Minimum detectable effect size for main outcomes (accounting for sample
design and clustering)
The power analysis calculations below were computed to examine if there is enough statistical power to detect effects if the Bonus program was in fact effective. Power analysis is typically used for impact evaluations and, as such, this power analysis corresponds only to RQ 2. We used the command power in STATA assuming power of 80% and alpha of 5%, and input outcome means and standard deviations using data from the previous SMHR 5-year program (Garneau-Rosner & Herzog, 2015-2020) to compute the minimum detectable effects of a change in means in key measures of relationship decision making, relationship attitudes, parenting outcomes, perceived stress, and personal resilience/optimism. Our calculations show that the minimum detectable effects for these measures range from 0.14 to 0.25 for the majority of the outcome variables. Please notice that Table 2 only includes variables that we collected in a previous program and not all of the variables that we will use in the present studyout of 12 outcome variables . The outcome variables that we will examine in the present study can be found in Appendix D. Consistent with the literature which typically uses a minimum detectable effect size of 0.25, and following Wolf (1986) who reported 0.25 as a strong enough effect to indicate a practical change in non-clinical samples when evaluating educational programs, our final sample size will have sufficient power to detect effects over the full course of the RCT.