I have a hypothetical field experiment that I have to design for my master thesis. It is about implementing different team goal difficulty levels (A, B, C) in a company that currently has goal difficulty level X. My hypothesis are about how A, B, C are increasing/decreasing team performance compared to X and compared to each other.
In this hypothetical company, performance is measured on a quarterly basis (in % of goal attainment). I want to measure performance for at least four quarters after the implementation so that I can see how the goals really effect performance for more than just one quarter. Randomization is done on unit level and stratas are created based on prior performance to ensure comparability between treatment and control groups.
Now I'm confused if I can do a panel analysis and use a fixed effects regression like this:
Perfit = ai + ß1A + ß2B + ß3C + TimeFE + TeamFE + eit
Perfit denotes the performance of team i in quarter t. Our panel data set comprises eight waves and thus, I included the development of team performance with eight separate quarterly performance levels in this analysis. A, B and C are binary variables for whether the team is part of one of the three treatment groups. ai is the unobserved heterogeneity, which are constant over time but vary across teams. eit is the idiosyncratic error term that varies over time and teams.
Do you think this is a good way to analyze the hypothetical experiment? I do not need to actually do it in my master thesis, but I need to find the ideal design. Do you have any recommendations here?
I would really appreciate your help!!!!