09-06-2012 11:44 AM
Can someone suggest to me the appropriate stat model in SAS for analysing data from a 4-way crossover study design?
4-way = 4-period, 4-treatment, 4-sequence, crossover.
I tried using GLM model but the test for sequence effect could not be computed.
09-07-2012 07:51 AM
A good source for this would be to look at the documentation for PROC PLAN, Example 68.7 Crossover Designs. Down towards the bottom, it presents an analysis using PROC ORTHOREG, that includes carryover (sequence) effects, that specifically depend on the previous treatment. Note that the analysis depends on using the EFFECT statement, so if you are not on a version of SAS/STAT that supports this statement (9.3 and 12.1 do), this may not be possible to run.
In any case, the model statement should give you what you need to get started.