Pairwise comparison by time point after repeated measure

[Giuseppe Giannotti]

Hi,
I'm eying trouble trying to figure out a way to do pairwise comparisons by time point after repeated measure Anova. I have 2 time points (time within) and 4 different groups each time point (between). I don't want to have all the pairwise comparison using pwmean, but instead multiple comparisons for each time using just one command.
Thank you,
Giuseppe

[Marcos Almeida]

Welcome to the Stata Forum/Statalist.

The best way to entail insightful replies, as recommended in the FAQ, is sharing data (full, abridged or mock).

You may do it by using Code delimiters or installing the SSC dataex.

[Giuseppe Giannotti]

Hi Marcos,
Thank you for your time. These are the data. I was using wsanova command to get a repeated measure anova result (3-way).

wsanova DepVar Time if Subjects, id(Subjects) bet (VAR1 VAR2 VAR1*VAR2)

I like to use wsanova because the code it's really simple. However, I get the same results with the following code:

anova DepVar VAR1##VAR2 / Subjects|VAR1#VAR2 Time##VAR1##VAR2, repeated(Time) bseunit(Subjects)

 
Subjects
VAR1
VAR2
Time
DepVar

1
1
1
1
39.7

5
1
1
1
50.3

8
1
1
1
25.3

12
1
1
1
41

18
1
1
1
40.3

19
1
1
1
57.3

22
1
1
1
39

31
1
1
1
55.3

37
1
1
1
72.3

41
1
1
1
40.3

46
1
1
1
43.7

4
1
2
1
49.3

9
1
2
1
46.7

13
1
2
1
43.3

16
1
2
1
18.3

23
1
2
1
38

26
1
2
1
56.3

27
1
2
1
29.7

29
1
2
1
45

33
1
2
1
59.7

48
1
2
1
43.7

3
2
1
1
40

10
2
1
1
16

14
2
1
1
64.7

17
2
1
1
42

25
2
1
1
49

28
2
1
1
39.3

2
2
2
1
32.7

11
2
2
1
73

15
2
2
1
49

20
2
2
1
44.3

30
2
2
1
35

32
2
2
1
49

34
2
2
1
57.3

39
2
2
1
48.7

43
2
2
1
53.3

1
1
1
2
51

5
1
1
2
36

8
1
1
2
51

12
1
1
2
164

18
1
1
2
125

19
1
1
2
119

22
1
1
2
98

31
1
1
2
177

37
1
1
2
120

41
1
1
2
157

46
1
1
2
130

4
1
2
2
41

9
1
2
2
73

13
1
2
2
60

16
1
2
2
12

23
1
2
2
54

26
1
2
2
125

27
1
2
2
58

29
1
2
2
120

33
1
2
2
98

48
1
2
2
33

3
2
1
2
108

10
2
1
2
20

14
2
1
2
57

17
2
1
2
59

25
2
1
2
79

28
2
1
2
91

2
2
2
2
155

11
2
2
2
70

15
2
2
2
96

20
2
2
2
127

30
2
2
2
84

32
2
2
2
178

34
2
2
2
139

39
2
2
2
250

43
2
2
2
109

however, if i ran

pwmean DepVar, over(Time VAR1 VAR2) mcompare(snk) pveffects

I get all the pairwise comparison within time point (1 and 2) and between time point (all the variable of time point 1 Vs. all the variables time point 2).
I was looking for a way to get multiple comparisons af all my dataset only within each time point.
Thank you
Giuseppe

[Giuseppe Giannotti]

I'm still struggling with it!

[Joseph Coveney]

Here, do this:

mixed DepVar i.(VAR1##VAR2##Time) || Subjects: , reml dfmethod(kroger) nolrtest nolog
contrast r.VAR1#VAR2@Time, mcompare(bonferroni) small noeffects