Question about latent class analysis(PROC LCA)

[lionking19063]

Hello, 

I am new user of PROC LCA (latent class analysis). I wonder if PROC LCA is applicable to categorical independent variables only. Can it be used in continuous variables as well? If yes, what statement in PROC LCA can be used for continuous variables? Thanks

[Rick_SAS · Accepted solution]

You can find the documentation of PROC LCA at

https://methodology.psu.edu/downloads/proclcalta

The documentation describes the mathematical model on p. 5.  It says " a covariate denoted X which may be either continuous or dichotomous." On p. 18 it says that the COVARIATE statment supports continuous variables.  See the doc if you have questions about other statements.

[lionking19063]

Supposed there is a dataset called test with two categorical (X1, X2) variables and 1 continuous variable (x3). There are only two levels in both categorical variables. I would like to run LCA with only two clusters on the data and wonder if the syntax makes sense or not. Thank you in advanced.

PROC LCA DATA=test;

NCLASS 2;

ITEMS x1 x2;

CATEGORIES 2 2;

COVARIATES x3;

SEED 100000;

RHO PRIOR=1;

RUN;