|
Message/Author |
|
|
Hi Mplus veterans! I ran a Two-Level Random Loadings IRT model. TITLE: RANp1 DATA: FILE = ... NAMES = id group y1-y5; USEVARIABLES = group y1-y5; CATEGORICAL = y1-y5; CLUSTER = group; ANALYSIS: ESTIMATOR = BAYES; PROCESSORS = 2; BSEED = 48932; BCONVERGENCE = .05; TYPE = TWOLEVEL RANDOM; Biterations = 50000000(1000); MODEL: %WITHIN% s1-s5 | f BY y1-y5; f@1; ; %between% f; y1-y5(v1-v5); s1-s5(v6-v10); MODEL PRIORS: v1-v5 ~ N(1, 0.09); v6-v10 ~ N(0, 1); OUTPUT: TECH1 TECH8; SAVEDATA: FILE=...; SAVE=FS(50); FACTORS=f; In the savedata output, I got the following variables: y1 -y5 F%W (mean, median, std, 2.5%, 97.5%) F%B (mean, median, std, 2.5%, 97.5%) S1-S5 (mean, median, std, 2.5%, 97.5%) B_Y1-B_Y5 (mean, median, std, 2.5%, 97.5%). Is S1 the between level lambda value? Is B_Y1 the between level tau value? Or are they deviations from the mean? How can I obtain group specific lambda and tau values? Thank you! SW |
|
|
S1-S5 are the factor scores for the random slopes (loadings) and B_Y1 etc are the between-level random intercepts of the y variables. So they are the group-specific values of lambda and tau (actually minus tau) that you ask about. If you add them to your FACTORS= statement you will obtain them all. |
|
Back to top |
|
|