In attempts to make assessments less threatening, children are sometimes asked to respond to questions about another child rather than about themselves. Little is known about how this manipulation of response format (self versus other) might influence children's responses. This study compared responses of 58 younger (5-7 years) and 68 older (8-11 years) boys who made attributions about child behavior. Boys reported either as they themselves would respond to a situation described in the measure (self condition) or as they thought a hypothetical other child (other condition) would respond. Some situations were positive and some were negative. Consistent with hypotheses, older boys provided more internal attributions than did younger boys; boys in both age groups attributed negative events more to internal causes if they responded as a hypothetical other child, as compared to responding as themselves. Suggesting specificity of the self-serving bias to attributions, boys who were responding as a hypothetical ''other'' child described engagement and involvement with their mothers more positively than did those who were responding as themselves. In addition, few differences were found in responses of boys diagnosed with ADHD compared to those without known problems. The implications for having children respond to rating scales as another child in both clinical and research settings are discussed.