YT YouTube2Text

YouTube2Text → Transcripts → Persiapan UAS Metodologi Penelitian | Buddy Tentir | Buddy Program ASP oleh Bu Nur Indah Lestari

Persiapan UAS Metodologi Penelitian | Buddy Tentir | Buddy Program ASP oleh Bu Nur Indah Lestari — Transcript

by Buddy Program ASP · 10,672 words · 1,639 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:01Can you see it?
0:03>> Yes,
0:04>> Ma'am.
0:05>> Okay. Alright. Okay. This um you have
0:13150 minutes, 2.5 hours, right? So make
0:18sure you answer them well. There are
0:23six questions in Quanti. Then in Quali,
0:28there are five yeah. 1 2 3 4 5. There
0:35are five questions. Now, the attachment
0:37is long. So, whether you like it or not
0:42, you have to understand what it
0:45contains. Mm I cannot explain
0:56everything from the start. Why? Because
1:00when talking about quantitative results
1:03, it inevitably depends on your
1:05experience working with Stata.
1:10Meanwhile, as far as I know, you in ASP
1:14don't have econometrics, and in data
1:21analytics, what do you use with Mr.
1:23Agung? What do you use?
1:26>> We use Orange, Ma'am.
1:27>> Okay, you use Orange with Mr. Agung. My
1:29former boss. That's why he even said, "
1:34Mbinda, just teach data analytics." Hmm
1:38. We'll see, Sir, it depends on the
1:41assignment. So, not everyone has
1:46experience with Stata. That's why if
1:51you want to know more, you'll have to
1:55try using Stata for the data you use in
1:58Orange so you can get familiar with it.
2:05Because if you don't, you won't get it.
2:09The questions from the last two years
2:12might not appear again, and they
2:15probably won't be like this either
2:18because if they were, they wouldn't be
2:22approved. There is still a coordinator,
2:28I suddenly forgot. So, the coordinator
2:32has to sign off that these questions
2:34haven't been given before or in class
2:37like that. Okay, I'll start from the
2:42first one; if there's anything you
2:44don't understand, feel free to ask
2:47right away. Because I don't know how
2:52far your understanding goes. Okay. Now,
2:58in attachment A, there is a long output
3:02. Should we discuss the questions first
3:05or look at the attachment first?
3:15>> I think we could discuss the questions
3:17first, Ma'am. And look at the
3:18attachment simultaneously.
3:20>> Okay. Alright. Because this, these are
3:24at 13, 14, 15, that's the line number
3:27where it's used, right. Okay, down here
3:31is a long attachment from beginning to
3:33end. I'll scroll down for a moment so
3:37we can go through it. Now, this is it,
3:40right. Um the first thing you can
3:45cross-check is that you can read this.
3:48This is Podes data. What is Podes? Does
3:57anyone know? Most likely village
4:03potential. Is that it? Now, it has
4:08these main variables. There's
4:10malnutrition, there's poor. Although I
4:13don't know if this "poor" is the number
4:16of poor people or the poverty rate.
4:22Because there's no description here.
4:24There's KUD, there's micro, there's
4:28savings. Then what is this "others"? I
4:30don't know either. And even if you
4:33don't know, that's natural because the
4:35question itself already had, uh, what
4:38do you call it? Uh, below it, it
4:40immediately went 13, 14, 15, right?
4:43Wait, it seemed like someone asked
4:44earlier. Oh, the Research Methodology
4:56final exam question. Okay, I thought
4:59those of you who don't have it yet were
5:00asking. Then, uh, total cooperatives,
5:04there are KUDs, micro, savings, and
5:06others. It seems this is talking about
5:08cooperatives and then there are the
5:10costs. Then there is the gen ln
5:13malnutrition, there is gen ln. Ln, for
5:18those of you who don't know, ln is
5:20natural logarithm, excuse me. Now,
5:25natural logarithm is actually a
5:28logarithmic form of a number. What
5:31numbers are used here? The numbers used
5:34are malnutrition, the numbers used are
5:37poverty data, total cooperatives, and
5:40total cooperatives squared.
5:48>> What do you call it? From this, it
5:50means we have this. Now, the question
5:54above in 13 to 14 is, why does it use
5:58ln? Why does it need to be
6:00ln-transformed? Here is the first one.
6:07Why does data need to be ln-transformed
6:09? Usually, it's not at the beginning.
6:13So, in terms of process, data is
6:16ln-transformed if later, after being
6:18regressed, it has a problem. There is a
6:23problem in the classical assumption
6:25test called normality or, uh,
6:27heteroscedasticity. That is, if there
6:33is a problem. Second, for example,
6:36there's no mention of a problem, but
6:38it's still using ln. Why? Because we
6:41want to see the elasticity. So, using
6:47ln, uh, there are several reasons. The
6:51first one earlier was because after
6:53regressing with the original data, it
6:54turned out to have problems in the
6:56classical assumption test. Now, for
7:00those of you who don't know, there are
7:02four classical assumption tests. There
7:06is normality, there is
7:08multicollinearity, there is
7:13heteroscedasticity. The last one is
7:16autocorrelation. Uh, for example, if
7:22all assumptions are safe or haven't
7:25been tested yet, why still use ln? Why
7:28is that? It's possible. The second
7:30reason is because what we want to see
7:33is its elasticity. So, we want to know
7:38which one is more elastic, you still
7:44remember economically, uh, there are
7:47elastic and inelastic ones, and that
7:51matters to the government. Why? Because
7:55if, for example, an item's behavior is
7:58inelastic, whatever booster the
8:01government gives won't have an effect,
8:06because it is inelastic. But if it is
8:14elastic, once boosted, it can jump. The
8:20economic growth could potentially
8:22skyrocket, for example like that. That
8:25is the second one. Or the third one,
8:28why use ln? Because, uh, the data from
8:31all the data used has too far a
8:34distance. Distance is range. The others
8:40, uh, the numbers, for example, are
8:43percentages only from 0 to 100 or
8:45ratios, even only from 0 to 1. Now,
8:51here is data that reaches millions like
8:53, for example, assets or the data is,
8:59uh, in the thousands. These are very
9:04different, like that. That is the
9:06reason why we use ln, uh, why it is
9:09logarithmically transformed. Well, but
9:11here, the thing being asked isn't about
9:14the logarithm itself. What's being
9:16asked is why use + 1 before taking the
9:22logarithm. First, can all data, all
9:27variables be log-transformed? Actually,
9:32yes, but it is not recommended for data
9:35that is already in ratio or percentage
9:38form. That's one thing. Second, for
9:42data that has a value of 0 or is
9:45negative, like that. So, if asked why
9:54add the number 1 before calculating the
9:57logarithm? Most likely because
10:01malnutrition, poverty, and total
10:03cooperatives have zero values in them.
10:08Why? Because the logarithm of 0 has no
10:12result, you see. So, from that, the
10:25reason is so that it can get a value
10:32that represents that. That's the first
10:38one. Second, no data is lost. It
10:45doesn't mean "lost" as in you deleted
10:48it, but it could be deleted by Stata
10:51automatically. If there's no plus 1,
10:55for example, gen len malnutrition
10:57equals ln open-bracket malnutrition. If
11:03it turns out there's zero data, after
11:05line 13, it will surely show something
11:08like 10 missing values. What does that
11:12mean? It means there will be 10 pieces
11:15of data whose malnutrition content is
11:18gone. Now, when it's gone, it becomes a
11:23missing value. It could mean it no
11:26longer depicts the whole picture, like
11:33that. So, in terms of data analysis,
11:39it's better to use plus one. That is so
11:42that the data can be representative.
11:50Because, for instance, in total
11:51cooperatives, there's one region that,
11:54oh, it turns out it has no cooperatives
11:56, so the sum result is still zero, well
11:58, that's just how it is. Most likely,
12:02this quantity—well, not quantity, the
12:04minimum value is definitely zero.
12:07Because if it's negative, adding 1
12:10might not be enough yet. So it's most
12:14likely because the data contains zero
12:16values. Any questions so far? If there
12:33aren't any, what are the implications?
12:41What are the implications? Uh,
12:46statistically, what is it? If there's a
12:53+ 1, it means we are shifting by one
12:58point. Everything increases by one. The
13:03minimum value should be 0. Then finally
13:07, the minimum value becomes one because
13:10of the plus. And usually, statistically
13:19, the most visible implication is that
13:25the data increases by one. So if the
13:32original data, for example, the total
13:35cooperatives was 100, after adding 1,
13:39the total cooperatives becomes 101. So,
13:45statistically, that's how it is
13:48regarding the zero value so it can be
13:52log-transformed. Then what else?
13:57Because usually, when adding one, how
14:03do I draw it? Do you guys know the bell
14:09curve? The one shaped like this. Now,
14:17when there's a + 1 or a 0 value,
14:20usually the curve isn't symmetrical in
14:23the middle like this, but it will be
14:26more like this. Like that. It is skewed
14:37to the right. I hope I'm not drawing it
14:43wrong. If I draw it wrong later, feel
14:45free to correct me. I don't have a
14:47problem with that. Okay. So that is one
14:53of, what do you call it, statistically,
14:58the data that was more right-skewed
15:03changes because of that plus one. Now,
15:13what about the interpretation? What is
15:16meant by substantive? Substantive is
15:18basically interpretation, right? Well,
15:21interpretation is more towards when you
15:24interpret the natural logarithm itself.
15:29Why? Because, as I said earlier, why
15:32use a natural logarithm? Because we
15:35want to talk about elasticity. Now,
15:37when talking about elasticity, it
15:39depends on the coefficient. Whether the
15:42coefficient is greater than 1, equal to
15:451, or less than 1, like that. So, in
15:58terms of transformation, the
16:01statistical implication is not just
16:04from the plus sign, but the statistical
16:07and substantive implications of the
16:10natural logarithm, like that. So,
16:15earlier, statistically, the natural
16:17logarithm was more towards the data,
16:23and then more towards the data
16:25distribution related to normality. Then
16:31, from the heteroskedasticity side, it
16:34means the variance is diverse, which is
16:38why it is natural logged. Now,
16:41substantively, it is more towards the
16:43interpretation. Because if the
16:47coefficient of a variable has a "ln" in
16:50front of it, whatever the unit is—for
16:53example, total cooperatives, the unit
16:56would be, say, five cooperatives. Now,
17:01once you use "ln" total cooperatives,
17:03the unit becomes a percentage. That is
17:08why what is often read, or what is
17:11often used, is elasticity. Is there
17:16anything you want to ask before I
17:18proceed to number two? None. If there
17:28isn't, I will go to number two. So,
17:33there is output below, from 23 to 24,
17:37there is global Y malnutrition, global
17:41X ln poverty, and global Z. Okay. Now,
17:53what does this mean? Explain the
17:55regression equation form and the
17:57command. Which one is the regression
17:59equation? The regression equation is
18:03the one at number 33, the RE. But the
18:07global command means it creates a
18:11symbol; instead of typing "ln
18:13malnutrition," it's better to type Y.
18:20So, global replaces this malnutrition
18:23variable with the symbol Y, like that.
18:30Why? Because Stata is case-sensitive.
18:35You might have written "ln malnutrition
18:37" correctly. But if you use a capital
18:40L. Then a red command will surely
18:42appear: variable not found. Why?
18:49Because the L is a capital L. So, to
18:51overcome typos between capital and
18:53lowercase letters, the global command
18:56is used. So, it has Y as malnutrition,
19:01X as poverty, ln poverty, Z1 as total
19:04cooperatives, and Z2 as total
19:06cooperatives squared. Then C, what is
19:12CTRL? Most likely, my assumption is
19:14that this control means, what is it?
19:17Control variable. The control variable
19:19is ln total schools. Now, what does the
19:27regression become? Regression of Z1 on
19:31X and Control. So, who is the dependent
19:36variable? Z1. Why use the dollar sign?
19:41Yes. If you use a global name, or use a
19:45global command, when calling a variable
19:47, you need to write a dollar sign in
19:50front of it. So, you cannot just write
19:55Z1 X Ctrl; it won't be found. Stata
20:01will later say variable not found.
20:03Because in the variable section, there
20:05is no Z1 variable; there is a
20:07ln_total_cooperatives variable. There
20:11is no X variable; there is a ln_poor
20:13variable. Why did it change to this?
20:17Because above, it has been given
20:20symbols; ln_poor is symbolized by the
20:23letter X, and total cooperatives by Z1.
20:31Like that. Then, describe the meaning
20:34of each variable in the context of the
20:36relationship between cooperatives,
20:37poverty, and education on malnutrition
20:39rates. Well, you just need to connect
20:44these according to the knowledge you
20:48have. What is the relationship between
20:50cooperatives and malnutrition? What is
20:52the relationship between poverty and
20:53malnutrition? What is the relationship
20:56between education and malnutrition?
20:57Like that. Does anyone want to try
21:01explaining the relationship between
21:03cooperatives, poverty, and education?
21:08>> Uh, excuse me, Ma'am. Does this mean
21:10the relationship in general terms,
21:11Ma'am? We don't need to look at the
21:13estimation results from Stata down
21:15there? No, not yet. This is just for
21:22each variable. So, honestly, if I call
21:26this your prior knowledge. What, in
21:29your opinion, is the relationship
21:31between cooperatives and malnutrition?
21:35Oh, if it turns out the relationship is
21:37A, then fine. Later, the Stata results
21:41will confirm whether or not it matches
21:44what you predicted. Simply put, it's
21:49like you are trying to state your
21:51hypothesis. Although in the context of
21:55research methodology, this should be
21:57constructed. Constructed means there
22:01are foundations, literature books, or
22:04journals for it. But this is an exam,
22:10you can't open books or look for
22:12journals, right? So, it's based on the
22:15prior knowledge you have. How do you
22:19think the relationship is? Does anyone
22:20want to try? Cooperatives, poverty,
22:23education?
22:25>> Uh, maybe I'd like to try, Ma'am. If
22:27>> it's allowed, please go ahead.
22:29Cooperatives should be able to reduce
22:31malnutrition rates because the
22:33assumption is that they can buy
22:36>> nutrients at the cooperative. Then, if
22:39>> okay
22:41>> Poverty can increase malnutrition,
22:43assuming that the residents are unable
22:47to buy healthy, nutritious food. And
22:51education can reduce malnutrition rates
22:54because when a family has education
22:57about what food is good or healthy,
23:00they can choose food that can avoid
23:03that malnutrition.
23:06>> Okay, nice. Like that. So, you just
23:10need to provide your explanation. Don't
23:15just say cooperatives increase or
23:19education decreases, for example, but
23:23explain it like what Azani just said,
23:27why that is the case. Okay, I will
23:32continue. Number T. Oh, this is on a
23:35separate page again. Okay, here we have
23:40the command estat hettest. Now, for
23:45those of you who don't know, earlier I
23:47mentioned that you should have all
23:52studied statistics, right? When you get
23:55regression results from your statistics
23:58, can you use them right away? Actually
24:01, no. Because you need to satisfy the
24:04classical assumption tests first. As I
24:08said at the beginning, there are four:
24:11normality, multicollinearity,
24:13heteroskedasticity, and autocorrelation
24:16. Now, the estat hettest command here
24:22is used to test for heteroskedasticity.
24:31It says Breusch-Pagan and Cook-Weisberg
24:33, who are the people that developed the
24:36test formula. Where is the result? The
24:41result is right down here. There. So,
24:46when you are asked to explain what is
24:48being tested, this means we are testing
24:51for heteroskedasticity. What is the
24:55hypothesis? The hypothesis is H0 is
25:00homoskedasticity, or no
25:03heteroskedasticity. H1 is
25:07heteroskedasticity. Let me repeat that.
25:14H0 is homoskedasticity, or in other
25:18words, no. No means there is no
25:21heteroskedasticity. Meanwhile, H1 is
25:26heteroskedasticity, or there is
25:28heteroskedasticity. So, the next
25:36question is, based on the probability
25:38value, which one is the probability?
25:40The one down here. Based on this value,
25:44what is the conclusion of the test
25:46result? Before reaching a conclusion,
25:50there is a decision first. Our
25:54threshold for H0 and H1 gives us only
25:57two choices: reject H0 or fail to
26:00reject H0. Now, the threshold is the
26:06alpha. The alpha commonly used is 5%.
26:120.00 means it is less than 5%. So, do
26:18we reject H0 or fail to reject H0?
26:21Reject H0.
26:23>> Okay. Reject H0, so the decision is to
26:26reject H0. What is the conclusion? If
26:30we reject H0, which one was true, H0 or
26:32H1? Earlier, the truth was H0 is no
26:42hetero. H0 is no hetero, right?
26:46>> H1 is there is hetero. From the result
26:51of 0.00, the decision was to reject H0.
26:55So, does that mean the conclusion is
26:57there is hetero or not?
27:00>> There is hetero.
27:01>> Yes.
27:02>> Exactly. There is hetero. So, what is
27:07the conclusion then? Distinguish
27:08between the two. Don't answer "reject
27:10H0" when you are asked for a conclusion
27:12. That is a decision, not a conclusion.
27:17The conclusion is that the model still
27:21contains heteroskedasticity, or it
27:24failed the heteroskedasticity test.
27:28What is the implication for the
27:30regression model being used? The
27:32implication. Well, I have to explain
27:36heteroscedasticity then. Okay. Uh, what
27:41is it called? Long story short, hetero
27:45is, uh, expecting from the model that
27:53it has the same bell shape. Earlier,
28:00those bells can be skewed to the right,
28:05skewed to the left, flat, or very steep
28:09. If I draw it, for instance, this is
28:14the normal one, but there are some that
28:18can be flat like this. Like that. Some
28:25drawings can be very steep. Is that
28:28possible? It is. So, if in one dataset
28:33you have all three, it means it is
28:36heteroskedastic. It is not uniform.
28:42Well, the expectation in a model is
28:44that it is uniform. It doesn't have to
28:47be uniformly flat, or middle-ground, or
28:49very sharp. No. It doesn't have to be
28:50like that. Any of them is fine. For
28:53example, if you want the data to be
28:55like this, that's okay. As long as they
28:57all lean to the right like this, that
29:03is called homoskedasticity. So, what
29:09happens to the model? In the model, it
29:13means, uh, it is diverse, the variance
29:20between, uh, the error and the
29:24variables. Then what happens to the
29:31model if it still contains
29:33heteroskedasticity? What happens is
29:40when uh, what do you call it? A model
29:46always has what is called a standard
29:48error. Now, the standard error usually
29:51becomes more biased. Like that. Now,
29:58when the standard error is biased, it
30:01means most likely the coefficients are
30:06not not stable, or rather the
30:11coefficients cannot be used because
30:13they are still biased, so the
30:15interpretation of the coefficients
30:18becomes, uh, ambiguous or could even be
30:21inconsistent. with, uh, reality as such
30:27. That is why it is necessary to, uh,
30:34check the autocorrelation results, the
30:38heteroskedasticity results, because if
30:41those are not met, then, uh, the
30:44regression needs to be fixed first.
30:49First, the error usually becomes larger
30:51. Second, uh, the coefficients cannot
30:56be used directly. Usually those two
31:00things. What is visible, uh, not always
31:10, but usually, uh, what is visible is
31:13that the R-squared is small, usually.
31:20Like that. So even if, for instance,
31:26the coefficients are interpreted, when
31:28applied to the real world, they become
31:31misleading. That is why, as I said,
31:35because the error is too large, it
31:38becomes ambiguous or misleading, like
31:40that. Is there anything you want to ask
31:45before I continue to number 4? Okay, if
32:07there isn't, I will continue. Number
32:10four, this, uh, comment above has a
32:13comment like this. Why is the global
32:18one always shown when the intention was
32:21only to ask about this regression? Yes,
32:24because we need to know who Y is, who
32:27Z1 is, who Z2 is, and who the control
32:31is. Now, from this equation, the
32:34question is which one undergoes a
32:36quadratic transformation. Which one
32:41undergoes a quadratic transformation?
32:43Does anyone know? Ln total cooperative
32:554.
32:57>> Which one?
32:58>> The ln total cooperative 2 that is
33:00squared.
33:01>> L I, okay, ln total cooperative 2,
33:03which is Z2, right?
33:05>> Yes, Ma'am.
33:06>> Yes, because Z2 is ln total cooperative
33:09squared. How do you know? You can look
33:12below or up here earlier, here it is,
33:14here it is. Z2. Above, uh, below
33:24earlier. Then write down the regression
33:28equation based on the estimation
33:31results in stage 2A. So what are these
33:33results? If the question is, uh, the
33:38regression based on these estimation
33:41results, it means you have to go down,
33:44go to the results on line 45. Where is
33:48line 45? This one. This is what you
33:52write. So which is the regression
33:55equation? The one right here. What is
34:00the equation? Ln malnutrition equals
34:040.29. Uh, this goes back to your
34:07respective lecturers, yes. Your
34:09lecturers prefer how many decimal
34:11places. If, for example, two digits,
34:15then uh 0.3 because this is 29, if
34:18rounded it becomes 0.3. Then plus 0.007
34:28or to keep it all three digits, that
34:31means ln malnutrition = 0.297 + 0.007*
34:37ln total cooperatives + 0.06 8 or 069
34:44ln total cooperatives squared + 0.0098
34:51that means 0.01 ln total schools. This
34:56is the equation that answers number 4.
35:06I'll go back up first here, yes. Here,
35:11based on the regression coefficients
35:13obtained, what is the shape of the
35:15squared variable curve? So what is the
35:18shape of the curve? Earlier, was
35:24variable Z2 positive or negative?
35:34>> Positive, Ma'am.
35:35>> Now, if it's positive, if asked what
35:38the image looks like for a quadratic,
35:41the image must be a parabola. It just
35:47depends on the coefficient. If the
35:48coefficient is positive, does that mean
35:50the parabola opens upward or downward?
35:54Now, try to remember your high school
35:56lessons when you learned quadratic
35:58equations. Come on, if it's positive,
36:00does it open upward or downward?
36:10>> Upward, Ma'am. Like a U shape.
36:12>> Okay. U shape. Okay, that's it. Then
36:20what's the next question? Wait, let me
36:22go back to number 4 here. Explain the
36:28shape of the curve in the context of
36:30the relationship between the number of
36:32cooperatives and the level of
36:33malnutrition. So, what does that mean?
36:44Y is malnutrition, right. So if you
36:46have a Cartesian coordinate axis, you
36:50will have something like this. This is
36:55malnutrition this is Y, and here is Z2
37:02like that. If it has a shape like this,
37:06then what does it mean in the context
37:08of the relationship between ln total
37:10cooperatives and uh malnutrition? So,
37:19uh, what is the condition of the
37:21cooperatives? If the number of
37:23cooperatives increases, what is the
37:28addition of the number of cooperatives
37:30to Y like? Here, it adds one point, it
37:34goes down by the amount of its power
37:42until it reaches the lowest point here,
37:48at what point? Now, what is the
37:53implication of this relationship shape
37:55for cooperative development in reducing
37:57malnutrition? So what needs to be done?
38:02What do you think if the cooperatives
38:05are left to run on their own? Uh, one
38:08moment. Okay, sorry, uh, I'm reminding
38:28my child to perform Asr prayer first.
38:33Now, this is what you need to
38:34cross-check. Because why? If the
38:37cooperatives are left to run on their
38:39own until a certain stage, they will
38:45reduce it, right? like that. Reduce
38:52malnutrition, but at another stage,
38:54they actually increase malnutrition.
38:56Right? So what must you have? You want
39:01more cooperatives, but malnutrition
39:03also rises. Now, that goes back to you
39:08on how to explain it okay. Both are
39:13mandates. like that, yes. So, uh, look
39:19at it visually. So it doesn't just stop
39:23at the uh coefficient. Oh, the
39:28coefficient is 0.007 for example. I
39:31don't remember what it was earlier. Oh,
39:34well, if the cooperatives increase,
39:37then the malnutrition will also
39:39increase by 0.007, like that. Now, how
39:45should we look at this first? Because
39:49if we're talking about total school
39:50cooperatives, they have a behavior like
39:51this. Even though when there is a
39:55quadratic like this, it cannot be
39:58discussed on its own, right? Why?
40:02Because there is a total of
40:03cooperatives that is not squared. So,
40:08it means we have ax² + bx + c, right?
40:15Well, those cannot be discussed
40:16separately, right? They must be
40:17discussed together. Both must be
40:21addressed because they are both total
40:22cooperatives. Linearly the effect is
40:26like this, non-linearly the quadratic
40:28effect is like that. So, you must have
40:35an idea of what aspect you want to
40:38discuss. If you want to limit it,
40:41roughly to what extent? Or, "Ma'am, we
40:45can't find the numbers because we don't
40:46have the data." Fine, then how will you
40:49present it here? Because you must know
40:53that, oh, at a certain point, when the
40:55number of cooperatives increases, it
40:57turns out malnutrition also increases.
41:01Okay, that's number four. Is there
41:03anything else you want to ask? If there
41:08isn't, I'll delete it. I'll move on to
41:11number 5. Now, this is number 5, there
41:17is a STA VIF command. What diagnostic
41:21test is performed with the STAT VIF
41:23command? As I mentioned earlier, there
41:28are 4 classic assumption tests, and VIF
41:31is for multicollinearity. Like that. So
41:38VIF is an assumption test related to
41:41multicollinearity. Why do we need to
41:45perform a multicollinearity test?
41:48Because we need to know if there is a
41:51relationship between X1, X2, X3, and X4
41:53. Is it not allowed, Ma'am? To have a
41:58relationship? It is allowed. But the
42:01relationship cannot be linear. That's
42:07why if you look down, let me scroll
42:09down first. It was 47, if I'm not
42:13mistaken. Now, this VIF, does it mean
42:16there is a relationship? There is a
42:19relationship. Because when there is a
42:22value, it means there is a relationship
42:25between the first total cooperatives,
42:27the second total cooperatives, and
42:29total schools. But the relationship is
42:33not not very linear or not very strong.
42:40That's why, where is the threshold? The
42:42threshold for VIF is 10. If the VIF is
42:5210, or if the VIF is above 10, then it
42:55has a multicollinearity problem. As
43:01long as it is below 10, it is free from
43:04multicollinearity problems or, in other
43:07words, it passes the multicollinearity
43:10test. So, is there still a relationship
43:14? Yes. But the relationship is not, not
43:17very strong. Like, for example, you
43:20guys, uh more than 20 people, for
43:24instance, you have a relationship with
43:26each other, right? I have a
43:28relationship with you, right? But it's
43:30not very strong. Roughly speaking, if,
43:35uh, one friend leaves, not everyone
43:37else follows and leaves, right? Well,
43:42it's that simple. It's different if the
43:46relationship is very strong, then if
43:48one friend leaves, everyone leaves.
43:51Well, that means the relationship could
43:54be more than 10 in terms of VIF. Then
44:03what else? Wait, let me scroll up. Oh,
44:07I missed it. Based on the VIF values
44:11shown, how do you assess whether or not
44:13there is a testing problem in the model
44:15? Earlier I already told you that for
44:18VIF, the threshold is 10. Like that. So
44:22, from 10, that means Because the
44:28average result was three, looking at
44:30the mean is fine, or if you want to
44:32look at each variable individually, you
44:34are welcome to do so. But if you want
44:38to see it quickly, you can just look at
44:40the average. Since the average is three
44:42, it means it is below 10. This means
44:45there is no multicollinearity issue in
44:48the proposed model. Explain your
44:52reasoning. Yes, you need to explain it.
44:56Uh, is there a relationship between
44:58total schools and total schools squared
45:00? Yes. But the relationship is
45:02quadratic, not linear, right? Yes,
45:03isn't it? Because total schools is x,
45:07and total schools squared—uh, total
45:10schools 2—is x². Meanwhile,
45:12multicollinearity becomes significant
45:14if the relationship is linear. Simply
45:18put, if you are told to do A, you do A;
45:20if you are told to do B, you do B. Oh,
45:22directly. Well, that becomes very
45:25linear. So, explain that, and what was
45:28the other one again? Total schools. The
45:31other one. What was the other one? Oh,
45:41wait. That was wrong. Total
45:43cooperatives. Are total cooperatives
45:46and total cooperatives squared related?
45:48Yes, but the relationship is non-linear
45:50. Meanwhile, is there any relationship
45:52between total cooperatives and total
45:55schools? Yes. There is no relationship,
45:58Ma'am. Schools are schools, and
46:00cooperatives are cooperatives. Unless
46:01it were a canteen. If it were a canteen
46:03, every school might have one. So, like
46:05that. Well, that is what you need to
46:08write down, since you don't have a
46:10basis and weren't provided with journal
46:12attachments either. So, what is the
46:15logic you use regarding the
46:17relationship between total schools and
46:20cooperatives? If it's just with
46:23cooperatives, the relationship is
46:26clearly distant, Ma'am. For instance,
46:29there could be five schools in one
46:32village but only one cooperative. So,
46:39there is no relationship regardless of
46:41whether the number of schools increases
46:42or not. So, that is what you all need
46:48to explain. Then, explain the
46:50relationship between those test results
46:52and the reliability of interpreting the
46:53regression coefficients in this
46:58analysis. From the side of the
47:01regression results for
47:03multicollinearity, if it is already
47:06safe, then each variable is reliable.
47:10What does reliable mean? It means
47:13dependable, usable; it's okay to use
47:16directly because it has passed the
47:19classical assumption test. In other
47:25words, if you have heard of BLUE—Best
47:32Linear Unbiased Estimator—it means
47:35the model obtained is already BLUE.
47:41That's it. Why? Because, uh, if it
47:47doesn't pass the multicollinearity test
47:50, the model might still be BLUE, but,
47:53um, not quite. The downside of
47:59multicollinearity—if there is a
48:00multicollinearity issue—is that it is
48:02difficult to use for prediction. Why is
48:05it difficult to use for prediction?
48:07Because...He...Where? Okay. Earlier,
48:22after eating, he was asked to talk
48:24about animals. Just got back from the
48:26mosque, he said, "I ran, I was chased
48:28by a monitor lizard." Okay, very well.
48:32That's why he paused here. Sorry, that
48:35was an intermezzo. So, that is related
48:38to the coefficients. So, if it is safe
48:41from multicollinearity, it is safe to
48:45be inferred. What is inference? If you
48:48want to interpret it, it can be used
48:51directly. That's it. Why? Because, er,
48:57it's settled; there is a relationship,
48:59but it’s not pulling against each
49:01other. Simply put, that's it. The last
49:06one, okay. The last one, number 6. Er,
49:10please look at the combined regression
49:12table on lines 767. There are three
49:15regression models. 1, 2, 3. What can
49:19you what can you conclude? Okay. The
49:24influence of the poverty variable.
49:26Meaning the poverty variable, right. Uh
49:27, wait a minute. Is this a question or
49:29information? Okay, if there were
49:31symptoms in the heteroskedasticity test
49:34, they can be overcome by using natural
49:37logarithms. What if there are symptoms
49:39of multicollinearity? Is there a way to
49:42overcome it, Ma'am, besides removing
49:44one of the variables that has symptoms?
49:47If the case makes it hard to find a
49:50replacement variable, or if removing it
49:52leaves too few independent variables,
49:54causing the R-squared to become low.
49:57There is there is a way. Like that.
50:01There definitely is, er, er, what? But
50:04I can't give an example, er, what is it
50:07, Mas Fauzan. So, it's true what Mas
50:11Fauzan said, if there are symptoms of
50:13multicollinearity, the first solution
50:15is, er, not just removing the variable.
50:18The main solution that Mas Fauzan
50:20mentioned afterward is looking for not
50:24looking for a replacement variable, but
50:27looking for other indicators for that
50:29variable. For example, here we use,
50:32there's poverty, right. As I mentioned
50:34earlier, I don't know if this poverty
50:37is the number of poor people or the
50:39poverty rate like that. If you use the
50:44number of poor people and it turns out
50:46to be multicollinear, try changing it
50:47to the poverty rate, like that. That's
50:50the first one. The second, you can add
50:53other variables or, er, remove them.
50:57Why add or remove? Because it could be
51:00that once another variable is added,
51:03the multicollinearity isn't high, the
51:05relationship isn't high, like that. But
51:09if it turns out you can't, er, if you
51:10want to remove it, Ma'am, there are
51:12only three variables, if one is removed
51:13, only two are left, that's very little
51:15. Well, there is a solution called mean
51:18centering like that. Er, er, I don't
51:24know if you were taught this in
51:25Research Methods, because these kinds
51:27of solutions should be discussed in
51:29Research Methods, right. Because when I
51:32teach econometrics in the Master of
51:34Notary program, the solutions aren't
51:36taught in econometrics. The solutions
51:39taught are the same as what Mas Fauzan
51:40knows. If you have to look for other
51:44indicators, add variables, remove
51:45variables, like that. You aren't told,
51:48er, the others. Because the focus is
51:50really on you understanding the context
51:52first, like that. So if there is a way,
51:54there is something called mean
51:56centering, like that. But I can't, er,
51:58explain it because it would become too
52:00long-winded while we want to discuss
52:02this. But if you really want to ask or
52:05have a case that requires Mas Fauzan to
52:08ask, feel free, okay, to contact me.
52:11It's just, er, what I can give as a
52:14hint is that it's called mean centering
52:17. So you, like, er, earlier, for
52:20example, we have long poverty data here
52:22. Now, we find the average. After that,
52:26each piece of data is subtracted by its
52:28average. Then the result is what we use
52:33for the regression. So poverty that has
52:37been mean-centered, like that, then
52:40malnutrition that has been
52:42mean-centered, like that. I hope that
52:45answers your question, Mr. Fauzan. Okay
52:50, if that's clear, I'll move on to
52:52number 6. Number 6, earlier at line 77.
53:08Here it is. Can you guys see this? This
53:21uh, there are three results. What
53:26distinguishes them? What distinguishes
53:29this number 1 is whose Y is it? Its Y
53:33is poor coop. I don't understand why
53:39its Y is poor coop. It has to be read
53:43from the top. The title is poor coop.
53:48This is mal coop mal poor. We'll
53:50cross-check later whose Y this is. Wait
53:54, let's just go up to the top, at least
53:56the last one. This is ln malnutrition.
54:04Ln malnutrition. Where are the other
54:13variable results? This is stage 2A.
54:16This is also ln malnutrition. It seems
54:22to be the same. This is stage one. Oh,
54:27this is total cooperatives. This is
54:31stage 1. Whose Y is this? You guys need
54:40to cross-check, okay. Whose Y it is.
54:44This is the estimate store for stage 2A
54:46. This is the Y, right here in the
54:50corner, this Y, the one in this box, is
54:53ln malnutrition, the Y for stage 2A
54:56using quadratic OLS. Then there is
54:59another regression result here. This is
55:05ln malnutrition. This is 2 SLS. Ln poor
55:11and total schools. This is stage 2B.
55:16Are there any more regressions? This is
55:21not one. This is you have and it wasn't
55:24saved. Now, okay. So it's here, right.
55:31If you look here at the top, you can
55:33see. Uh, so, uh, if if you want to see
55:36where this actually is, ma'am? You look
55:41at the sequence here, in the results
55:44sequence, look for the one that says
55:46estimate store. Estimate store stage 2B
55:50. So this is in number 3. So what is
55:56stage 2B? Quadratic 2 SLS. This uses a
56:01star. This star is not a comment. This
56:05star is just to indicate that this is
56:07stage 2B, what is it talking about?
56:09Quadratic 2 SLS. So it's not just OLS,
56:16there is 2 SLS, there is 3 SLS as well,
56:19like that. So you look here, there's
56:22estimate store stage 2B. So whose Y is
56:26it? The Y is malnutrition. Because,
56:28wait, let me show you guys first.
56:32Estimate store stage 2A here. Whose Y
56:35is it? The Y is also ln malnutrition,
56:37so what distinguishes it? Stage 2A is
56:40quadratic OLS. Then where is the other
56:44one? Earlier there was an estimate
56:46store stage 1 label. Now, in estimate
56:54store stage 1, whose Y is it? The Y is
56:57the same. Uh, the Y is different. Sorry
57:02. Here, total cooperatives, this is the
57:05one that's different. This is total
57:09cooperatives. So when looking at this,
57:14you guys need to be careful because the
57:16previous question was how is the effect
57:19of poverty on each dependent variable,
57:22right? Now, the dependent variable is
57:27not poor coop, not mal coop, not mal
57:29poor, no. The dependent variable. You
57:34guys must look up to see what the
57:35dependent variable is earlier. In this
57:39first one, it turns out the dependent
57:42variable is—let me write it here—if
57:45I’m wrong, please remind me, I think
57:48it was cooperatives. Like that. Total
57:55cooperatives, if I’m not mistaken.
57:58Total cooperatives. There. Meanwhile,
58:03in the second and the third ones here,
58:06the Y is the same, which is
58:08malnutrition. I put this in the middle
58:11because they—oh my goodness, can this
58:13be moved? Hmm. My mistake, I meant to
58:19move this. So, I placed this in the
58:22middle so that for these 2 and 3, the Y
58:25is malnutrition. So, the Y is not
58:28cooperatives for poverty, not
58:29malnutrition for cooperatives, not
58:31malnutrition for poverty. You all must
58:34return them to their original state.
58:41Now, the question was, if I recall
58:43correctly—or if others want to remind
58:46me while I'm at it—how does poverty
58:48influence each of these dependent
58:50variables from these three results, if
58:53I'm not mistaken. So, how is it in the
58:59first equation? The first equation is
59:01this one, right? Which one is the
59:04coefficient? The coefficient is this
59:07one, 0.054 is the coefficient. So, if
59:13the hypothesis talks about influence,
59:14what is the influence? The influence is
59:16positive. Is it significant or not
59:19significant? How do you know? There are
59:22these stars. The stars represent the
59:29level of significance. It's noted here,
59:33if there’s only one star, it means
59:35it’s significant at 10%. In other
59:38words, it uses an alpha of 10%. If
59:42there are two stars, it’s significant
59:44at 5%. Like what we usually use in
59:47statistics, an alpha of 5%, you know.
59:51If there are three stars, it’s
59:53significant at 1%. So, the influence of
59:57poverty on total cooperatives is
1:00:01positive and significant. Like that.
1:00:09Then, in the second equation, can it be
1:00:11checked? No, it can't, because in the
1:00:14second equation, there is no poverty
1:00:16variable. In the third variable, what
1:00:24is the relationship or influence of
1:00:27poverty on malnutrition? So, poverty
1:00:31has a positive and significant
1:00:33influence, both at 1%. That’s it.
1:00:40What else is the question? Can anyone
1:00:45show it to me? Or read it?
1:00:47>> The question is, what can you conclude
1:00:50regarding the influence of the ln
1:00:52poverty variable on the dependent
1:00:54variable based on the coefficients
1:00:57shown in columns 1 and 3, Ma'am?
1:01:00>> And why are the ln poverty coefficient
1:01:02values different between the two
1:01:03columns even though the variables
1:01:05tested are the same?
1:01:07>> Okay, very well. Okay, first, how about
1:01:12the coefficient? When talking about
1:01:16coefficients, it comes back to
1:01:22interpretation. Here, it uses a natural
1:01:26logarithm. Meanwhile, the Y is total
1:01:29cooperatives. The Y above here for
1:01:32total cooperatives, is it also a
1:01:34natural logarithm? Wait, let me go up
1:01:38first. Because if the Y is also a
1:01:41natural logarithm, then it’s safe. If
1:01:45not, then I have to add another
1:01:47explanation. Oh yes, it’s the same,
1:01:50ln total cooperatives, both use ln.
1:01:52Malnutrition also uses ln. Wait,
1:01:56scrolling up first. So what does this
1:01:59mean? Oh, it means if poverty increases
1:02:03by 1%. How do you know it's a percent,
1:02:09Ma'am? Because it uses ln. So when you
1:02:12guys change the coefficient—wait,
1:02:15change the variables into natural
1:02:18logarithms, the unit automatically
1:02:20becomes a percentage. So a 1%increase
1:02:24in poverty will increase total
1:02:27cooperatives by 0.05%. Meanwhile, in
1:02:33the third equation, a 1%increase in
1:02:35poverty will increase or will raise
1:02:39malnutrition by 0.1%. Like that, so uh
1:02:51automatically the influence of poverty
1:02:54will be greater on malnutrition
1:02:59compared to total cooperatives right?
1:03:04One is 0.1%, this one, uh, the total
1:03:07cooperatives is only 0.05%like that. So
1:03:12it has more influence on malnutrition,
1:03:14the effect is greater. Then why is it
1:03:18different? Yes, the first one is
1:03:20different because the dependent
1:03:22variables are clearly different right.
1:03:26The first reason is because the
1:03:28dependent variables are clearly
1:03:29different. One's dependent variable is
1:03:32total cooperatives, the other's is
1:03:33malnutrition. If they were the same,
1:03:36wouldn't it be a bit strange instead?
1:03:40Why would it be strange if they were
1:03:41the same, Ma'am? Yes, because the first
1:03:44one is uh it's a regular OLS while the
1:03:54third equation is 2SLS. So, the
1:04:00variables here, how do I explain 2SLS?
1:04:06Uh in terms of in terms of equations,
1:04:132SLS is like, for example, you have a
1:04:15first equation. The first equation is
1:04:20from X to Y like that. Then you have a
1:04:27second equation from Y to Z like that.
1:04:36Then what do you cross-check? You only
1:04:40cross-check from Y to Z. But inside it,
1:04:44you indicate that Y equals X. That's
1:04:49why—wait, let me close this first. If
1:04:52you guys go up here the result for the
1:04:58second stage A where was it? Stage 2A.
1:05:03Now, this one, for the quadratic, it's
1:05:07just y xx, right? This is the dependent
1:05:11variable, and these here are the
1:05:13independent ones. Meanwhile, for stage
1:05:162B you guys can see the dependent is Y
1:05:23and the independent seems to only be
1:05:26the controls, it seems. But actually,
1:05:34there is X. Now, this poverty X is
1:05:37depicted as another equation, which is
1:05:41cooperatives and cooperatives squared.
1:05:48See. So actually, there are two
1:05:50regressions. You first regress poverty
1:05:54with X as cooperatives and cooperatives
1:05:56squared. After that, the result of
1:06:00poverty is used as an independent
1:06:02variable along with control variables
1:06:05to measure malnutrition C, like that.
1:06:09So it's like this, but the first one is
1:06:12implicit and not visible. That's why
1:06:15it's put in parentheses here like that.
1:06:22So the first one is indeed a different
1:06:25dependent variable. The second one is
1:06:28because—oh, missed it. The second one
1:06:31is because the method is different. The
1:06:34first one uses the Ordinary Least
1:06:36Squares method, OLS. The second one
1:06:38uses the Two-Stage Least Squares method
1:06:40. So it's like having two equations.
1:06:45Not like, it actually has two equations
1:06:48, only one of them is run implicitly.
1:06:52Like in Indonesian, it's like implied
1:06:55versus explicit. What's explicit is
1:07:02only Y influencing Z. This is the
1:07:04second equation. But it is implied
1:07:08within it that Y is the result of a
1:07:11regression from X xxx xxx like that.
1:07:14That is allowed. Done there. Number 6.
1:07:29Okay, if you are ready Should we move
1:07:34on to qualitative, or is there anything
1:07:36you want to ask first or anything from
1:07:38these results that you don't understand
1:07:40yet? Because you won't be asked about
1:07:50that later. Is there anything? There is
1:07:57. Okay. If there isn't, I'll just add
1:08:19what's here. What's here earlier was
1:08:23only vif for multicollinearity, then
1:08:26hettest for heteroskedasticity, and one
1:08:29more here, numbers 48, 49. This is used
1:08:34for the normality test, even though it
1:08:40can't be concluded, it can't. That's
1:08:43why it didn't become a question because
1:08:44there's no prob. Right. So, as I said
1:08:54earlier, there are four; the ones that
1:08:58haven't been discussed are normality
1:09:01and autocorrelation. Now, for normality
1:09:04, the command here is sktest. This is
1:09:09for the residuals, and then there will
1:09:10be the results. Does the command always
1:09:13have to be sktest? No. The command can
1:09:15be swilk, the command can be sfrancia.
1:09:23So there are several commands,
1:09:25including hettest. Is the command
1:09:27really only hettest? No. There is the
1:09:29mtest command. There's a command I
1:09:37forgot because it once appeared on a
1:09:40research methods exam—was it for MKN
1:09:43or MAP? Because it appeared once, and
1:09:45suddenly I was asked after the exam. "
1:09:47Ma'am, what is this command?" I forgot,
1:09:49I didn't teach it in class. So the
1:09:53commands aren't just this, except for
1:09:55multicollinearity. Multicollinearity
1:09:56has no other commands; it's only vif.
1:10:00But for the rest, there are several
1:10:02commands. Now, for normality, what is
1:10:04the H0? The H0 is not normally
1:10:07distributed. Eh, that's reversed. The
1:10:09H0 is that the error is normally
1:10:11distributed. The H1 is that the error
1:10:13is not normally distributed. What
1:10:16should be cross-checked, Prof? The one
1:10:18right here. This should have content.
1:10:21Because it has no content, it's
1:10:22definitely not turned into a question.
1:10:25Who is the research methods course
1:10:26coordinator? Is it still Mr. Yun? Or
1:10:41does anyone know who the research
1:10:42methods coordinator is? Nobody knows
1:10:45anymore. What about you guys? I have
1:10:50been teaching research methods with Mr.
1:10:52Yun since 2014. The coordinator is Mr.
1:10:54Yun. Then, well, it's just that because
1:10:59Mr. Yun is in ASP, then for MKN and MAP
1:11:02, it's automatically different. Unless
1:11:06it's being merged into one in the same
1:11:08semester, that usually goes back to Mr.
1:11:11Yun. Like that.
1:11:14>> Oh yes, Ma'am. In the RPS, it's Mr.
1:11:16Yuniarto.
1:11:17>> Okay. Still okay. That's it. So looking
1:11:21at it here, the benchmark is the same
1:11:24as the hettest earlier. Look at the
1:11:26prob here to see what the result is.
1:11:27Compare it with 5%. So, what is the
1:11:31solution if it's not normally
1:11:33distributed? The solution is to add
1:11:36data. But there are times when we can't
1:11:39add data; it's difficult, Ma'am, to
1:11:41find questionnaires, for example. Well,
1:11:44if you really can't add data, then you
1:11:46must cross-check how much data is
1:11:49already available. That's why the
1:11:53natural logarithm is also one of the
1:11:55solutions for normality. If the data's
1:11:59errors, after being regressed, are not
1:12:02normally distributed, another way is to
1:12:04log-transform it. But if it has been
1:12:08log-transformed, like this one, it’s
1:12:10already been ln-transformed, so why
1:12:12doesn't the normality test show up? Yes
1:12:15, because my feeling is that the data
1:12:17must be too small. It isn't visible
1:12:21here because there is no summary. But
1:12:24personally, whenever I teach data
1:12:26processing, the first thing, number 2
1:12:29or 3 after inputting data, I always
1:12:31require a summary so you know the
1:12:33behavior of the data. There, so the
1:12:37cross-check is here. Now, what if it
1:12:40has been log-transformed but there are
1:12:41still no satisfactory results—in
1:12:44other words, it didn't pass the
1:12:46normality test, what should I do, Ma'am
1:12:47? Like that. Just like the
1:12:49multicollinearity earlier, what is the
1:12:51solution, Ma'am? Well, the solution is
1:12:53for you to use theory. Use a theory
1:12:55called the Central Limit Theorem. Read
1:12:58the book again; it should be in there.
1:13:03It states that if your n is already
1:13:07more than 30, it will be considered
1:13:10normally distributed. But you still
1:13:14have to go through the process first,
1:13:16okay. You can't just jump to the
1:13:18conclusion that you don't need a
1:13:19normality test because it's already
1:13:20above 30. It doesn't work like that.
1:13:25That's all regarding the classical
1:13:27assumption tests. Is there anything
1:13:29else here? Uh, no. That's all I see.
1:13:35It's rare for Mr. Yun not to include an
1:13:38OLS test or not to include uh well, who
1:13:46knows what you guys will get. I don't
1:13:48know. Is there anything else, uh,
1:13:52quantitative? None. If there isn't,
1:14:05I'll move on to qualitative, okay.
1:14:06There are attachments down here. Okay.
1:14:10Uh, transcripts. Yeah. There aren't any
1:14:18, Mr. Yun. He didn't give anything
1:14:19strange. He's being kind again, Mr. Yun
1:14:28. Or maybe this was made in a rush.
1:14:34Okay, done. If there's nothing else,
1:14:36I'll move on to number two. Uh, number
1:14:40two, the second method related to, uh,
1:14:44qualitative. When dealing with
1:14:48qualitative, I always tell students,
1:14:51whether for their thesis or when I
1:14:54teach research methods, uh, don't
1:14:56forget to cross-check the recording.
1:15:03Once you have the recording, hurry up
1:15:05and get the wording done, the interview
1:15:07transcript. It's tiring. Yes, it is
1:15:11really tiring because there is so much,
1:15:13although actually there is there is
1:15:18assistance, uh, using NVivo or, uh,
1:15:21what is it called? Uh. If you guys
1:15:27upload it to YouTube and ask for it to
1:15:29be transcribed, later you can download
1:15:32the transcript. Now, that can be one
1:15:36method, it's just that it still has to
1:15:39be, uh, listened to again. Now, from
1:15:46there, once it's finished like what is
1:15:48below—even though that definitely
1:15:50isn't everything, because if it were,
1:15:52your 2.5 hours wouldn't be enough. Now,
1:15:56you are asked to recap the coding
1:15:59results into a format like this. This
1:16:01is the category, this is the code, this
1:16:03is where the interview quote is located
1:16:05. It is only written by line. That is
1:16:09why, whenever I make an interview
1:16:11transcript, I always teach to, uh, have
1:16:14one side on the left or the right,
1:16:17whichever you are more comfortable with
1:16:20. Put numbers 5, 10, 15, and so on. So
1:16:25you don't take too long looking for
1:16:28which line it is. Uh, Ma'am, I just
1:16:32want to highlight something. Yes, feel
1:16:33free to highlight it if you want. But
1:16:35when you want to input it into the
1:16:38coding, the coding should ideally be a
1:16:40new Excel file. You shouldn't be
1:16:44working in Word anymore. Because it
1:16:47should have been moved already. Now,
1:16:50when moving it, why bother copying such
1:16:53long interview quotes? So, that's why
1:16:56you only need to write which line it's
1:16:58on, or for example, informant one, row
1:17:01such-and-such. Or, informant two, row
1:17:05such-and-such, like that. So, you just
1:17:09have to look at this; whether you like
1:17:12it or not, you must cross-check it
1:17:14downwards, you have to read your
1:17:16previous methodology. Mm, doing
1:17:21qualitative research, be grateful
1:17:23because it means you are already
1:17:25experienced. You will certainly be
1:17:29faster compared to those who chose, for
1:17:33example, quantitative, as they aren't
1:17:36as experienced as you are, like that.
1:17:40So, there are pros and cons. From here,
1:17:45you look for what exactly? Look for
1:17:49things related to the theme. If the
1:17:55theme is cooperatives, feel free, for
1:18:02example, the launch of village
1:18:06cooperatives. So, if it's the launch of
1:18:12the 80,081 village cooperative
1:18:13institutions, what does this likely
1:18:15mean? Creating something new, right? So
1:18:20, where do you want to put that
1:18:25interview and its coding? Try to ensure
1:18:32that coding isn't done just once; it
1:18:36should be raw first. "Raw" in quotes
1:18:39means it's still long, and later it
1:18:41will be summarized to be smaller. Why?
1:18:44Because you have to group them: oh,
1:18:46this is related to new cooperatives,
1:18:48new cooperatives, new cooperatives. Oh,
1:18:51this is related to the attitudes of
1:18:54leaders, like that. Oh, this is related
1:18:57to technical matters in the field; only
1:19:00then can it be named. Here, you are
1:19:03indirectly forced to determine the
1:19:11category, the code, and the interview
1:19:16quote. Actually, you should start from
1:19:21the table on the far right. You have
1:19:25the important interview quotes first,
1:19:27what will you code them as? Oh, this is
1:19:31a new cooperative. Oh, this is the
1:19:34nature of the cooperative. Then, after
1:19:37that, you categorize it. So, the order
1:19:43is actually from the interview quote,
1:19:45then you code it. Then, from those
1:19:48codes, you categorize them. Let's take
1:19:51this as an example. Wait, wait. This is
1:19:56the launch of the Merah Putih
1:19:58Cooperative, including...Wait, where
1:20:04was the result? Like I said earlier,
1:20:09for example, this is the launch of the
1:20:11Village Cooperative institution. This
1:20:14is the interview, the line...oh,
1:20:17there's an A here, so A3. This is A3,
1:20:22the launch of 80,081 Merah Putih
1:20:27Village and Subdistrict Cooperatives.
1:20:30If you want to stick to "Village
1:20:31Cooperative," just go ahead. And what
1:20:35is the code for this? What is the
1:20:38coding? What do you want to code it as?
1:20:43For example, what do you mean by "new
1:20:47cooperative"? Or just write "
1:20:52cooperative creation," for instance.
1:20:55Then what else? "Cooperatives are a
1:20:57tool for economic struggle for the weak
1:20:59." So what is this? Economic struggle.
1:21:06So, it's about improving the economy,
1:21:07right? What is the coding? improving
1:21:10the economy, if there are many related
1:21:13to, uh, cooperatives, they improve the
1:21:16economy, and then cooperatives, uh uh,
1:21:22the concept is simple, one stick has no
1:21:24power, but joined together it becomes a
1:21:25strength. What does that mean?
1:21:28Cooperatives unite because this is in a
1:21:33region. Now, when that is made into,
1:21:35what is the category? the category of
1:21:38cooperative goals. What was the goal of
1:21:40the cooperative earlier? Improving the
1:21:42economy, uniting the surrounding
1:21:44community, like that. And cooperatives
1:21:48are not for those who are strong and
1:21:49established. The strong ones have
1:21:52surely already created a holding, uh, a
1:21:54company, a PT. These cooperatives are
1:21:57for those who don't have access yet,
1:21:59right? So the goal, uh, the goal, what
1:22:01is the coding? uh, opening access not
1:22:07yet having economic power. So it's
1:22:09similar to the one before, right,
1:22:11earlier, uh, improving the economy,
1:22:13this one is strengthening the economy.
1:22:15Now, that means, uh, there are
1:22:17different codings. Improving the
1:22:21economy, strengthening the economy,
1:22:23adding access. Later, when it is turned
1:22:26into a category. What is the category?
1:22:28The category is cooperative goals. like
1:22:33that. So if you make, what do you call
1:22:35it, if it branches out, why did I
1:22:38suddenly forget things like this,
1:22:42brainstorming. Now, in brainstorming
1:22:46you have one idea at the start and then
1:22:48you break it down to the right. Oh, how
1:22:50about the location? Oh, if this, well,
1:22:53if this is reversed, you have it on the
1:22:55outside. The interview results, you
1:22:57have them on the outside. Later you
1:23:00enter one, enter another one. That's
1:23:05messy, right. Earlier A3, there was
1:23:11line A5 6 7, A7, then there was line A8
1:23:14, there was line A11, like that. Now,
1:23:23those are entered one by one. Hm. Then
1:23:29what's this? There are forces that do
1:23:32not want the country to grow strong and
1:23:34, uh, the country of Indonesia to grow
1:23:37strong independently, like that. There
1:23:41are big forces that do not want a
1:23:43country like Indonesia to grow strong
1:23:45independently. So what is the category?
1:23:53a country's lack of independence, like
1:23:55that. What else is the category
1:24:01competitors in the trading world, like
1:24:05that. Later, when entered into a
1:24:09category, what is the category?
1:24:11Obstacles for cooperatives like that.
1:24:16Now, indeed, in qualitative work,
1:24:18friends, you must read a lot because if
1:24:21you don't read a lot, you won't find
1:24:23what the category is, like that.
1:24:29Because from coding to category, a new
1:24:32word should emerge see. Now, here I can
1:24:41only give examples, right, earlier,
1:24:42cooperative functions. But in reality,
1:24:47it's not like that. It still must be
1:24:50dissected one by one. Uh, missed it.
1:24:55Here. This is an example of the
1:24:56distribution center function, the
1:24:58interview quote. Yes, you have to read
1:25:00A50, what was the interview quote
1:25:02earlier. Why did the code become
1:25:04distribution center? Just this, uh,
1:25:08when I teach qualitative, uh, coding
1:25:10usually has at least three like this.
1:25:14interview quote, coding one, this
1:25:16coding two, but if the language is
1:25:18still very complicated, it could be
1:25:20four or five because sometimes from the
1:25:22interview quote we cannot yet get what
1:25:25the code is. Can't find it, Ma'am, they
1:25:29say, to represent the code, right. If
1:25:32the words used to represent the codes
1:25:34are essentially just you not finding
1:25:36how to search for the categories yet.
1:25:40So, you really do need to read a lot so
1:25:41you can get it. Well, that is the way.
1:25:47So, like it or not, you all have to
1:25:50make them one by one. Don't be too
1:25:54detailed because I don't know how many
1:25:56interview attachments there are. I only
1:25:57opened A earlier. How many attachments
1:26:00are there?
1:26:02>> There are two, Ma'am.
1:26:03>> So A and B?
1:26:05>> Yes. Okay. Well, here are A and B. Two
1:26:11pages. Okay. Oh, two pages is still
1:26:14safe. It's just that the speed depends
1:26:20on each of you. Later, once you get
1:26:25many codes, you'll think, "Oh, this is
1:26:28related to the function, oh, this is
1:26:31related to the goal, oh, these are the
1:26:34cooperative's obstacles, oh, this could
1:26:37be an innovation, oh, what else is
1:26:40there?" Oh, this...for the future, what
1:26:47is it? Target. From there, there will
1:26:55be many items that become categories
1:26:58which will later appear in the next
1:27:01questions. That's why in the next
1:27:04question, if I'm not mistaken, I read
1:27:06about latent constructs. Yes, exactly.
1:27:14What latent constructs can be
1:27:16identified from the text? Mention and
1:27:19explain briefly. So, what is a latent
1:27:25construct? A latent construct is, in
1:27:27quantitative terms, like asking, "What
1:27:29is the variable?""What are the
1:27:31indicators?" Yes, something like that.
1:27:36What do you think you have? Oh, I see.
1:27:39There is the cooperative's function,
1:27:40the cooperative's goal, and the
1:27:41cooperative's obstacles. Hmm, from
1:27:44there, what else? Why? Because if you
1:27:52already have several categories that
1:27:55can be made into one variable, or one
1:27:58construct, then from here you should be
1:28:02able to make arrows. Arrows that, in
1:28:06quotation marks, if in quantitative,
1:28:07the arrows just go from X to Y, right?
1:28:10At most, X2 to Y, they all point to Y.
1:28:13Now, this is the same, where do they
1:28:15point? It's just that it becomes more
1:28:22...that's why it's written as a
1:28:25flowchart, more like a story because
1:28:27it's not just one-way like quantitative
1:28:30, because it will say, "Oh, from this
1:28:33cooperative, there's its function, its
1:28:35goal, and its obstacles.""So, what do
1:28:39you want to discuss?""Oh, what's being
1:28:40discussed are the obstacles.""How do
1:28:42you control them?" So it's only here.
1:28:44Then the story continues, the flowchart
1:28:48. Anything not discussed in detail, no
1:28:51need to include. Well, so that's...
1:29:05what's it called?...in short......yeah,
1:29:12that's it. You really have to work on
1:29:17it one by one. Then, while I
1:29:22cross-check number 10, question number
1:29:309 is about the findings framework and
1:29:34empirical interpretation that you have
1:29:37outlined from questions 1 to 6 in
1:29:40method 1. So from here, later you all
1:29:45will have...that's my assumption. I
1:29:49haven't read the interview results to
1:29:51the end yet. But my assumption is that
1:29:55later below there will definitely be
1:29:57someone talking about, for example,
1:30:00things related to schools related to,
1:30:05well, as mentioned above, there are
1:30:09total operations, schools, and
1:30:12malnutrition. So there will be a
1:30:17connection here. Now, from that
1:30:20flowchart, connect it to what was above
1:30:27and answering this is like you all are
1:30:32doing research using mixed methods. So
1:30:37there is the quant part and the quali
1:30:39part. Now, this quali part is
1:30:43supporting. If earlier it said
1:30:46malnutrition is influenced by—wait,
1:30:48reversed—poverty influences
1:30:50malnutrition, well, that's my guess, I
1:30:53haven't read it in detail yet. My guess
1:30:56is that below there will surely be one
1:30:58sub-theme categorizing the causes for
1:31:02the emergence of the cooperative idea.
1:31:06One of them could be poverty. It could
1:31:12possibly be, um, what is it, not
1:31:14graduating from school or even not
1:31:17going to school. That means it's the
1:31:20education level, right? Even though
1:31:23when combined with the quant results,
1:31:25there's no education level in the quant
1:31:28, Ma'am. Having the total number of
1:31:30schools is fine. Why? Because the total
1:31:34number of schools indirectly depicts
1:31:37how the surrounding community obtains
1:31:39education. If there are, say, 1,000
1:31:45people in the community and only one
1:31:48school, are you sure those 1,000 can
1:31:51all be educated, even if one class only
1:31:55has 40 people? Yes, where are the
1:31:59others supposed to go? Like that. So
1:32:03this will be explained here the
1:32:07connection between what was mentioned
1:32:09earlier, surely below it will appear:
1:32:10there's poverty, there's malnutrition,
1:32:12there's government intervention. Why?
1:32:14Because there were obstacles earlier,
1:32:16right? At least that’s what I think,
1:32:19and, um, I don’t know the answer key.
1:32:23And as far as I’ve worked, as far as
1:32:26I’ve worked with Mr. Yun, Mr. Yun
1:32:28never gives an answer key that is
1:32:30strictly rigid. What is given are the
1:32:35boundaries. Um, why? Because especially
1:32:40with qualitative, in qualitative you
1:32:43can give different category names, you
1:32:46can give different coding clauses, and
1:32:49that can be justified as long as it's
1:32:53within those boundaries. Like that. Now
1:33:01, including number 10, um, you are
1:33:04asked for the implications. Why?
1:33:09Because for you, simply put, in a
1:33:11thesis, this is the conclusion and
1:33:13suggestions. Like that. That’s it. So
1:33:22if you want more details, come on, feel
1:33:25free, whoever wants to try, but not me.
1:33:31I shouldn't be the one doing everything
1:33:33, right? Why? Because you all need to
1:33:36know it, right? These are the best
1:33:42mango farmers, but there are no trucks,
1:33:44no one to buy, so they end up rotting.
1:33:49So what does that mean, Ma'am? Does
1:33:51that mean accommodation or
1:33:53transportation? Or distribution. Now,
1:33:57you all will have different codes for
1:34:00that. Subsidized fertilizer is scarce,
1:34:05regulations are convoluted. So what
1:34:08does this fall under? Oh, I made the
1:34:10category government policy, Ma'am. Like
1:34:12that. Go ahead. Oh, I made the category
1:34:15rules, that's allowed. Because nothing
1:34:19is wrong. like that. the price drops
1:34:24like that. Now, what is the context
1:34:29here? From the context of the results.
1:34:33If before we were talking about the
1:34:35function, earlier we mentioned the
1:34:37function of cooperatives, the goals,
1:34:39and the obstacles. Now, this is from
1:34:42the agricultural context. that. Because
1:34:49, if the goal of the cooperative was,
1:34:51if I recall correctly, to increase
1:34:53access. Now, there is no access. No
1:34:57trucks, no one to buy. That means there
1:35:01is no access if there are no trucks.
1:35:06this. The people are losing 10 trillion
1:35:09per year. manipulating rice quality,
1:35:12playing with prices like that. Greed,
1:35:16so what is this that? That means, what
1:35:26do you call it, the problem, right?
1:35:34what is this, but immediately at 83,000
1:35:39points. Cooperatives are formed in
1:35:4383,000 locations. This is on, uh, line
1:35:4743. So it goes along with which one?
1:35:51Along with the number of cooperatives
1:35:54earlier. It's, uh, switching back and
1:35:59forth. If you want to read it easier
1:36:05after taking notes like this. If it's
1:36:08on a computer it's easy, you can write
1:36:11based on the transcripts below and then
1:36:14just cut and paste above. But if you're
1:36:18writing by hand on a PDF, how many
1:36:20lines do I actually need? Because if I
1:36:25provide five lines for five codes, five
1:36:28interview transcripts, but then when
1:36:30reading further down, oh, there's more.
1:36:34Like that. It's hard to insert. like
1:36:42that. Okay, any questions so far? Look,
1:37:08this is about pregnant women being able
1:37:10to get enough protein intake. So what
1:37:12does that mean? This is, uh, me wanting
1:37:16. If I want, it means a hope. So what
1:37:20is the condition? Here, you can find
1:37:23the condition of malnutrition. So, yes,
1:37:33as I mentioned earlier, there is
1:37:35implied and explicit. So you have to
1:37:47cross-check them one by one, slowly.
1:37:53Please try it while we're here, if you
1:37:56have any questions, feel free. So you
1:38:02must cross-check one by one. Now, here
1:38:09there is also 80,000 plus village
1:38:10cooperatives. So what does that mean,
1:38:13the number of cooperatives? Well, it's
1:38:29finished. Down here, I don't see it. It
1:38:33doesn't seem to be related to
1:38:35malnutrition. If I read it briefly, I'm
1:38:40just skimming, I only got the
1:38:42malnutrition context above. That's it,
1:38:55everyone. Any questions? If there are
1:39:11none, I will stop sharing for now. Okay
1:39:24, sure. Uh, any questions? Anything
1:39:37else? If not, do you want to continue,
1:40:11or what else is there? Or is it enough,
1:40:17or what? It seems enough, Ma'am.
1:40:25Because you've explained the
1:40:27quantitative and qualitative aspects
1:40:29very well, and based on the guide, it's
1:40:31just those two, the coding and how to
1:40:33read the stats.
1:40:36>> Okay, ready. Alright then, if that's
1:40:41all, hopefully it helps you all, sorry
1:40:44I can't discuss both. Otherwise, we'd
1:40:47end up with 3 credits. So, I'll stop
1:40:52here for now, please read through it
1:40:55again. If there’s still anything
1:40:59you're not sure about regarding the
1:41:01quant or qual, feel free to ask. Okay,
1:41:05I’ll wrap it up for now. That's all,
1:41:07good afternoon everyone. Peace be upon
1:41:11you.
1:41:12>> And peace be upon you too.
1:41:13>> And peace be upon you too. Thank you so
1:41:15much for your time, Ma'am.
1:41:18>> You're welcome. Okay, I'm going to head
1:41:20out now. Thanks, Fat Nazani and
1:41:22everyone. Yeah.

About this transcript

This page contains the full transcript of Persiapan UAS Metodologi Penelitian | Buddy Tentir | Buddy Program ASP oleh Bu Nur Indah Lestari by Buddy Program ASP, generated from the public captions YouTube serves with the video. The transcript has 10,672 words across 1,639 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.