Hello, I'm trying to run a multiple mediator model. paam and efam are L1 mediators between jcSR and eng (also L1)(fixed slopes). day is a control variable at L1. I want to use the ESTIMATOR=BAYES option but get the following error:
*** ERROR in MODEL command Unrestricted x-variables for analysis with TYPE=TWOLEVEL and ESTIMATOR=BAYES must be specified as either a WITHIN or BETWEEN variable. The following variable cannot exist on both levels: JCSR
Syntax is below: USEVARIABLES ARE id day paam efam jcSR eng; within=day; CLUSTER IS id; ANALYSIS: TYPE IS TWOLEVEL RANDOM; ESTIMATOR=BAYES; MODEL: %WITHIN% paam ON jcSR(aw1); efam ON jcSR(aw2); eng on paam(bw1); eng on efam(bw2); eng ON jcSR; paam ON day; efam ON day; eng ON day; paam WITH efam; %BETWEEN% paam efam eng; eng ON paam(bb1); eng ON efam(bb2); eng ON jcSR; paam WITH efam; paam ON jcSR(ab1); efam ON jcSR(ab2); MODEL CONSTRAINT: NEW(abw1 abw2 abb1 abb2 conw conb); abw1=aw1*bw1; abw2=aw2*bw2; abb1=ab1*bb1; abb2=ab2*bb2;
Unlike ML, with Bayes estimation the latent variable decomposition of the independent variables (JCSR in this case) is not done. You have to create a between-level version of this variable (you can use the cluster-mean option) and specify JCSR as a within variable.