Minimum detectable effect size for main outcomes (accounting for sample
design and clustering)
We have the following conservative MDEs for a main outcome. We assume 80% power and a significance level of 0.05. We also account for clusters of 10 (10 resumes per participant) and 250 participants, with treatment and control (female and male resumes) presented in equal proportion. We use standard deviation estimates from our pilot.
-Log final total compensation: 7.9% (clustered with an ICC of 0.3), 4.1% (unclustered)