SPSS Statistics

Random Effect Coefficients for Dummy Variables

  • 1.  Random Effect Coefficients for Dummy Variables

    Posted 2 days ago
    I am running a model with a grouping variable (Control, Experimental) and Time (0,1,2,3). I want to treat Time as categorical (dummy variables). In addition, I would like to be able to have random slopes. So ... I expect my output to show the variances of the slopes for each dummy variable (e.g., var(0v3), var(1v3), etc.). However, I have been unable to find the correct code/covariance structure that will produce the desired output.
    I can get the proper output from R; below we can clearly see the level 1 residual SD, random intercept SD, and the SDs for each dummy variable "slope". Time 3 is treated as the reference category.
    m <- lme(dv ~ Time*Group, random = ~ Time|Subj, method="REML", data=d)
    summary(m)
    Random effects:
    Formula: ~Time | Subj
    Structure: General positive-definite, Log-Cholesky parametrization
                          StdDev       Corr
    (Intercept) 59.79464 (Intr) Time0 Time1
    Time0         93.60771 -0.653
    Time1         72.42573 -0.277 0.710
    Time2         63.31527 -0.462 0.765 0.926
    Residual     22.84535
    Any help would be greatly appreciated!


    ------------------------------
    Robert Cribbie
    ------------------------------