Message/Author |
|
|
I ran LCA analyses and found 7 classes in our data. We are wondering if it is possible to run some kind of factor analysis like procedure to see if the classes form a hierarchical structure (basically can we further group them). Is this possible? If so, can you point me towards the correct analysis? |
|
|
To add structure on the classes you can add a second class variable - for example you can explore two class variables c1(2) c2(4) - see user's guide examples 7.18, 8.12, 8.15 for some ideas. |
|
|
Does this syntax look correct? I am also getting an error re: memory space. Is there a workaround? variable: names are ID y1-y21; usevariables are y7-y12 y14 y16-y20; CATEGORICAL = y7-y12 y14 y16-y20; classes = c1(2) c2(2) c3(2) c4(2) c5(2) c6(2) c7(2); missing is all (99); auxiliary=id ; Analysis: type = mixture; algorithm=integration; STARTS=200 50; model: %overall% f1 BY; f1@1; c2 c6 c7 on f1*1(1); %overall% f2 BY; f2@1; c3 c5 on f2*1 (1); %overall% f3 by; f3@1; c1 c4 on f3*1 (1); |
|
|
I don't see how this model should be interpreted - for instance, it is not clear to me how the 7 c variables are distinguished. |
|
|
Here was the original LCA syntax that ran fine. But, I am trying to figure out how to test if there are any second order relationships between the classes. The "eye ball" test seems to show some potential to group them based on the indicators, but I want to actually test that. The syntax I posted above was just a shot in the dark based on example 7.18, but obviously is not right. variable: names are ID y1-y21; usevariables are y7-y12 y14 y16-y20; CATEGORICAL = y7-y12 y14 y16-y20; classes = c(7); missing is all (99); auxiliary=id ; Analysis: type = mixture; STARTS=200 50; |
|
Back to top |