Hypothesizing an effect size by considering individual variation

Andrew Gelman, Amy Krefman, Lauren Kennedy, Jessica Hullman

When designing and evaluating an experiment or observational study, it is useful to have a realistic hypothesis regarding the average treatment effect. We present an approach to conceptualizing this average by first considering a distribution of effects. We demonstrate with examples in medicine, economics, and psychology.

Read on ELI