Minimum detectable effect size for main outcomes (accounting for sample
design and clustering)
For stunting, the aforementioned design yields a minimum detectable effect size of 8.6 percentage points (a 14% decline over a baseline incidence of 62%) when comparing Control and any of the two treatment arms. For comparisons between the two treatment arms, the estimated effect size is 9.2 pp. For height-for-age z-score, the corresponding figures are -0.19 SD and -0.20 SD, respectively. The relevant parameters are fixed as follows: 80% power, 5% significance, and intra-cluster correlations of 0.07 (stunting) and 0.10 (height-for-age z-score) (based on data on rural Sindh from Pakistan DHS 2017-18).