Hi I have a data set of 124 individuals who are nested within (i.e. make up) 62 couples. All variables are observed and continuous.
I'm interested in predicting 2 outcomes Y1 and Y2 from individual level predictors I1, I2 and I3, and couple-level predictors C1 and C2.
I've succesfully specified my model as
NAMES ARE COUPLE I1 I2 I3 C1 C2 Y1 Y2; USEVARIABLES ARE COUPLE I1 I2 I3 C1 C2 Y1 Y2; MISSING=BLANK; WITHIN = I1 I2 I3; BETWEEN = C1 C2; CLUSTER = COUPLE;
ANALYSIS: TYPE = TWOLEVEL RANDOM
MODEL: %WITHIN% Y1 ON I1 I2 I3; Y2 ON I1 I2 I3; %BETWEEN% Y1 ON C1 C2; Y2 ON C2;
However, responses to one of my DV's (Y1) and to one of my IVs (I2) are likely to be non-independent within couples. How do I model that dependency? I need to allow for a correlation between the observations of these variables within couples?