The paper actually explicitly addresses this: > However, all intervention categories and domains apart from “finance” show evidence for heterogeneity, which im

The paper actually explicitly addresses this: > However, all intervention categories and domains apart from “finance” show evidence for heterogeneity, which implies that some nudges might be effective, even when there is evidence against the mean effect.