The meta-analytic by Mertens et al. (1) interprets nudges as a generally effective technique for increasing desirable decision-making, with an overall pooled effect size of d = 0.43. This research also reports large systematic variations (meta-analytic heterogeneity) in effects, primarily attributed to moderators such as the domain, as well as asymmetrically distributed effects, interpreted as moderate publication bias. Apart from publication bias, non-normality and high heterogeneity may be problematic for the representativeness of meta-analytic means (2). Here, we reanalyze the corrected data made available by Mertens et al. (1), finding evidence that nudges have more limited than general effectiveness. We show that effects are clearly left-truncated, likely due to substantial publication bias, consistent with another reanalysis (3). We also find that most of the pooled effects as reported in Mertens et al. (1) are overestimated and hence unrepresentative.
展开▼