Minimum detectable effect size for main outcomes (accounting for sample
design and clustering)
Due to high attrition between phases, this study expects to be underpowered for detecting changes in the self-reported growth mindset measures (typical MDES=0.3).
Considering baseline measures of choices for competition pooling both women and men in the study (0.38), this study is expected to be sufficiently powered for an MDES of 0.2p.p. for the treatment group.
Considering baseline measures of female choices for competition, this study is sufficiently powered to find an increase of at least 0.25p.p. on competition choices by women in the treatment group, closing a substantial portion of the gender gap found at baseline (0.31p.p.).
Since we do not have a baseline measure of the challenge-seeking choice exercise used in phase 3, we can only estimate the ranges of MDES associated with the intervention. For the entire sample, MDES ranges include 15p.p. for a control mean of 0.1, 20p.p. for a control mean of 0.5, and 0.08p.p. for a control mean of 0.9. For the sample of women, in particular, MDES ranges include 22p.p., 27p.p., and 18p.p. for control means of 0.1, 0.5, and 0.8 respectively.