

Withinperson repeated measures corre... 

Message/Author 

joz posted on Friday, July 24, 2020  8:02 am



I am using DSEM without random slopes (T is small; N large). I am doing withinperson correlations between the following (assessed daily) P and X; P and Y; N and X; N and Y X and Y are highly correlated. 1) Is the following syntax correct? 2) Should I use TYPE = TWOLEVEL or TWOLEVEL RANDOM? 3) how do I control for Y in withinperson correlations of P and X in the %within model (e.g. P with X)? CLUSTER = id; USEVAR = P N X Y; Missing = ALL(999); ANALYSIS: TYPE = TWOLEVEL; MODEL: %WITHIN% P with X; P with Y; N with X; N with Y; P with N; %BETWEEN% P N X Y; P with X; P with Y; N with X; N with Y; P with N; output: standardized cinterval; 


1) If you want to run a DSEM model you have to add something like that lagvar=P(1); MODEL: %WITHIN% P on P&1; Also you are missing X with Y; in your model 2) It doesn't matter 3) You can use interaction modeling https://www.statmodel.com/download/LVinteractions.pdf http://statmodel.com/download/WebNote23.zip but honestly I am not sure you need to go there. Try first P on X Y; 

Back to top 

