Latent Class analyses and repeated measures

[Tommaso Salvitti]

input int ID byte(TIME_POINT A B C)
1 0 5 10 10
1 1 25 25 25
1 2 50 50 25
2 0 50 25 25
2 1 25 75 75
2 2 . . .
3 0 25 50 50
3 1 50 95 50
3 2 50 50 75
4 0 5 25 5
4 1 10 10 25
4 2 5 10 90
5 0 . . .
5 1 25 5 10
5 2 50 90 10
6 0 . . .
6 1 5 5 5
6 2 5 5 5
7 0 . . .
7 1 10 5 5
7 2 10 5 5
8 0 . . .
8 1 . . .
8 2 75 95 25
9 0 . . .
9 1 5 5 5
9 2 5 5 5
10 0 10 10 5
10 1 10 5 10
10 2 10 10 25
11 0 25 50 10
11 1 5 95 10
11 2 50 95 5
12 0 50 95 75
12 1 50 95 50
12 2 50 95 50
13 0 50 95 25
13 1 50 95 25
13 2 50 95 50
14 0 50 95 5
14 1 50 95 25
14 2 50 95 25
15 0 50 95 .
15 1 50 95 50
15 2 95 75 25
16 0 . . .
16 1 75 25 75
16 2 75 50 75
17 0 90 50 95
17 1 50 25 95
17 2 . . .
18 0 . . .
18 1 50 5 75
18 2 25 10 75
19 0 25 50 50
19 1 25 50 50
19 2 50 50 75
20 0 . . .
20 1 . . .
20 2 50 25 25
21 0 75 95 95
21 1 95 75 95
21 2 75 50 95
22 0 95 75 50
22 1 75 50 75
22 2 90 25 75
23 0 25 50 5
23 1 . . .
23 2 5 25 5
24 0 . . .
24 1 . . .
24 2 25 5 10
25 0 25 50 5
25 1 25 10 25
25 2 . . .
26 0 75 10 10
26 1 95 25 95
26 2 95 50 95
27 0 50 50 10
27 1 10 50 10
27 2 10 50 10
28 0 75 10 25
28 1 25 25 5
28 2 5 5 5
29 0 . . .
29 1 90 25 75
29 2 90 25 5
30 0 . . .
30 1 25 25 25
30 2 25 10 50
31 0 . . .
31 1 75 95 95
31 2 . . .
32 0 . . .
32 1 . . .
32 2 5 5 10
33 0 50 75 10
33 1 . . .
33 2 . . .

Good morning everyone,
I am attempting to develop a latent growth curve model repeated measures (LCARM)
Above mydataset.
I have 3 TIME_POINT (0, 1 and 2) and A,B,C are the three variables measured for each child at 3 time points. Can anyone tell me what instruction to write to build the model?
Thank a lot Tommaso Salvitti

[Tommaso Salvitti]

Someone could help me?

[Weiwen Ng]

Your title may be misleading. It appears that this has nothing to do with latent class analysis.

I am not familiar with the term latent growth curve model. There is an example for the SEM command that demonstrates how to fit a latent growth model. This would require you to reshape the data into wide format. You could also fit a mixed model, which I think is likely equivalent; you don't have to reshape the data in this case. You should probably clarify what you are modeling, e.g. are you looking to estimate the change in A, controlling for B and C, or something else?

Also, please don't post the same question multiple times. We also prefer posters not to bump a thread too often.