A measure of prosocial choice in story dilemmas was created by adding 23 stories to an original 4 studied by Eisenberg-Berg and Hand (1979). Each story called upon the child to make either a prosocial or a hedonistic choice in the hypothetical situation. Adding more items improved the internal consistency of the measure, as predicted by the Spearman-Brown formula. In our Head Start sample, the children's prosocial responses significantly predicted more prosocial ratings of child behavior by parents. The same relationship held, although less consistently, with behavior ratings carried out by observers in the Head Start classroom. Parents' ratings of themselves on mental health and parenting variables were not related to the measure of prosocial choice in story dilemmas. These findings as a whole provide some exception to past findings of lack of correspondence between responses to hypothetical situations and real-life behavior.