Spanking and externalizing problems: Examining within-subject associations.

This study examined the effects of spanking on externalizing on a within-subject level, while excluding causally irrelevant between-subject variance. Results from two longitudinal studies which used participants from the Child Development Project (n = 585) were reanalyzed with a random-intercept cross-lagged panel model using yearly measurements over ages 6-8. After removing between-subject variance, there were no significant effects of general spanking on externalizing (β = .06, .07). However, when done without objects and at a rate of about once per month or less, spanking showed beneficial effects (β = -.17, -.21). Results suggest that previous findings may be due to a failure to separate between-subject and within-subject variance. Additionally, results illustrate the need to examine limited spanking separately from more general forms of physical punishment.