Repeated measures analyses (within-subjects designs)

[Sascha Saly]

Hello,

I am conducting a field research on 8 franchise coffee shops that are in the same geographical region and are owned by the same individual. Below is a brief description of my setting:

• I implemented a team-based intervention program whereby all employees at any of the 8 locations could earn recognition based on restaurant-level performance.
• The recognition program lasted for 16 weeks with recognition awarded every two weeks (i.e., 8 recognition periods) based on their performance for the previous two-week period. Because recognition was not based on relative performance, each location could be recognized up to 8 times over the 16 weeks of the program.
• I test my hypotheses using non-parametric (i.e., Wilcoxon signed-rank z-test) tests that compare dependent measures during the post-recognition periods (i.e., the 16 weeks subsequent to the recognition program) to dependent measures during the pre-recognition periods (i.e., 16 weeks prior to the recognition program). I used Wilcoxon signed-rank z-test because of the small sample size and the non-normal distribution of many of my dependent measures.

Basically, I have repeated measures design. I have two questions (1) Is the Wilcoxon signed-rank z-test descried above appropriate for this design?, and (2) If not, what are the most appropriate tests.


Thank you,