College of Natural Sciences
 
FAQs
This is for IE7 to hold div open

General FAQ #20: Simple main effects tests

Question:

How can I carry out a simple main effects test using either SAS or SPSS?

Answer:

You can use either SAS or SPSS to conduct these tests. For example, let's say that you had a fairly straightforward completely between-subjects design, with a dependent variable called Y and two categorical predictors, A and B, each with two levels. Suppose that you wanted to test the effect of A for each level of B.

In SAS release 6.11 or higher, you can use the SLICE option in PROC GLM, like this:

PROC GLM ;
  CLASS a b ;
  MODEL y = a b a*b ;
  LSMEANS a*b / SLICE=b;

If you are using SPSS, you can use the MANOVA command to test the same hypotheses as the SAS program shown above:

MANOVA y BY a(1,2) b(1,2)
  /design = B A within b(1) A within b(2).

More complex designs are testable with the recent SPSS GLM and SAS MIXED procedures, including within-subjects simple main effects tests.

If you have further questions, send E-mail to stats@ssc.utexas.edu.