We are looking at whether the teacher-specific intervention and 3 take home interventions (t1, t2, t3) will influence math, reading, and writing. Between-subjects = ses. IV: teacher & t1, t2, t3 (coded as 3 separate dummy variables) DV: math, reading, writing Mod: ses id = individual child
*note:repeated samples design,every child receives all levels of IVs.
My goal is to get slopes for the moderation for all levels of both independent variables.
Is there any way to test for the simple interaction of teacher x ses & t1,t2,t3 x ses? Please advise. Thank you.
NAMES ARE id, sestot, teacher, math, read, writ, t1, t2, t3; cluster = id; WITHIN = teacher, t1, t2, t3; BETWEEN = sestot;
ANALYSIS: TYPE = TWOLEVEL RANDOM;
writ ON t1 (b1); writ ON t2, t3; writ ON teacher; math ON t1, t2, t3; math ON teacher; read ON t1, t2, t3; read ON teacher;
%BETWEEN% writ on sestot; read on sestot; math on sestot;
MODEL CONSTRAINT: NEW(LOW_moder MED_moder HIGH_moder SIM_slope_low SIM_slope_med SIM_slope_high);
A related question is, take-home intervention is really 3 dummy variables. That is, it is coded as 3 variables (t1, t2, t3). How can I combine this categorical predictor so I can look at "take home intervention" as a whole (one variable) and not 3 separate groups (i.e. t1, t2, t3). I want to test the moderation on take-home intervention collectively?
Sorry, I wasn't clear on my question. What I was referring to was that it is a repeated samples design for t1, t2, t3. They are actually different time points, or rather, categorical (or 0/1) representations of different interventions. So, the dummy variables work such that t1 is only 1 for rows where that subject received t1. 0 for rows where t1 wasn't given. t2 is only 1 for rows where that subject received t2.
There are no rows where t1, t2, and t3 are all 1. Only one is active at a time.
Is there functionality in MPLUS which allows one to collapse 3 time points like this into a single categorical variable? The 3 dummy variable approach requires me to analyze every group separately. Thanks for all the help.