Research Findings: Prosocial behavior in early childhood can be assessed by various informants and methodologies, but the information each provides may differ in meaningful ways. The present study measured and compared young children's (N = 486; age: M = 48.51 months, SD = 8.22) prosocial behavior based on parent reports, teacher reports, and child behavior on the Dictator Game. Special consideration of the multidimensional nature of prosocial behavior was given by including questionnaire subscales for helping, sharing, and comforting behaviors, and differences across gender were investigated using a multigroup framework. Analyses found differences in the pattern of responding between parents and teachers and no agreement across informants. Observed sharing was not associated with either parent- or teacher-reported prosocial behavior. Gender differences were minimal, with girls rated as demonstrating more comforting behavior than boys by both parents and teachers and more overall prosocial behavior by teachers. Practice or Policy: Results suggest that methodology has a strong impact on the measurement of prosocial behavior. Children's prosocial behavior may differ across contexts or vary by informant perspective, and this must be considered when interpreting results. These findings thus support the need for multi-informant and multimethod approaches in child assessment.