothesis testing based
5:33:25on the types of data and values
5:33:26available are discussed here
5:33:29the value of alpha can be assumed to be
5:33:315
5:33:32or
5:33:330.05
5:33:35suppose you want to check for the
5:33:36average height of a population
5:33:39north american males are selected as the
5:33:41population here
5:33:43from the population 117 men are gathered
5:33:46as the sample and the readings of their
5:33:48height are taken
5:33:50the null hypothesis is that the average
5:33:53height of north american males is 165
5:33:56centimeters
5:33:57and the alternate hypothesis is that the
5:34:00height is lesser or greater than 165
5:34:04centimeters
5:34:05consider the sample size n as 117 for z
5:34:09test and sample size n is 25 for t test
5:34:14sample average or x bar is
5:34:16164.5 centimeters
5:34:19using the data given let us calculate
5:34:22the z calc value and t calc value
5:34:25the population height is 165 centimeters
5:34:29with a standard deviation of 5.2
5:34:31centimeters and the average height of
5:34:33the sample group is
5:34:35164.5 centimeters the test for
5:34:38significant difference should be
5:34:40conducted
5:34:42first let us compute z calc value using
5:34:45the formula given on the screen
5:34:47hence the z calc is 1.04
5:34:51which is less than 1.96 or t critical
5:34:55therefore the null hypothesis cannot be
5:34:58rejected
5:34:59since z
5:35:000.05 equals 1.96
5:35:03the null hypothesis is not rejected at
5:35:06five percent level of significance
5:35:09the statistical notation is shown on the
5:35:11screen
5:35:12thus a conclusion based on the sample
5:35:14collected is that the average height of
5:35:16north american males is 165 centimeters
5:35:20if the population standard deviation is
5:35:23not known a t-test is used
5:35:26it is similar to the z-test
5:35:28instead of using the population
5:35:30parameter or sigma the sample statistic
5:35:33standard deviation or s is used
5:35:36in this example the s value is 5.0
5:35:40let us now compute t value using the
5:35:42formula given on the screen
5:35:45the statistical notation to reject null
5:35:47hypothesis is shown on the screen
5:35:50the t critical value is 2.064
5:35:53and we know the t calc value is 0.5
5:35:57which is less than 2.064
5:36:00therefore the null hypothesis cannot be
5:36:02rejected at five percent level of
5:36:04significance
5:36:06thus a conclusion based on the sample
5:36:08collected is that the average height of
5:36:10north american males is 165 centimeters
5:36:14the conclusion of not rejecting the null
5:36:17hypothesis
5:36:18is based on the assumption that the 25
5:36:21males are randomly selected from all
5:36:23males in north america
5:36:25null and alternative hypotheses are same
5:36:29for both z-test and t-test
5:36:32in both the examples the null hypothesis
5:36:34is not rejected
5:36:36in the next screen we will understand
5:36:38the hypothesis test for variance with an
5:36:41example
5:36:43in hypothesis test for variance
5:36:45chi-square test is used
5:36:48in the case of a chi-square test the
5:36:50null and alternate hypotheses are
5:36:52defined and the values of chi-square
5:36:55critical and chi-square are calculated
5:36:59to understand this concept with an
5:37:01example click the button given on the
5:37:03screen
5:37:05the null hypothesis is that the
5:37:07proportion of wins in australia or
5:37:10abroad is independent of the country
5:37:12played against
5:37:14the alternate hypothesis is that the
5:37:16proportion of wins in australia or
5:37:18abroad is dependent on the country
5:37:21played against
5:37:22chi-square critical is 6.251
5:37:26and chi-square calculated is 1.36
5:37:30since the calculated value is less than
5:37:32the critical value the proportion of
5:37:34wins of the australia hockey team is
5:37:37independent of the country played or
5:37:40place
5:37:42in this screen we will discuss
5:37:44hypothesis test for proportions with an
5:37:46example
5:37:47the hypothesis test on population
5:37:50proportion can be performed
5:37:52to understand this with an example click
5:37:55the button given on the screen
5:37:59let us perform hypothesis test on
5:38:01population proportion
5:38:03the null hypothesis is that the
5:38:05proportion of smokers among males in a
5:38:07place named r is 0.10
5:38:11represented as p 0.
5:38:14the alternative hypothesis is the
5:38:16proportion is different than 0.10
5:38:19in notation it is represented as null
5:38:22hypothesis is p equals p zero against
5:38:26alternative hypothesis is p different
5:38:29than p zero
5:38:31a sample of 150 adult males are
5:38:34interviewed and it is found that 23 of
5:38:37them are smokers
5:38:38thus the sample proportion is 23 divided
5:38:42by 150 which is 0.153
5:38:46substituting this value in the
5:38:48expression of z given on the screen
5:38:51gives the result of 1.80
5:38:54you can reject the null hypothesis at
5:38:56level of significance alpha if z is
5:38:59greater than z alpha for five percent
5:39:02level of confidence the z value should
5:39:05be 1.96
5:39:07since the calculated z value is more
5:39:09than what is required for five percent
5:39:11level of confidence the null hypothesis
5:39:14is rejected
5:39:16hence it can be concluded that the
5:39:18proportion of smokers in r is greater
5:39:21than 0.10
5:39:24in this screen we will focus on
5:39:26comparison of means of two processes
5:39:29means of two processes are compared to
5:39:32understand whether the outcomes of the
5:39:34two processes are significantly
5:39:36different
5:39:37this test is helpful in understanding
5:39:39whether a new process is better than an
5:39:41old process
5:39:43this test can also determine whether the
5:39:45two samples belong to the same
5:39:47population or different populations
5:39:50it is especially required for
5:39:52benchmarking to compare an existing
5:39:54process with another benchmarked process
5:39:58let us proceed to the next screen to
5:40:00learn about the paired comparison
5:40:01hypothesis test for means
5:40:04the example of two mean t-tests with
5:40:06unequal variances is discussed here
5:40:10null and alternate hypotheses are
5:40:12defined
5:40:13the average heights of men in two
5:40:15different sets of people are compared to
5:40:17see if the means are significantly
5:40:19different
5:40:20for this test the sample sizes means and
5:40:23variances are required to calculate the
5:40:26value of t
5:40:28two samples of sizes n one of one
5:40:31hundred twenty five and n two of one
5:40:33hundred ten are taken from the two
5:40:36populations
5:40:37the mean value of sample size one is one
5:40:40hundred 167.3
5:40:42and sample size 2 is 165.8
5:40:47the standard deviation for sample sizes
5:40:501 and 2 are 4.2 and 5.0 respectively
5:40:55using the formula given on the screen
5:40:58the t value is derived as 2.47
5:41:02the null hypothesis is rejected if the
5:41:05calculated value of t is more than the
5:41:07required value of t in other words
5:41:11reject null hypothesis at level of
5:41:13significance a if computed t value is
5:41:16greater than t of df a divided by two
5:41:21with a t-test we're comparing two means
5:41:24and the population parameter sigma is
5:41:26unknown
5:41:27therefore we're pulling the sample
5:41:29standard deviations in order to
5:41:31calculate t
5:41:33the variances are weighted by the number
5:41:35of data points in each sample group
5:41:38since t 223 and
5:41:410.025 equals 1.96
5:41:44the null hypothesis is rejected at 5
5:41:48level of significance
5:41:50the test used here is known as the
5:41:52paired t-test and is considered a very
5:41:55powerful test in the next screen we will
5:41:58look into the example of the paired
5:42:00comparison hypothesis test for variance
5:42:03of f-test
5:42:05it is important to understand the
5:42:07different types of tests through an
5:42:09example
5:42:10susan is trying to compare the standard
5:42:12deviation of two companies
5:42:15according to her the earnings of company
5:42:17a are more volatile than those of
5:42:20company b
5:42:21she has been obtaining earnings data for
5:42:23the past 31 years for company a
5:42:26and for the past 41 years for company b
5:42:30she finds that the sample standard
5:42:32deviation of company a's earnings is
5:42:34four dollars forty cents
5:42:36and of company b's earnings is three
5:42:39dollars ninety cents
5:42:41determine whether the earnings of
5:42:42company a have a greater standard
5:42:45deviation than those of company b at
5:42:47five percent level of significance click
5:42:50the button given on the screen to know
5:42:53the answer
5:42:55susan has the data of the earnings of
5:42:58the companies
5:42:59distributions rarely have the same
5:43:01spread or variance
5:43:03when processes are improved one of the
5:43:06strategies is to reduce the variation
5:43:09it is important to be able to compare
5:43:11variances with each other
5:43:13a null hypothesis would indicate no
5:43:16change has occurred
5:43:17if it can be rejected and the variance
5:43:20is lower one can claim success
5:43:23the statistical notation for this
5:43:25example is given on the screen
5:43:28suppose one has to compare two sets of
5:43:30company data
5:43:32susan has looked at the earnings of two
5:43:34companies
5:43:35she has been studying the effects of
5:43:37strategy management styles and
5:43:39leadership profiles on the earnings of
5:43:42these companies
5:43:43there are significant differences in
5:43:45these kpiv's
5:43:47she wants to know if they have an effect
5:43:49on the variance in the earnings
5:43:51she has sample data over several decades
5:43:54for each company by the given data it
5:43:57can be concluded that earnings of
5:43:59company a have a greater standard
5:44:01deviation than those of company b
5:44:04in calculating the f-test statistic
5:44:07always put the greater variance in the
5:44:09numerator
5:44:11let us look at the f test example of
5:44:13hypothesis test for equality of variance
5:44:16in this screen
5:44:18the degrees of freedom for company a and
5:44:20company b are 30 and 40 respectively
5:44:24the critical value from f table equals
5:44:261.74
5:44:28the null hypothesis is rejected if the f
5:44:32test statistic is greater than 1.74
5:44:36the calculated value of f test statistic
5:44:38is 1.273
5:44:41and therefore at the 5 significance
5:44:44level the null hypothesis cannot be
5:44:46rejected
5:44:48the next screen will focus on hypothesis
5:44:50tests f-test for independent groups
5:44:55a restaurant which wants to explore the
5:44:57recent overuse of avocados suspects
5:45:00there is a difference between two chefs
5:45:03and the number of avocados used to
5:45:05prepare the salads
5:45:06the data shown in the table is the
5:45:08measure of avocados in ounces
5:45:11the weight of avocado slices used in
5:45:13salads prepared by two different chefs
5:45:16is to determine if one chef is using
5:45:18more avocados than the other
5:45:21perhaps the restaurant's expenditures on
5:45:23avocados is greater this month than the
5:45:26average of the past 12 months
5:45:28this is assuming there is no change in
5:45:30avocado prices or the amount of avocados
5:45:33being used
5:45:35click the tab to learn to conduct an
5:45:37f-test in ms excel
5:45:41the f-test is conducted in ms excel
5:45:44through the following steps
5:45:46open ms excel
5:45:48click data click data analysis
5:45:52please follow the facilitator
5:45:53instruction on how to install add-ins
5:45:56select f-test to sample for variances
5:46:00in variable one range select the data
5:46:02set for group a and select data set for
5:46:05group b in variable 2 range
5:46:08click ok
5:46:10the screen shot of the f test window is
5:46:12also shown here
5:46:14in this screen we will discuss the f
5:46:17test assumptions
5:46:18before interpreting the f test the
5:46:21assumptions to be considered are
5:46:23null hypothesis
5:46:25there is no significant statistical
5:46:27difference between the variances of the
5:46:29two groups thus concluding any variation
5:46:32could be because of chance
5:46:34this is common cause of variation
5:46:37alternate hypothesis
5:46:39there is a significant statistical
5:46:41difference between the variances of the
5:46:44two groups
5:46:45thus concluding that variations could be
5:46:47because of assignable causes too
5:46:50this is special cause of variation
5:46:53the following screen will focus on
5:46:55f-test interpretations
5:47:00the interpretations for the conducted
5:47:02f-test are
5:47:04from the excel result sheet the p-value
5:47:07is 0.03
5:47:09if p-value is low or below 0.05 the null
5:47:14must be rejected
5:47:15thus no hypothesis with 97 confidence is
5:47:20rejected
5:47:21also the fact that variation could only
5:47:23be due to common cause of variation is
5:47:26rejected it is inferred from the test
5:47:29that there could be assignable causes of
5:47:32variation or special causes of variation
5:47:36excel provides the descriptive
5:47:38statistics for each variable
5:47:40it also gives the degrees of freedom for
5:47:42each
5:47:43f is the calculated f statistic f
5:47:46critical is a reference number found in
5:47:48a statistics book table
5:47:50p of f less than or equal to f is the
5:47:54probability that f really is less than f
5:47:57critical
5:47:58or that the null hypothesis would be
5:48:00falsely rejected
5:48:02since the p-value is less than the alpha
5:48:05the null hypothesis can be confidently
5:48:08rejected
5:48:09alongside conducting a hypothesis test
5:48:12a meaningful conclusion from the test
5:48:14has been drawn
5:48:16the following screen will focus on
5:48:18hypothesis test t-test for independent
5:48:21groups
5:48:23as discussed earlier the table shows the
5:48:25measure of avocados in ounces and the
5:48:27significant difference in their means
5:48:30needs to be inspected
5:48:32if a significant amount of difference is
5:48:34found it can be concluded that there is
5:48:36a possibility of special cause of
5:48:38variation
5:48:40the next screen will demonstrate how to
5:48:42conduct the two-sample t-test
5:48:46the two-sample independent t-test
5:48:48inspects two groups of data for
5:48:50significant difference in their means
5:48:53the idea is to conclude if there is a
5:48:55significant amount of difference
5:48:58if there is a statistical evidence of
5:49:00variation
5:49:01one can conclude a possibility of
5:49:03special cause of variation the steps for
5:49:06conducting a two-sample t-test are
5:49:10open ms excel click data and click data
5:49:13analysis
5:49:15select two-sample independent t-test
5:49:18assuming unequal variances
5:49:20in variable 1 range select the data set
5:49:23for group a and select the data set for
5:49:26group b in variable 2 range
5:49:29keep the hypothesized mean difference as
5:49:320.
5:49:33click ok
5:49:35in the following screen we will focus on
5:49:37two sample independent t test
5:49:39assumptions
5:49:42the assumptions for a two sample
5:49:44independent t test are
5:49:46null hypothesis
5:49:48there is no significant statistical
5:49:50difference between the means of the two
5:49:52groups thus concluding any variation
5:49:54could be because of chance
5:49:57this is common cause of variation
5:50:00alternate hypothesis
5:50:02there is a significant statistical
5:50:04difference between the means of the two
5:50:06groups
5:50:07thus concluding that variations could be
5:50:09because of assignable causes too
5:50:12this is special cause of variation
5:50:15the null hypothesis states the mean of
5:50:17group a is equal to the mean of group b
5:50:21the alternate hypothesis states that the
5:50:23mean of group a is not equal to the mean
5:50:26of group b
5:50:28note that alternate hypothesis tests two
5:50:30conditions
5:50:32mean of a is less than mean of b
5:50:35and mean of a is greater than mean of b
5:50:38thus a two-tailed probability needs to
5:50:40be used
5:50:42before we interpret the t-test results
5:50:45let us compare the two-tailed and
5:50:47one-tailed probability in the next
5:50:49screen
5:50:50two-tailed probability and one-tailed
5:50:53probability are used depending on the
5:50:55direction of the alternate hypothesis
5:50:58if the alternate hypothesis tests more
5:51:00than one direction either less or more
5:51:03use a two-tailed probability value from
5:51:06the test
5:51:07example
5:51:08if mean of a is not equal to mean of b
5:51:12then it is two-tailed probability
5:51:15if the alternate hypothesis tests only
5:51:17one direction
5:51:19use a one-tailed probability value from
5:51:21the test
5:51:22example if mean of a is greater than
5:51:26mean of b then it is one-tailed
5:51:28probability
5:51:30in the next screen let us look at the
5:51:32two-sample independent t-test results
5:51:35and interpretations
5:51:37the results are shown in the table on
5:51:39the screen
5:51:41the inference is as the two-tailed
5:51:43probability is being tested the p-value
5:51:46of two-tailed probability testing is
5:51:480.24
5:51:49which is greater than 0.05
5:51:52if p-value is greater than 0.05 the null
5:51:56hypothesis is not rejected
5:51:59this means one cannot reject the fact
5:52:01that there is no significant statistical
5:52:03difference between the two means similar
5:52:06to the f-test excel provides the
5:52:08descriptive statistics for each group or
5:52:10variable
5:52:11the t stat is shown
5:52:13excel also shows one-tailed or
5:52:15two-tailed data
5:52:17for the one-tailed test the alpha is
5:52:190.05
5:52:21the error is expected to be in one
5:52:23direction
5:52:24for the two-tailed test the error is
5:52:27alpha slash two or zero point zero two
5:52:30five
5:52:32in this example t stat or t calculated
5:52:35is less than either t criticals
5:52:37therefore the null hypothesis cannot be
5:52:39rejected
5:52:40thus it can be inferred that both the
5:52:42groups are statistically same
5:52:45we will discuss the paired t-test in the
5:52:47next screen
5:52:49paired t-test is another hypothesis test
5:52:52from the family of t-tests
5:52:55the following points will help in
5:52:56understanding the paired t-test in
5:52:58detail
5:53:00the paired t-test is one of the most
5:53:01powerful tests from the t-test family
5:53:05the paired t-test is conducted before
5:53:07and after the process to be measured for
5:53:10example a group of students score x in
5:53:13cssgb
5:53:14before taking the training program
5:53:17post the training program the scores are
5:53:19taken again one needs to find out if
5:53:21there is a statistical difference
5:53:23between the two sets of scores
5:53:26if there is a significant difference the
5:53:28inference could be that the training was
5:53:30effective it is important to note that
5:53:33the paired t-test interpretation shows
5:53:36the effectiveness of the improvement
5:53:37measures
5:53:39this is the main reason why paired
5:53:41t-tests are often used in the improve
5:53:43stage let us learn about sample variance
5:53:46in the following screen
5:53:48sample variance is defined as the
5:53:50average of the squared differences from
5:53:53the mean
5:53:54the sample variance that is s square can
5:53:57be used to calculate and understand the
5:53:59degree of variation of a sample
5:54:02it can also be used in statistics
5:54:05however it cannot be used or explained
5:54:07directly because its value does not
5:54:10provide any information
5:54:12to use the value you have to first
5:54:14convert it into standard deviation and
5:54:17then combine it with the mean
5:54:19click the button to know the steps for
5:54:21calculating the sample variance
5:54:26step one calculate the mean or average
5:54:29of the sample
5:54:30step two subtract each of the values
5:54:33from the mean
5:54:35step three calculate the square value of
5:54:38the results
5:54:39step four take the average of the
5:54:42squared differences
5:54:45let us understand how to calculate the
5:54:47sample variance with the help of an
5:54:49example in this screen
5:54:52consider the sample of weights
5:54:54the mean value is one hundred forty
5:54:57when you subtract the individual values
5:54:59from the mean
5:55:00take the square value of the results and
5:55:03then take the average of the squared
5:55:05differences
5:55:06you will get 1936
5:55:09as the sample variance
5:55:12this number is not useful as it is
5:55:14in order to get the standard deviation
5:55:17take the square root of the sample
5:55:19variance
5:55:20square root of 1936 equals forty four
5:55:24the standard deviation in combination
5:55:26with the mean will tell you how much the
5:55:29majority of the people weigh
5:55:31in this example if your mean is one
5:55:33hundred forty and your variance is forty
5:55:36four
5:55:37you can conclude that the majority of
5:55:40people weigh between 96 pounds
5:55:43mean minus 44
5:55:45and 184 pounds
5:55:47mean plus 44.
5:55:50let us proceed to the next screen that
5:55:52focuses on the analysis of variance or
5:55:55anova which is the comparison of more
5:55:58than two means
5:56:00a t-test is used for one sample
5:56:03two sample tests are used for comparing
5:56:05two means
5:56:07to compare the means of more than two
5:56:08samples use the anova method
5:56:12anova stands for analysis of variance
5:56:15anova does not tell the better mean it
5:56:18helps in understanding that all the
5:56:19sample means are not equal
5:56:21the shortlisted samples based on anova
5:56:24output can further be tested
5:56:26one important aspect of anova is it
5:56:29generalizes the t test to include more
5:56:31than two samples
5:56:33performing multiple two-sample t-tests
5:56:36would increase the chance of committing
5:56:37a type one error
5:56:39hence anovas is useful in comparing two
5:56:42or more means
5:56:43the next screen will help in
5:56:45understanding this concept through an
5:56:47exam
5:56:48as an example consider the takeaway food
5:56:50delivery time of three different outlets
5:56:54is there any evidence that the averages
5:56:56for the three outlets are not equal
5:56:59in other words can the delivery time be
5:57:01benchmarked on the outlet
5:57:04the null hypothesis will assume that the
5:57:06three means are equal
5:57:08if the null hypothesis is rejected it
5:57:11would mean that there are at least two
5:57:13outlets that are different in their
5:57:14average delivery time
5:57:17in minitab one can perform anova in one
5:57:20of the statistics packages
5:57:22ensure that the data of the table is
5:57:24stacked in two columns
5:57:26in the main menu go to stat anova and
5:57:29then one way
5:57:31the left column of the table will have
5:57:32the outlets and the right column will
5:57:34have the time in minutes
5:57:37this is similar to the table shown on
5:57:39the screen
5:57:40in the one-way analysis of variance
5:57:42window
5:57:43select the response as delivery time and
5:57:46factor as outlet and click ok
5:57:50the output of this process is shown here
5:57:53notice the p value which is much higher
5:57:56than 0.05
5:58:00the steps to perform anova in excel are
5:58:02as follows
5:58:04after entering the data to a spreadsheet
5:58:07select the anova single factor test from
5:58:09the data analysis tool pack select the
5:58:12array for analysis designate that the
5:58:15data is in columns and select an output
5:58:17range
5:58:19excel shows the descriptive statistics
5:58:21for each column in the top table
5:58:24in the second table the anova analysis
5:58:26shows whether the variation is greater
5:58:28between the groups or within the groups
5:58:32it shows the sum of squares or ss
5:58:34degrees of freedom df
5:58:37means of squares ms or sum of squares
5:58:39divided by n minus 1 or variance
5:58:42the f statistic
5:58:44ms between divided by ms within
5:58:47p-value and f-critical from a reference
5:58:49table
5:58:51the f and p are calculated for the
5:58:52variation that occurs within each of the
5:58:54groups and between the groups
5:58:57if the conditions in the groups are
5:58:59significant it would be expected to see
5:59:02the between groups ss much higher and
5:59:05the ms slightly higher let us now
5:59:08interpret the minitab anova results
5:59:11since the p-value is more than 0.05
5:59:15the null hypothesis is accepted
5:59:18this means there is no significant
5:59:20difference between means of delivery
5:59:22time for the three outlets
5:59:25based on the confidence intervals it is
5:59:27found that the intervals overlap
5:59:29which means there is little that
5:59:30separates the means of the three samples
5:59:34in one-way anova where there was only
5:59:36one factor to be benchmarked that is the
5:59:39outlet of delivery
5:59:40if there are two factors you may use the
5:59:43two-way anova
5:59:45in this screen we will learn in detail
5:59:48about chi-square distribution
5:59:50the chi-square distribution is one of
5:59:53the most widely used probability
5:59:55distributions in inferential statistics
5:59:58it is also known as hypothesis testing
6:00:01and the distribution is used in
6:00:03hypothesis tests
6:00:05when used in hypothesis test needs one
6:00:08sample for the test to be conducted
6:00:11the chi-squared distribution is also
6:00:13known as chi squared
6:00:15it has k1 degrees of freedom and is the
6:00:18distribution of a sum of the squares of
6:00:20k independent standard normal random
6:00:22variables suppose in a field for nine
6:00:25players player one comes in and can
6:00:27choose amongst all nine positions
6:00:29available
6:00:30player two can choose only amongst eight
6:00:33and so on
6:00:34after all the eight players have chosen
6:00:36their positions the last player gets to
6:00:39choose the last position left
6:00:42the eight players are free to choose in
6:00:44a playing field of nine eight is the
6:00:46degree of freedom for this example
6:00:48conventionally degree of freedom is n1
6:00:51where n is the sample size
6:00:54for example if w x y and z are four
6:00:57random variables with standard normal
6:01:00distributions
6:01:01the random variable f which is the sum
6:01:03of w square x square y square and z
6:01:07square has a chi-square distribution
6:01:10the degrees of freedom of the
6:01:11distribution or the df equals the number
6:01:15of normally distributed variables used
6:01:17in this case df equals four
6:01:21the formula for chi-squared distribution
6:01:23is shown on the screen
6:01:25it is important to note that f of o
6:01:28stands for an observed frequency and f
6:01:31of e stands for an expected frequency
6:01:34the next screen will explain chi-square
6:01:36test through an example
6:01:39suppose the australian hockey team
6:01:41wishes to analyze its wins at home and
6:01:44abroad against four different countries
6:01:47the data has two classifications and the
6:01:49table is also known as a two by four
6:01:52contingency table with two rows and four
6:01:55columns
6:01:56the expected frequencies can be
6:01:58calculated assuming there is a
6:02:00relationship
6:02:02thus expected frequency for each of the
6:02:04observed frequency is equal to product
6:02:07of row total and column total divided by
6:02:10overall total
6:02:11one has to find out how to calculate the
6:02:14expected frequency
6:02:16if the observed frequency is three wins
6:02:18against south africa in australia then
6:02:21it would convert to total winds at home
6:02:23which is 21 divided by the total number
6:02:26of wins or 31 and is the result
6:02:29multiplied by five
6:02:31the result is 3.39
6:02:34similarly the expected population
6:02:37parameters for all cases are found
6:02:40in this step all the information of the
6:02:42previous screen is combined and the
6:02:44table is populated
6:02:46the estimated population parameters are
6:02:49calculated and added
6:02:51the formula estimates the observed
6:02:52frequency to calculate the final
6:02:54chi-square index which in this case is
6:02:571.36
6:02:59it is important to note that there is a
6:03:01different chi-squared distribution for
6:03:04each of the different numbers of degrees
6:03:05of freedom
6:03:07for the chi-square distribution the
6:03:09degrees of freedom are calculated as per
6:03:11the number of rows and columns in the
6:03:13contingency table
6:03:15the equation for degrees of freedom
6:03:17should be noticed
6:03:18the number of degrees of freedom is
6:03:20equal to three
6:03:22assuming an alpha of ten percent the
6:03:24chi-squared distribution in the
6:03:26chi-square table is noticed and a
6:03:29critical chi-square index of 6.251 is
6:03:32arrived at chi-square calculated value
6:03:35is 1.36
6:03:38both the values of the chi-square index
6:03:40should be plotted
6:03:42the critical chi-square distribution
6:03:44divides the whole region into acceptance
6:03:47and rejection
6:03:48while the calculated chi-square
6:03:50distribution is based on data and
6:03:52conveys whether the data falls into an
6:03:55acceptance or rejection region
6:03:58therefore as the calculated value is
6:04:00less than the critical value
6:04:02and falls in the acceptance region
6:04:05the proportion of wins of the aussie
6:04:06team at home or abroad has nothing to do
6:04:09with the opponent being played against
6:04:12let us proceed to the next topic of this
6:04:15lesson in the next screen in this topic
6:04:17we will learn in detail about hypothesis
6:04:20testing with non-normal data
6:04:23let us begin with the mon whitney test
6:04:25in the next screen
6:04:27the mon whitney test also known as the
6:04:30wilcoxon rank sum test is a
6:04:32non-parametric test which is used to
6:04:35compare two unpaired groups
6:04:37in this test the value of alpha is by
6:04:40default set at 0.05 and the rejection
6:04:44and acceptance condition remains the
6:04:46same for different cases
6:04:48that is if p is less than alpha reject
6:04:51the null hypothesis
6:04:53if p is greater than alpha reject the
6:04:56alternate hypothesis
6:04:58the aim of this test is to rank the
6:05:00entire data available for each condition
6:05:03and then compare the total outcome of
6:05:05the two ranks
6:05:06click the button to know the steps to
6:05:08perform the mon whitney test
6:05:12to perform the mon whitney test first
6:05:15rank all the values from low to high
6:05:17without paying any attention to the
6:05:19group to which each value belongs
6:05:21the smallest number gets a rank of one
6:05:24the largest number gets a rank of n
6:05:26where n is the total number of values in
6:05:29the two groups
6:05:30if there are ties continue to rank the
6:05:33values anyway pretending they are
6:05:35slightly different
6:05:37then find the average of the ranks for
6:05:39all the identical values and assign that
6:05:42rank to all these values
6:05:44continue this till all the whole number
6:05:46ranks have been used
6:05:48next sort the values into two groups
6:05:52these can now be used for the mon
6:05:53whitney u test
6:05:55summate the ranks for the observations
6:05:57from sample one and then summate the
6:05:59ranks in sample two larger group
6:06:04let us look at an example of the mon
6:06:06whitney test in this screen
6:06:09suppose you have two sets of data g1 and
6:06:12g2
6:06:13the g1 values are 14 2 5 16 and 9
6:06:19and the g2 values are 4 2 18 14 and 8.
6:06:24now combine the g1 and g2 values sort
6:06:28them in ascending order and mention the
6:06:30group name against each value
6:06:33next rank the groups from 1 to 10 and
6:06:36check if any values are identical
6:06:39take an average of the ranks of the
6:06:40identical values and place it against
6:06:43the identical values in the final rank
6:06:45column
6:06:46hence the average final rank is 1.5 for
6:06:49ranks 1 and 2.
6:06:51similarly the average final rank is 7.5
6:06:55for ranks 7 and 8.
6:06:57next calculate r1 and r2 by adding the
6:07:00ranks of the groups 1 and 2 respectively
6:07:04in this example the r 1 value is 28 and
6:07:07our 2 value is 27
6:07:10from the given data 5 is the value of
6:07:12both n1 and n2
6:07:16the formula for the mon whitney u test
6:07:19for n one and n two values is
6:07:21u one equals n one multiplied by n two
6:07:25plus n one multiplied by n one plus one
6:07:28whole divided by two minus r one
6:07:31similarly u2 equals n1 multiplied by n2
6:07:35plus n2 multiplied by n2 plus 1 whole
6:07:39divided by 2 minus r2
6:07:42in this example the value of u1 is 12
6:07:45and u2 is 13.
6:07:48now the u value can be calculated by
6:07:51taking the minimum value among 12 and 13
6:07:54which is 12.
6:07:55look up the mon whitney u test table for
6:07:58n 1 equals five and n two equals five
6:08:01you will get the critical value of u as
6:08:04two
6:08:05to be statistically significant the
6:08:07obtained u has to be equal to or less
6:08:10than this critical value
6:08:12our calculated value of u is 12 which is
6:08:15not less than two that means there is no
6:08:18statistical difference between the means
6:08:20of the two groups
6:08:22in this screen we will learn about the
6:08:24kruskal-wallis test
6:08:26the kruskal-wallis test is named after
6:08:28william kruskal and w alan wallace
6:08:31it is also a non-parametric test used
6:08:34for testing the source of origin of
6:08:36samples for example whether the samples
6:08:39originate from the same distribution or
6:08:41not
6:08:42the characteristics of the
6:08:43kruskal-wallis test are as follows
6:08:46it is the only way to analyze variants
6:08:49by ranks
6:08:50this test compares the medians of two or
6:08:53more samples to find out if the samples
6:08:55are from different populations
6:08:58since this test is a non-parametric
6:09:00method it does not assume the normal
6:09:03distribution of the residuals unlike the
6:09:06analogous one-way analysis of variance
6:09:09for this test the null hypothesis is
6:09:12that the medians of all groups are equal
6:09:15and the alternate hypothesis is that at
6:09:18least one population median of one group
6:09:21is different from the population median
6:09:23of at least one other group
6:09:25let us learn about the moods median in
6:09:27the next screen
6:09:30the mood's median is also a
6:09:32non-parametric test that is used to test
6:09:34the equality of medians from two or more
6:09:37different populations
6:09:39this test works when the output y
6:09:41variable is continuous and discrete
6:09:44ordinal or discrete count
6:09:46while the input x variable is discrete
6:09:48with two or more attributes
6:09:51click the button to view the steps
6:09:52involved in the mood's median test
6:09:56following are the steps in the mood's
6:09:58median test
6:10:00first find the median of the combined
6:10:02data set
6:10:03next find the number of values in each
6:10:05sample that are greater than the median
6:10:07and form a contingency table
6:10:10then find the expected value for each
6:10:12cell and finally find the chi-square
6:10:14value
6:10:16we will learn about the freedmen test in
6:10:18this screen the friedman test is another
6:10:21form of a non-parametric test
6:10:24it does not make any assumptions about
6:10:26the specific shape of the population
6:10:28from which the sample is drawn
6:10:31and therefore allows smaller sample data
6:10:33sets to be analyzed
6:10:35unlike anova the friedman test does not
6:10:38require the data set to be randomly
6:10:40sampled from normally distributed
6:10:42populations with equal variances
6:10:45this test uses a two-tailed hypothesis
6:10:48test
6:10:49where the null hypothesis is that the
6:10:52population medians of each treatment are
6:10:54statistically identical to the rest of
6:10:56the group
6:10:58in the next screen we will learn about
6:10:59the one sample sign test
6:11:03the one sample sign test is the simplest
6:11:06of all the non-parametric tests that can
6:11:08be used instead of a one-sample t-test
6:11:11it is similar to the concept of testing
6:11:14if a coin is fair in showing heads or
6:11:16tails
6:11:17here the null hypothesis represented as
6:11:20h o
6:11:21is the hypothecated median or assumed
6:11:24median of the sample which belongs to
6:11:26the population
6:11:27click the button to view the steps
6:11:29involved in the one sample sign test
6:11:33following are the steps in a one sample
6:11:36sign test
6:11:37first count the number of positive
6:11:39values
6:11:40these are the values that are larger
6:11:42than the hypothesized median
6:11:44next count the number of negative values
6:11:47these are the values that are smaller
6:11:49than the hypothesized median
6:11:51finally test the values to check if
6:11:53there are significantly more positive
6:11:56values or negative values than expected
6:12:00this screen will focus on the one sample
6:12:02wilcoxon test
6:12:04the one sample wilcoxon test also known
6:12:07as the wilcoxon signed rank test
6:12:10is another form of a non-parametric test
6:12:13this test is equivalent to the
6:12:15parametric one-sample t-test and more
6:12:18powerful than the non-parametric
6:12:20one-sample sign test
6:12:22let us discuss the characteristics of
6:12:24this test in the following screen
6:12:27some of the characteristics of this test
6:12:29are as follows
6:12:31this test assumes that the sample is
6:12:33randomly taken from a population
6:12:35with a symmetric frequency distribution
6:12:37around the median
6:12:39also in this test the symmetry can be
6:12:42observed with the histogram or by
6:12:44checking if the median and mean are
6:12:46approximately equal
6:12:48the conclusion in this test is that if
6:12:50the value is on the midpoint you can
6:12:53continue and accept the null hypothesis
6:12:56if not you need to reject the alternate
6:12:58hypothesis
6:13:00click the button given on the screen to
6:13:02view an example
6:13:05let us consider an example
6:13:07the median customer satisfaction score
6:13:09of an organization has always been 3.7
6:13:13and the management wants to see if this
6:13:15has changed
6:13:17they conduct a survey and get the
6:13:18results grouped by the customer type
6:13:21the conclusion will be as follows
6:13:24if the median value is 3.7 the null
6:13:27hypothesis h o can be accepted if not
6:13:31the alternate hypothesis needs to be
6:13:33rejected the alpha value will be 0.05
6:13:38choose from over 300 in-demand skills
6:13:40and get access to 1 000 plus hours of
6:13:43video content for free visit scale up by
6:13:45simply learn click on the link in the
6:13:47description to know more
6:13:49this lesson will focus on the improved
6:13:51phase of the dmacc process
6:13:54the improved phase comes after the
6:13:56analyze phase
6:13:57in the analyze phase the data was
6:13:59analyzed and some patterns were found to
6:14:02identify where the problem lies
6:14:05design of experiments or doe consists of
6:14:08a series of planned and scientific
6:14:10experiments that test various input
6:14:12variables and their eventual impact on
6:14:15the output variable
6:14:17design of experiments can be used as a
6:14:19one-stop alternative for analyzing all
6:14:22influencing factors to arrive at a
6:14:24successful model
6:14:26doe is applicable where multiple input
6:14:28variables known as factors affect a
6:14:31single response variable
6:14:33an output variable is the variable which
6:14:35may get affected due to multiple input
6:14:38variables
6:14:39doe is preferred over one factor at a
6:14:42time or o fat experiments because it
6:14:44does not miss interactions
6:14:47with techniques like blocking
6:14:48experimental error can be eliminated
6:14:51the trials should be randomized to avoid
6:14:54concluding that a factor is significant
6:14:56when the time at which it is measured or
6:14:57the sequence followed may have
6:14:59influenced the responses result
6:15:01an example of blocking is highlighted in
6:15:03the table given on the screen
6:15:06with techniques like replication many
6:15:08experiments can be conducted to ensure a
6:15:10robust model
6:15:12we will understand the concept of design
6:15:14experiments through an example in the
6:15:16upcoming screen
6:15:18to understand doe and the main effects
6:15:21consider the following example
6:15:23suppose the objective of the experiment
6:15:26is to achieve uniform part dimensions at
6:15:28a particular target value to reduce
6:15:30variations
6:15:32the inputs x or factors that affect the
6:15:35output are cycle time mold temperature
6:15:38holding pressure
6:15:39holding time and material type
6:15:42the process is the molding process and
6:15:45the output or the response of the
6:15:46experiment is the part hardness
6:15:49the components of the doe in this
6:15:51example will be described in the next
6:15:53screen
6:15:57output response factors levels and
6:16:00interactions are the components of the
6:16:01doe in the given example
6:16:04click each component to learn more
6:16:07the response variable is the part
6:16:09hardness and is measured as a result of
6:16:11the experiment and is used to judge the
6:16:14effects of factors
6:16:16factors of this experimental setup are
6:16:19cycle time mold temperature holding
6:16:21pressure holding time and material type
6:16:26factors can be varied and are called
6:16:28levels the molding temperature can be
6:16:30set at 600 degrees fahrenheit or 700
6:16:34degrees fahrenheit
6:16:35plastic type can be fillers and no
6:16:38fillers and the material type has two
6:16:40levels of nylon or acetyl
6:16:45interactions refer to the degree to
6:16:47which factors depend on one another
6:16:50some experiments evaluate the effect of
6:16:52interactions
6:16:54in the molding example the interaction
6:16:56between cycle time and molding
6:16:58temperature is critical
6:17:01the best level for time depends on what
6:17:03temperature is set
6:17:05if the temperature level is higher the
6:17:07cycle time may have to be decreased to
6:17:10achieve the same response from the
6:17:12experiment
6:17:14let us understand full factorial
6:17:17experiments through an example
6:17:19full factorial experimental design
6:17:22contains all combinations of all levels
6:17:24of all factors
6:17:26this experimental design ensures no
6:17:29possible treatment combinations get
6:17:30omitted
6:17:32hence full factorial designs are often
6:17:34preferred over other designs
6:17:37the table shown here is for a two-way
6:17:39heat treatment experiment
6:17:41there are two factors
6:17:43oven time x2 and the temperature x1 at
6:17:47which the material is drawn out of the
6:17:49oven
6:17:50the output y of the experiment is the
6:17:52hardness of the material
6:17:54each of the factors has two levels
6:17:57this example illustrates the concepts of
6:18:00main factor and interaction effects
6:18:03from the table it is clear that without
6:18:05repetition the experiment will have four
6:18:07different outcomes based on the changes
6:18:10in levels of factors
6:18:12each experimental trial here is repeated
6:18:15to give a total of eight values
6:18:17let us now analyze the mean effect
6:18:20an analysis of the means helps in
6:18:23understanding how a change in
6:18:24temperature at which the material is
6:18:26drawn creates a difference in the
6:18:28average part hardness
6:18:30this affects the output and is called
6:18:32the main effect
6:18:34analysis of means also tells how a
6:18:36change in oven time creates a difference
6:18:38in the average part hardness
6:18:40this is also the main effect
6:18:43analysis of means explains how
6:18:45interaction between temperature and time
6:18:48affects the average part hardness
6:18:50this is known as the interaction effect
6:18:53let us next understand the concept of
6:18:56main effect for calculating the main
6:18:59effect the means have to be calculated
6:19:02hence to calculate the main effect of
6:19:04draw temperature the mean of the
6:19:06hardness values has to be calculated
6:19:09the values are populated in the
6:19:10corresponding columns of draw
6:19:12temperatures
6:19:13the columns have been labeled a1 and a2
6:19:17the value of the mean of a1 is 91
6:19:20and of a2 is 82.
6:19:23plotting the data on a graph shows that
6:19:26changing draw temperatures changes the
6:19:28average hardness
6:19:30similarly we calculate the mean of
6:19:32hardness values in b1 and b2
6:19:36the values are 87 and 86 which are
6:19:39plotted on a graph
6:19:41it can be seen that changing the oven
6:19:43time does not affect the average
6:19:45hardness
6:19:47now let us understand how the
6:19:49interaction between temperature and time
6:19:51affects the average part hardness to
6:19:54check how draw temperature and oven time
6:19:56interact
6:19:57the mean values are calculated by taking
6:20:00the repetition response
6:20:02hence the cell a1 b1 has the mean of the
6:20:05values 90 and 87. the cell a2 b1 has the
6:20:10mean of the values 84 and 87.
6:20:14after the mean values are calculated
6:20:16they are plotted on a graph
6:20:18the graph shows that to reduce
6:20:20interactions
6:20:21low temperature and high oven time
6:20:24should be selected to have the desired
6:20:26output of high hardness
6:20:28also called renault hardness
6:20:30if low hardness is the desired output
6:20:33the experimental setup should have high
6:20:36draw temperature and high oven time
6:20:39the ideal case is represented by the
6:20:41parallel lines which give the desired
6:20:44output based on the main effect without
6:20:46being affected by the interaction
6:20:48between the factors the parallel lines
6:20:51are shown as a dotted line
6:20:53the mean of the factors are also
6:20:55calculated and shown in the small table
6:20:59in this we will introduce the concept of
6:21:01runs in design of experiments
6:21:04the numbers of experiments in a doe
6:21:06setting is known as runs
6:21:09a full factorial experiment without
6:21:11replication on five factors and two
6:21:13levels is two raised to the power of
6:21:16five which equals 32 runs
6:21:19a full factorial experiment with one
6:21:21replication on five factors and two
6:21:24levels is 32 plus 32 which equals 64
6:21:28runs
6:21:29a half fractional factorial experiment
6:21:32without replication on five factors and
6:21:35two levels is two raised to the power of
6:21:37five minus one which equals sixteen runs
6:21:42a half factorial experiment with one
6:21:44replication on five factors and two
6:21:46levels is 16 plus 16 which equals 32
6:21:50runs
6:21:52the number of combinations can be
6:21:53determined using the formula l to the
6:21:56power of f where l is the number of
6:21:58levels and f is the number of factors
6:22:01half fractional factorial is calculated
6:22:03using the formula l to the power f minus
6:22:06one
6:22:07at three levels five factors full
6:22:10factorial experiment would amount to 243
6:22:13trials and half factorial experiments
6:22:16would require 81 trials
6:22:18the difference between full factorial
6:22:20and half fractional factorial
6:22:22experiments can be seen from the number
6:22:24of runs
6:22:26let us proceed to the next topic of this
6:22:28lesson in the following screen
6:22:31in this topic we will discuss root cause
6:22:34analysis in detail
6:22:36we will learn about residuals analysis
6:22:38in the following screen
6:22:42while performing the regression analysis
6:22:45of a linear or non-linear model
6:22:47you will get a model with the predicted
6:22:49values
6:22:50some of the data might fit within that
6:22:52model whereas others may be scattered
6:22:54across the modeled equation predicts one
6:22:57value for y at level x
6:23:00however the actual value for y observed
6:23:03at that level of x is different from the
6:23:05predicted value
6:23:07this difference between the observed
6:23:09value of the dependent variable y
6:23:12and the predicted value is called
6:23:14residual
6:23:15the formula to calculate residual is
6:23:18observed value minus predicted value
6:23:21residuals are considered to be errors
6:23:24and each data point has one residual
6:23:26you can validate the assumptions on
6:23:28random errors as they are independent
6:23:31exhibit normal distribution have a
6:23:33constant various sigma square for all
6:23:35the settings of the independent
6:23:37variables and finally have a mean as
6:23:40zero
6:23:41in the next slide we will continue to
6:23:43discuss residuals analysis
6:23:46as discussed in the previous screen
6:23:49while performing any regression analysis
6:23:51you will observe that not all the data
6:23:53fits into the linear model
6:23:56as the linear regression model is not
6:23:58always appropriate for the data
6:24:00therefore you should assess the
6:24:02appropriateness of the model by defining
6:24:05residuals and examining the residual
6:24:07plots
6:24:09if all assumptions are satisfied the
6:24:11residuals should randomly vary around
6:24:14zero
6:24:15and the spread of the residuals should
6:24:17be the same throughout the plot
6:24:19that is no systematic patterns are
6:24:21observed
6:24:23remember in residuals analysis both the
6:24:26sum and the mean of the residuals are
6:24:28zero
6:24:30residuals and diagnostic statistics
6:24:32allow you to identify patterns that
6:24:34either poorly fit in the model with a
6:24:37strong influence on the estimated
6:24:38parameters
6:24:40or have a high leverage
6:24:42it is helpful to interpret these
6:24:44diagnostics together to understand any
6:24:46potential problems with the model
6:24:49in the next screen we will learn about
6:24:51data transformation using the box cox
6:24:54method the available data must be
6:24:56transformed when it does not exhibit the
6:24:58normal distribution
6:25:00box and cox in the year 1964 developed a
6:25:03procedure for estimating the best
6:25:05transformation to normality within the
6:25:07family of power transformation
6:25:10it works by taking the current y data
6:25:12and raising it to the power known as
6:25:14lambda
6:25:15the formula for transformation of y is
6:25:18represented as y asterisk equals y to
6:25:21the power lambda minus one
6:25:24the whole divided by lambda
6:25:26this formula is used where the value of
6:25:28lambda is not zero if the value of
6:25:31lambda is zero you can use natural
6:25:33logarithm to transform y
6:25:36the family of power transformations can
6:25:38be used for the following
6:25:40for converting a data set so that
6:25:42parametric statistics can be used
6:25:44here lambda is a parameter to be defined
6:25:47from the data
6:25:48for any continuous data greater than
6:25:50zero this will not work when the values
6:25:53are less than or equal to zero
6:25:55transforming specs
6:25:57note that the use of the transformation
6:25:59does not guarantee normality
6:26:02in the next screen we will continue the
6:26:04discussion on data transformation using
6:26:06box and cox
6:26:09the table on the screen shows how the
6:26:11data can be transformed using lambda
6:26:14the first column lists down values of
6:26:16lambda and the second column shows the
6:26:18transformed value
6:26:20if the value of lambda is negative 2 it
6:26:23becomes y to the power negative two
6:26:25after the transformation which is one
6:26:27divided by y square
6:26:29similarly if the value of lambda is
6:26:32negative one after transformation it
6:26:34becomes y to the power of negative one
6:26:36which is one divided by y and so on note
6:26:39that you will use a different formula
6:26:41when you have the value of lambda as
6:26:43zero wherein you will take natural log
6:26:46of the value y
6:26:48similarly transform values are also
6:26:51shown on the screen
6:26:52click the example button to know more
6:26:56let us look at an example of how data
6:26:58transformation is done using box
6:27:02the difference between original data and
6:27:04the data transformed using box is
6:27:06shown on the screen
6:27:08figure one shows the original data
6:27:10plotted on a histogram
6:27:12here you can see that this data is
6:27:14abnormal
6:27:16in figure two the box and cox procedure
6:27:18is applied on the original data and it
6:27:21is transformed
6:27:23you can see that the data in the second
6:27:25figure is more normal than figure one
6:27:28let us learn about process input and
6:27:30output variables in the following screen
6:27:34process improvement has a few
6:27:36prerequisites before a process can be
6:27:39improved it must first be measured to
6:27:41assess the level of improvement required
6:27:44the first step is to know the input
6:27:46variables and output variables and check
6:27:49for any relationship
6:27:51the sipoc map and the cause and effect
6:27:54matrix are very helpful
6:27:56there are many ways to measure the key
6:27:58process variables
6:28:00metrics such as the percent defective
6:28:02operation costs elapsed time
6:28:05backlog quantity and documentation
6:28:08errors can be used
6:28:10critical variables are best identified
6:28:12by the process owners
6:28:14once they are identified cause and
6:28:16effect tools are used to establish the
6:28:18relationship between variables
6:28:21a cause and effect matrix is shown on
6:28:23the screen
6:28:24the key process input variables have
6:28:27been listed vertically
6:28:28and the key process output variables
6:28:31horizontally
6:28:32for each of the output variables a
6:28:34prioritization number is assigned
6:28:37numbers which reflect the effect of each
6:28:39input variable on the output variable
6:28:42are entered in the matrix
6:28:44the process output priority is
6:28:47multiplied with the input variables to
6:28:49arrive at the results for each input
6:28:51variable
6:28:52the values are added to determine the
6:28:54results for each input variable
6:28:57for process input variable 1 the output
6:29:00variables are 3 4 and 7 with a
6:29:04prioritization value of 4 7 and 11
6:29:07respectively
6:29:08therefore multiplying the output
6:29:10variables with their corresponding
6:29:12prioritization numbers and adding those
6:29:15gives one hundred seventeen
6:29:17which is around thirty three percent of
6:29:18the total effect
6:29:20the process input variables results are
6:29:23compared to each other to determine
6:29:25which input variable has the greatest
6:29:27effect on the output variables
6:29:30click the cause and effect matrix
6:29:32template button to view another template
6:29:36a sample of the cause effect matrix or
6:29:38ce matrix is shown here
6:29:41the ce matrix gives the correlation
6:29:43between input and output variables
6:29:47in this screen we will discuss the steps
6:29:49to update the ce matrix
6:29:52the steps for updating the cause and
6:29:53effect matrix r
6:29:56list the input variables vertically
6:29:58under the column process inputs
6:30:00list the output variables horizontally
6:30:03under the numbers 1 to 15.
6:30:06these output variables are important
6:30:08from the customer's perspective
6:30:10one can refer to either the qfd or the
6:30:13ctq tree to know the key output
6:30:16variables
6:30:17rank the output variables based on
6:30:20customer priority
6:30:22these numbers can also be taken from the
6:30:23qfd
6:30:25the input variables with the highest
6:30:27score become the point of focus in the
6:30:29project
6:30:30another method to establish the cause
6:30:32effect relation is the cause and effect
6:30:35diagram
6:30:36this is explained in detail in the
6:30:38following screen
6:30:39the cause and effect diagram is used to
6:30:42find the root cause and the potential
6:30:44solutions to a problem
6:30:46a cause and effect diagram breaks down a
6:30:49problem into bite-sized pieces and also
6:30:52displays the possible causes in a
6:30:54graphic manner
6:30:55it is also known as the fishbone the 4m
6:30:59or the ishikawa diagram
6:31:02it is commonly used to examine effects
6:31:04or problems to find out the possible
6:31:06causes and to indicate the possible
6:31:09areas to collect data
6:31:11the steps involved in the cause and
6:31:13effect diagram are
6:31:15all the possible causes of the problem
6:31:17or effects selected for analysis are
6:31:19brainstormed
6:31:21the major causes are classified under
6:31:23the headings of materials methods
6:31:26machinery and manpower
6:31:29the cause and effect diagram is drawn
6:31:31with the problem at the point of the
6:31:33central access line and the causes on
6:31:35the diagram are written under the
6:31:37classifications chosen
6:31:40the next screen illustrates the cause
6:31:42and effect diagram with the help of an
6:31:44example
6:31:47the diagram shows the cause and effect
6:31:49diagram for the possible causes of
6:31:51solder defects on a reflow soldering
6:31:53line
6:31:54this diagram helps in collecting data
6:31:57and discovering the root cause
6:31:59during brainstorming the group looked at
6:32:02all the major causes and then grouped
6:32:04them under the main headings
6:32:06under materials causes like types of
6:32:09solder paste components and the
6:32:11components packaging used are considered
6:32:14the major causes under methods are
6:32:16technology and preventive maintenance
6:32:19similarly operator and schedule are
6:32:21placed under manpower
6:32:23while tools and oven are grouped under
6:32:26machinery causes
6:32:28the next screen will discuss another
6:32:30root cause analysis tool in detail
6:32:335y is one of the tools used to analyze
6:32:36the root cause of a problem
6:32:38the responsibility of the root cause
6:32:40analysis lies with the 5y analysis team
6:32:44the technical experts have a great
6:32:46responsibility as the conclusion will be
6:32:48drawn from the way the drill down of the
6:32:50symptoms is carried out
6:32:53the 5y is a very simple tool as it poses
6:32:56the y question to every problem till the
6:32:59root cause is obtained
6:33:01it is important to know that the 5y tool
6:33:04does not restrict the interrogation to 5
6:33:06questions
6:33:08y can be asked as many times as required
6:33:11till the root cause for the problem is
6:33:13found
6:33:14it can be used along with the cause and
6:33:16effect diagram
6:33:18the following screen will explain the
6:33:19process of the five why technique
6:33:23the process for the 5-y technique is
6:33:27identify the problem and emphasize the
6:33:29problem statement
6:33:31arrange for a brainstorming session with
6:33:33the team including subject matter
6:33:35experts process owners and team members
6:33:40explain the purpose and the problem
6:33:41statement
6:33:43analyze scenarios working backwards from
6:33:45the problem
6:33:47ask why for the answers obtained until
6:33:50the root cause is found
6:33:52normally reasons like insufficient
6:33:55resources and time become the root
6:33:57causes
6:33:58if the drill down in brainstorming is
6:34:00carried out in the right direction
6:34:03it is often found that the root cause is
6:34:05related to the process
6:34:07therefore the occurrence of a problem is
6:34:10often due to the process
6:34:12and not an individual or a team
6:34:15in the next screen we will understand
6:34:17the concept of the five-wide technique
6:34:20with the help of an example
6:34:23the process for the 5y technique is
6:34:26identify the problem and emphasize the
6:34:28problem statement
6:34:30arrange for a brainstorming session with
6:34:32the team including subject matter
6:34:34experts process owners and team members
6:34:39explain the purpose and the problem
6:34:41statement
6:34:42analyze scenarios working backwards from
6:34:45the problem
6:34:46ask why for the answers obtained until
6:34:49the root cause is found
6:34:52normally reasons like insufficient
6:34:54resources and time become the root
6:34:56causes
6:34:58if the drill down in brainstorming is
6:35:00carried out in the right direction
6:35:02it is often found that the root cause is
6:35:04related to the process
6:35:06therefore the occurrence of a problem is
6:35:09often due to the process
6:35:11and not an individual or a team
6:35:15in the next screen we will understand
6:35:17the concept of the five why technique
6:35:19with the help of an example
6:35:22in this topic we will discuss lean tools
6:35:25in detail
6:35:26let us learn about lean techniques in
6:35:28the following screen
6:35:30the eight lean techniques are kaizen
6:35:32bokayoke 5s just in time kanban jadoka
6:35:38tactime and hijanka
6:35:40click each technique to know more
6:35:44kaizen or continuous improvement is the
6:35:47building block of all lean production
6:35:49methods
6:35:50kaizen philosophy implies that all
6:35:52incremental changes routinely applied
6:35:55and sustained over a long period of time
6:35:57results in significant improvements
6:36:02the second technique is pokayoka it is
6:36:04also known as mistake proofing
6:36:06it is good to do it right the first time
6:36:09and even better to make it impossible to
6:36:11do it wrong the first time
6:36:13the prompt received to save the word
6:36:15document before closing it without
6:36:17saving is an example of bokoyoka
6:36:215s is a set of five japanese words which
6:36:24translate to sort
6:36:27set in order shine
6:36:29standardize and sustain
6:36:32this is a simple and yet powerful tool
6:36:34of lean
6:36:35the sort principle refers to sorting
6:36:38items according to a rule
6:36:40the rule could be frequency of use or
6:36:43time of use
6:36:44after sorting the objects are set in
6:36:46order
6:36:47the place for everything is defined and
6:36:50everything is placed accordingly
6:36:52cleaning of the area refers to the shine
6:36:55principle
6:36:56the fourth step requires formation and
6:36:58circulation of a set of written
6:37:00standards
6:37:01the last step refers to sustaining the
6:37:04process by following the standards set
6:37:06earlier
6:37:085s is useful as a framework to create
6:37:10and maintain the workplace
6:37:14just in time or jit is another lean
6:37:18technique
6:37:19this technique philosophizes about
6:37:21producing the necessary units in the
6:37:23necessary quantity at the necessary time
6:37:26with the required quality
6:37:29as an item is removed from a shelf of a
6:37:31super mart the system confirms it and
6:37:34automatically sends a note for
6:37:35replenishment
6:37:37this kind of technique can be used in an
6:37:39organization to prevent accumulation of
6:37:41inventory
6:37:44the fifth technique is known as kanban
6:37:46which means signboard in japanese
6:37:49kanban utilizes visual display cards to
6:37:52signal movement of material between the
6:37:54steps of a product process
6:37:57this is one of the examples of visual
6:37:59control in lean
6:38:02the next technique is jiduka
6:38:05it means automation with human touch and
6:38:07is sometimes known as autonomation
6:38:10jiduka implements supervisory function
6:38:12in the production line and stops the
6:38:14process as soon as a defect is
6:38:16encountered
6:38:18the process does not start again till
6:38:20the root cause of the defect is
6:38:21eliminated
6:38:24tack time is the maximum time in which
6:38:26the customer demands need to be met
6:38:29for example a customer needs 100
6:38:31products and the company has 420 minutes
6:38:34of available production time
6:38:36tax time equals time available divided
6:38:39by demand
6:38:41in this case the company has a maximum
6:38:43of 4.2 minutes per product
6:38:46this will be the target for the
6:38:47production line
6:38:50the final technique is hijanka which
6:38:52means production leveling and smoothing
6:38:55it is a technique to reduce waste
6:38:58occurring due to fluctuating customer
6:39:00demand
6:39:02let us understand the concept of cycle
6:39:04time reduction in this screen
6:39:07cycle time reduction refers to the
6:39:08reduction in the time taken for a
6:39:10complete process
6:39:12implementing lean techniques reduces
6:39:15cycle time
6:39:16and releases resources faster than any
6:39:19other method
6:39:20low cycle time increases productivity
6:39:23and throughput
6:39:24lean techniques release resources early
6:39:27achieving more production with the same
6:39:29machinery
6:39:30internal and external waste is reduced
6:39:33and the operational process is
6:39:35simplified with a decrease in product
6:39:38damage
6:39:39all these factors help in satisfying the
6:39:41customer and staying ahead in
6:39:43competition
6:39:44the following screen describes the
6:39:46concept of cycle time reduction through
6:39:48an example
6:39:50the changes brought by implementing lean
6:39:53techniques on an existing process are
6:39:55illustrated in the given diagram things
6:39:58to be noticed are number of operators
6:40:01used work allocation to the operators
6:40:04path or the movement in the process and
6:40:06flow of the process
6:40:08notice the changes brought about by
6:40:10implementing lean techniques on the old
6:40:12process
6:40:13first the path followed by the material
6:40:15in between the process is considerably
6:40:18reduced
6:40:19this decreases the cycle time for the
6:40:21entire process
6:40:23second the number of operators is
6:40:25reduced to 3 when compared to 5 in the
6:40:28old process
6:40:29operator 1 can now work on process 1 and
6:40:32process 4.
6:40:34similarly operator 2 can work on process
6:40:372 and process 3.
6:40:39there is an increased productivity of
6:40:41the operators and the remaining skilled
6:40:43operators can be used in some other
6:40:45process or system
6:40:48the next screen will introduce the
6:40:49concept of kaizen and kaizen blitz
6:40:53kaizen means good change in japanese
6:40:57kaizen is a continuous improvement
6:40:59method to improve the functions of an
6:41:01organization
6:41:02the improvements could be in process
6:41:04productivity quality technology and
6:41:07safety
6:41:08it brings in small incremental changes
6:41:11to the process
6:41:12kaizen blitz is known as kaizen event or
6:41:16kaizen workshop if the event is tightly
6:41:19defined and the scope is evident for
6:41:21implementation
6:41:22processes can be easily changed and
6:41:25improved
6:41:26teams could improve processes through
6:41:28creative problem solving methods in
6:41:30structured workshops over a short time
6:41:32scale
6:41:33the next screen will provide the
6:41:35differences between kaizen and kaizen
6:41:38blitz
6:41:39the differences between kaizen and
6:41:41kaizen blitz are
6:41:44kaizen is a method that brings
6:41:46continuous improvement in the
6:41:47organization
6:41:49while kaizen blitz is a workshop or an
6:41:51event that brings in change
6:41:54kaizen brings in small incremental
6:41:56changes in the organization
6:41:58there are no major changes made within
6:42:01the processes
6:42:02kaizen blitz is applied when a rapid
6:42:05solution is required
6:42:07the kaizen method follows a step-by-step
6:42:09process
6:42:11it standardizes measures and compares
6:42:13the process with the requirement before
6:42:16improving it
6:42:17kaizen blitz plans for the event
6:42:20executes it arrives at a solution and
6:42:22follows it through
6:42:24all the people of the organization are
6:42:26involved in kaizen whereas kaizen blitz
6:42:29is led by the top management and others
6:42:31are invited to participate
6:42:34the decision making lies with the upper
6:42:36management
6:42:37in kaizen the process is standardized
6:42:40and measurements are regularly collected
6:42:42and compared before the decision is
6:42:44taken this relatively delays the process
6:42:48of decision making
6:42:49in kaizen blitz decisions are taken soon
6:42:53and the process change is wrapped in
6:42:55three to five days
6:42:57kaizen is a continuous improvement
6:42:59method whereas the kaizen blitz is part
6:43:02of the improving process
6:43:04kaizen follows pdca in essence plan do
6:43:08and check and act for the improvement
6:43:10process
6:43:11kaizen blitz uses pdca for execution
6:43:15where the events are planned conducted
6:43:17decided implemented and followed up
6:43:21the following screen will elaborate on
6:43:22the concepts of kaizen and kaizen blitz
6:43:25through examples
6:43:28kaizen and kaizen blitz are practiced in
6:43:30many organizations across the world
6:43:33the examples of kaizen and kaizen blitz
6:43:36method are shown here
6:43:37click each tab to know more
6:43:41the toyota production system is known
6:43:43for kaizen practices
6:43:45in toyota if any issue arises in the
6:43:48production line the line personnel sees
6:43:51all the production until the issue is
6:43:53resolved
6:43:54once the solution is implemented the
6:43:56team resumes the production cycle
6:44:00a wood window company in the state of
6:44:02iowa us uses the kaizen blitz method to
6:44:06redesign their shop floor and replace
6:44:08expensive non-flexible automation with
6:44:11low-cost highly flexible cellular
6:44:14applications
6:44:15eliminating scraps reorganizing work
6:44:18areas and reducing inventory are some of
6:44:20the examples of quick implementation
6:44:23through kaizen blitz
6:44:25the term lien refers to creating more
6:44:27value to customers with fewer resources
6:44:30it means reducing unwanted activities or
6:44:33processes that do not add value to the
6:44:35product or service for the customer the
6:44:38lean philosophy is to provide perfect
6:44:40value to the customer through a perfect
6:44:43value creation process that has zero
6:44:45waste
6:44:46while the ultimate goal is to achieve
6:44:48zero waste you may not always get that
6:44:51in the first couple of tries
6:44:53however you will achieve minimum waste
6:44:55and continue to move towards zero waste
6:44:57eventually
6:44:59hence lean is the path towards
6:45:01perfection
6:45:03lean is about optimizing the process
6:45:05from beginning to end
6:45:06eliminating non-value-adding activities
6:45:09in vas and increasing flow to ensure
6:45:12that parts and services are provided to
6:45:14customers more quickly
6:45:16if quality is the word to describe six
6:45:18sigma then speed is the word to describe
6:45:21lean
6:45:22let's understand the importance of lean
6:45:26there are many benefits of lean and some
6:45:28of them are reduced cost reduced cycle
6:45:30time
6:45:31more throughput and increased
6:45:33productivity
6:45:35despite all of these benefits lean is
6:45:37not implemented by most of the
6:45:38organizations because of the
6:45:40misconception that it is only suited in
6:45:42manufacturing areas
6:45:44the reason for this misconception is the
6:45:46beginning of lean it began and grew in
6:45:49popularity in the manufacturing areas
6:45:51starting with toyota
6:45:53in recent years one can notice more
6:45:56applications of lane in other areas such
6:45:58as healthcare and the transactional
6:46:00space however the truth is that lean
6:46:04concepts can be applied in any business
6:46:06and in any process
6:46:08on the next screen let's discuss how
6:46:10lean and six sigma combine
6:46:13lean and six sigma are two different
6:46:15principles or methodologies that combine
6:46:17to form and create one powerful
6:46:20continuous improvement methodology
6:46:23they have various overlapping goals
6:46:24toward the improvement with the aim of
6:46:27creating the most efficient system
6:46:29though the approaches are different the
6:46:31methods complement each other
6:46:34lean six sigma takes the power and rigor
6:46:36of six sigma methodology and combines it
6:46:39with lean concepts leading to faster
6:46:41results better quality and improved
6:46:44customer satisfaction
6:46:47let's look at the differences between
6:46:48lean and six sigma
6:46:50lean focuses on efficiency by
6:46:53identifying value from the customer's
6:46:55point of view removing unnecessary steps
6:46:58in the process and improving process
6:47:00speed or velocity
6:47:02on the other hand
6:47:04six sigma focuses on effectiveness with
6:47:06the help of breakthrough processes
6:47:09identifying root causes
6:47:11and reduction in variation
6:47:14therefore when six sigma is combined
6:47:16with lean it is possible to achieve
6:47:18business transformation
6:47:20so remember lean is about speed with a
6:47:23focus on efficiency six sigma is about
6:47:26quality with a focus on effectiveness
6:47:28and lean six sigma brings the best of
6:47:30both together
6:47:32to yield a better result first implement
6:47:35lean to streamline the process
6:47:37this helps to understand the chronic
6:47:38problems and the ways to handle them
6:47:41quickly
6:47:42once the problem is identified use six
6:47:45sigma methodology to analyze the issues
6:47:48and provide business improvement
6:47:49transformation
6:47:50in other words lean is used to reduce
6:47:52the waste and six sigma is used to
6:47:55reduce the variation
6:47:57thanks to our training experts it was a
6:47:58real pleasure having you with us
6:48:01and with that we have come to an end of
6:48:02this full course video on six sigma i
6:48:05hope it was informative and interesting
6:48:07if you have any questions related to the
6:48:09topics that we covered in this video
6:48:11please ask away in the comment section
6:48:13below our team of experts will be
6:48:14delighted to hear it from you thanks for
6:48:16watching stay safe and keep learning
6:48:21[Music]
6:48:23hi there if you like this video
6:48:25subscribe to the simply learn youtube
6:48:26channel and click here to watch similar
6:48:29videos turn it up and get certified
6:48:31click here

About this transcript

This page contains the full transcript of Six Sigma Full Course in 7 Hours | Six Sigma Green Belt Training | Six Sigma Training | Simplilearn by Simplilearn, generated from the public captions YouTube serves with the video. The transcript has 56,888 words across 10,288 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.