Four controlled trials, published between 1981 and 1990, exert outsized influence in the debate about spanking among academics, professional organizations, lawmakers, and jurists. The trials have also been analyzed in five meta-analyses. Advocates of conditional spanking claim that the trials validate its effectiveness. As such, they argue, parents should consider conditional spanking as a disciplinary strategy and there should be no blanket legal ban. This article evaluates the controlled spanking trials, individually and in composite, to determine their quality and evaluate their effect sizes. The article evaluates the internal and external validity of the trials and compares the impact of spanking on child compliance through a series of meta-analyses. The objective is to determine whether the trials support claims that spanking more effectively reduces child noncompliance than other conditions assessed in these trials. We reviewed the trials, including formal evaluation of risk for bias. We also surveyed the five meta-analyses that incorporate the trials. Finally, we conducted a series of meta-analyses, comparing spanking to comparator conditions assessed in the trials. We found that the trials lack internal and external validity. The meta-analyses that incorporate these trials usually do not assess spanking against other cooperation-eliciting strategies, are sometimes flawed, and reach inconsistent conclusions. The current analyses showed no significant difference between spanking and other cooperation-eliciting strategies combined; g = 0.24, standard error = 0.26, t (27) = 0.91, p = .37).
Atkinson et al. (Tue,) studied this question.