The reputation of psychological science depends on the adequacy of the science underlying its policy recommendations. This commentary raises concerns about the science used by Heilman et al. (2021) in their recent narrative (not meta-analytic) review that encourages spanking bans worldwide. By reviewing controlled longitudinal studies, Heilmann et al. provided stronger causal evidence than the two meta-analyses of unadjusted correlations most frequently cited to support spanking bans. However, the two previously published meta-analyses of controlled longitudinal studies of spanking do not support spanking bans, due to the trivial size of the average adverse-looking effect of customary spanking in those studies. Moreover, several lines of evidence indicate that this trivial average effect is likely due to inadequate statistical controls rather than an actual adverse causal effect of typical spanking. We need stronger causal evidence for policy recommendations for both the welfare of children and the reputation of psychological science.