Raw data can lead to half-baked conclusions. Harriet MacMillan and colleagues1 isolated a single marker, spanking, from the complexities of childhood. They then established a correlation with adult psychiatric disorders. Considering the multitude of "confounding variables unaccounted for in [their] study" it is surprising to me that the study was even published. Just think of some of them: parental alcoholism or psychiatric disorder, poverty, family size, birth order, socioeconomic status and education.
What is infinitely more disturbing to me is to see Murray Straus pick such a lame-duck study, which itself admits that "[its] nature ... precludes comment on the causal role of slapping and spanking for psychiatric disorder," call it "definitive evidence ... on the potential benefits of not spanking"2 and use it to suggest making spanking illegal.