Objectives: It is hypothesized that parents who have physically abused a child or who are at high risk to physically abuse a child form more negative/less positive impressions about children than nonabusive/low-risk-for-abuse parents. However, to date, the overall evidence from the several individual studies testing this hypothesis has been inconclusive. Method: The current meta-analysis compared the impressions that abusive/high-risk-for-abuse parents and nonabusive/low-risk-for-abuse parents form about children (k = 27, N = 1,645). Results: Overall, we estimated the difference in parents’ impressions of children to be about 1/3 of 1 standard deviation in the hypothesized direction (d = 0.32, 95% confidence interval [0.19, 0.44]). This difference was found among studies using a variety of procedures and a variety of dependent variables. Further, an analysis of the distribution of p values from the published studies corroborated this finding. Conclusions: Collectively, the current meta-analysis confirms differences between abusive/high-risk-for-abuse parents’ and nonabusive/low-risk-for-abuse parents’ impressions of children. However, despite the clear evidence for the existence of an effect, the current results suggest the magnitude of this effect is modest. Future research should determine whether parents’ impressions are meaningfully related to parents’ child-directed behaviors. Such research would be informative as to whether parents’ impressions of children should be targeted in interventions aimed at reducing physically abusive parenting behaviors.