Factorial Design (Part A): Scenario & Descriptive Statistics — Transcript

Full transcript

1. 0:00let's take a look at the 2x two
2. 0:02factorial design that where we have two
3. 0:05factors that could influence a dependent
4. 0:07variable and we'll look at a scenario as
5. 0:10well as descriptive
6. 0:12statistics to get started I want to let
7. 0:14you know that I was asking my students
8. 0:15in class what variables could perhaps
9. 0:18influence a student's weight and several
10. 0:20possibilities that you can read here
11. 0:22were offered and for each of these
12. 0:23possibilities we identified the scale of
13. 0:25measurement whether it was nominal which
14. 0:27means categories uh that you do not rank
15. 0:30they're all equal so to speak uh ordinal
16. 0:33which would be categories that you would
17. 0:35rank and then there were some things
18. 0:37that were um scale scale meaning ordinal
19. 0:40or or ratio uh and by scale I mean that
20. 0:43these are uh variables where the numbers
21. 0:46are are not categorical they're actual
22. 0:48numbers the way you would think you
23. 0:49could add subtract multiply and divide
24. 0:51with them okay so students could uh take
25. 0:54a look at this list of variables that
26. 0:56might affect a person's weight and as a
27. 0:59class we chose two variables to focus on
28. 1:02uh the variables were gender which is a
29. 1:04nominal variable male and female and
30. 1:06also level of stress and possible levels
31. 1:09of stress might be low medium and high
32. 1:11and again this is categorical but it's
33. 1:13categories that you can rank so that
34. 1:14makes it ordinal our design in this
35. 1:17particular uh illustration would have
36. 1:19been a 2x3 factorial design our first
37. 1:22variable gender had two levels our
38. 1:25second Factor level stress had three
39. 1:28levels so at three says there's three
40. 1:30levels for the second
41. 1:31variable an initial survey of my
42. 1:34students however indicate there were no
43. 1:35low stress students perhaps it was
44. 1:37because finals was approaching so we
45. 1:39went to a 2X two factorial design two
46. 1:42levels for gender two levels for level
47. 1:45of stress our dependent variable again
48. 1:48being the students weight to collect
49. 1:50data we had a simple Anonymous survey
50. 1:52students reported their gender their
51. 1:54stress level and their
52. 1:56weight in terms of recording the data in
53. 1:59SPS s uh SPSS use a number to indicate a
54. 2:03category so one we use to indicate male
55. 2:05two for female nominal data for stress
56. 2:09or ordinal data a two minant mid-level
57. 2:11stress and a three would be even more
58. 2:13stress high level
59. 2:15stress our null hypotheses whenever you
60. 2:17do a 2X two uh factorial design using uh
61. 2:21and analyzing your data with in no Nova
62. 2:23you don't really get to pick what your
63. 2:25research and null hypotheses will be
64. 2:27they're kind of a given so for our uh
65. 2:30factor of uh gender the N hypothesis is
66. 2:34no difference for our factor of stress
67. 2:37null hypothesis no difference and in
68. 2:39terms of a possible interaction between
69. 2:40gender and stress our null hypothesis no
70. 2:43interaction so n hypothesis is all about
71. 2:46no difference no
72. 2:48interaction the research hypothesis is
73. 2:50that there will be a difference in
74. 2:52weight between males and
75. 2:53females that there will be a difference
76. 2:55in weight between mid and highle
77. 2:57stressed people and that there will be
78. 2:59an interaction between stress and gender
79. 3:01on weight notice that when we're talking
80. 3:03about our factors of gender and stress
81. 3:06we're saying there'll be a difference
82. 3:08it's a two-tail hypothesis we don't get
83. 3:10to pick uh whether we want one tail or
84. 3:12two tail with the Inova the data
85. 3:14analysis is always twail that is you
86. 3:16simply get to say that there will be a
87. 3:18difference you you do not get to predict
88. 3:19in which direction for the interaction
89. 3:22research hypothesis is simply there will
90. 3:23be an
91. 3:25interaction so when we look at our three
92. 3:27hypotheses first hypothesis say that
93. 3:30there will be a main effect of gender in
94. 3:32terms of people's weight our second
95. 3:33hypothesis says that there will be a
96. 3:35difference uh that is a main effect in
97. 3:37terms of stress level on people's weight
98. 3:40and a third hypothesis is that there
99. 3:42will be an interaction between our two
100. 3:44factors gender and
101. 3:47stress okay here are the possible
102. 3:49descriptive statistics for our results
103. 3:51uh this is what SPSS gives us if we
104. 3:54under options choose descriptive
105. 3:55statistics and this has a lot of
106. 3:57information it's kind of nice to put in
107. 3:58easier format for your reader to to see
108. 4:01um so we grab midlevel stress for males
109. 4:05and we plop it in we grab the highle
110. 4:07stress for males and put it in and for
111. 4:09females mid level of stress we put it in
112. 4:12and high level of stress we put it in a
113. 4:14graph is a great way to go to share with
114. 4:17people uh your results so notice that uh
115. 4:20males uh are weighing on average more
116. 4:22than the females so that's clearly a
117. 4:24main effect of gender also notice that
118. 4:27people who are highly stressed way more
119. 4:29more than people who are mid-level
120. 4:31stress so that's another main effect
121. 4:33also notice that if you are um male take
122. 4:37a look instead of just having the same
123. 4:40type of weight gain that a female
124. 4:42experienced uh so that these two lines
125. 4:44would have been parallel the males
126. 4:45actually gained considerable more weight
127. 4:48uh when they were highly stressed
128. 4:50compared to females when they were
129. 4:51highly stressed so that's an indicative
130. 4:53of an interaction