This study examined the reliability and construct validity of two types of measures of mastery motivation for elementary school children: a new version of the Dimensions of Mastery Questionnaires (DMQ) and behavioral mastery tasks. Participating were 64 mostly middle class and Caucasian 7- and 10-year-olds living in a middle-sized western city. Mothers and teachers rated children on the DMQ; the DMQ scales were also administered orally to the children. Children also completed four sets of individualized mastery tasks, each set with five difficulty levels. Mothers and children completed the scholastic, athletic, and peer acceptance scales of Harter's Self-Perceived Competence scales. Preference for challenge and independent mastery were rated by the teachers and child. Findings indicated that most of the correlations across raters between parallel DMQ persistence scales and Harter Perceived Competence Scales were significant, but modest, with higher parent-teacher correlations than for adults with child's self-reports. There were modest, significant correlations between child-report of DMQ total persistence and total persistence at the behavioral tasks and preference for challenge. Child and teacher DMQ mastery pleasure scales were correlated with observed pleasure during the tasks. Child DMQ total persistence, parent DMQ cognitive/object persistence, and the child's rating of peer acceptance as unimportant combined to predict the child's overall behavioral task persistence. Children's task behaviors and parent DMQ ratings predicted school behavior as indicated by teacher ratings, but children's DMQ ratings did not. (KB)