Minimum detectable effect size for main outcomes (accounting for sample
design and clustering)
Firm-Side: We measure each participant's prediction error (predicted minus true compensation), standardized so effects are in SD units. We average each participant's 3 errors within each condition, yielding one observation per person per condition, and test for a difference in means between treatment and passive control via a paired t-test. With 50 participants, ICC = 0.3, 80% power, and α = 0.05, the MDE is 0.27 SD.