Minimum detectable effect size for main outcomes (accounting for sample
design and clustering)
On average over all discounts, the share of subjects choosing the item with the shorter expiration date is 11% (=MDES) higher in the treatment group than in the control group, amounting to a relative difference of 22%.