This study was designed to clarify findings from the parents' beliefs literature concerning the nature and the effects of parental accuracy. Subjects were 60 second- and fifth-grade children (mean ages = 8–5 and 11–5, respectively), their mothers, and teachers. Each child responded to 4 cognitive tasks, predicted how he or she would perform on each task, made self-assessments of preferences and personality traits, and predicted the response of 2 peers to all of the measures. Both the child's mother and the child's teacher were shown the tasks and made similar predictions of the child's probable response. As in previous studies, mothers were above chance but far from perfect in predicting their children's cognitive performance and preferences and traits. Teachers were as accurate as mothers in judging cognitive abilities; they were less successful, however, on the preference and personality items. The children's predictions, whether for self or peers, were less accurate than those of the adults on all tasks. The accuracy of the mothers' predictions correlated positively with external measures of the child's cognitive performance. This finding is compatible with the Hunt “match hypothesis,” although other bases for the correlations also exist.