Is there a way to conduct contrast tests in the GENLIN procedure

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 1
Joined: Sun Mar 27, 2016 3:20 am

Is there a way to conduct contrast tests in the GENLIN procedure

Postby eoc124 » Sun Mar 27, 2016 3:55 am


I am currently conducting some longitudinal mixed modeling in a dataset with both continuous and binary outcomes. For this research, we conducted an intervention and we are interested in assessing the effects of the intervention on various outcome variables. For the continuous outcomes I am using the MIXED function in SPSS and for the binary outcomes I'm using the GENLIN function to run Generalized Estimating Equations. For these analyses, we were interested in comparing the mean difference between our intervention vs. control groups (Tx: Intervention = 1, Control = 0) in the amount that each group changed in our outcome variables (e.g., depression) from the baseline assessment (Time = 0) to the time that each participant completed various follow-up assessments (3.8M, 9.1M and 16.2 respectively). For time - we are using the exact time that participants completed each wave (as months since they completed the baseline assessment) and we are including a quadratic term.

For the continuous variables, I have been able to use the TEST subcommand in SPSS to calculate whether there are mean differences between our treatment and control group in the amount that each group has changed from baseline to each follow-up assessment (see syntax below).

MIXED Depression WITH TX Time
/FIXED Tx Time Time*Time Tx*Time Tx*Time*Time
/Test 'Group Difference in Change in Depression at Time 3.8M' Tx*Time 3.8 Tx* Time*Time 14.44
/Test 'Group Difference in Change in Depression at Time 9.1M' Tx*Time 9.1 Tx* Time*Time 82.81
/Test 'Group Difference in Change in Depression at Time 16.2M' Tx*Time 16.2 Tx* Time*Time 262.44

When I looked at the syntax commands in SPSS, there doesn't seem to be the TEST subcommand option or the LMATRIX subcommand option for either the GENLIN or GENLINMIXED procedures. Does anyone have any suggestions of how I can run these contrasts tests? (I know in SAS the comparable function would be the ESTIMATE function - should I just use another statistical software package?)

Any help would be greatly appreciated!

Who is online

Users browsing this forum: No registered users and 1 guest