Delforterie posted on Tuesday, April 19, 2011 - 5:26 am
I have 11 items loading on 1 factor, and I want to detect measurement bias by including an interaction with a dichotomous variable (say gender). Is this the correct way to do this:
TITLE: SEM DATA: FILE IS 13042011.dat; VARIABLE: NAMES ARE y1-y11 gender; ANALYSIS: TYPE = RANDOM; ALGORITHM = INTEGRATION; MODEL: f1 BY y1*-y11; f1 @1; gender @.5; f1 with gender; [f1 @ 0]; [gender @0]; f1xgender | f1 XWITH gender; y1 ON f1xgender; OUTPUT: TECH1 TECH8;
I have copied this from example 5.13 in the manual, but I am not sure this is applicable to my situation (i.e. a dichotomous interaction).