Hello,
I am working on a pre-analysis plan for an experiment that I want to conduct. The experiment is about productivity behaviour. Participants solve matrices and the number of matrices solved my primary outcome variable (i.e. productivity).
There are 2 groups:
1.) Treatment Group: Working under stimulus
2.) Control Group: Working without stimulus
In the beginning of the experiment, I conduct a baseline-phase to check whether potential differences in productivity may stem from differing baseline abilities.
My hypothesis states that productivity in both groups is the same.
What is the best way to investigate the hypothesis?
a) First check for differences in baseline ability and then conduct a nonparametric / parametric test?
b) Use a linear regression model, use a dummy on the treatment group and include baseline productivity as a control variable?
c) Is it even better to conduct both (e.g. Mann-Whitney U-Test and a subsequent linear regression) to arrive at more compelling results? Or would that approach even be counterproductive?
Kind regards and thanks for you help!