The overjustification hypothesis predicts decreased intrinsic motivation when persons are paid to perform an interesting task. The factors of reward experience, socioeconomic status (SES), and sex are examined while testing conflicting predictions of the hypothesis and reinforcement theory. Children from grade 1 at two public elementary schools worked on a counting task. Half of the children attended a school which had a 100% minority enrollment (low SES), and half attended a school which had a 95% Caucasian enrollment (middle SES). A Baseline 1, treatment session, Baseline 2 paradigm was used to assess the effects of both reward introduction and withdrawal. Reward experience (group), SES, and sex were manipulated in a 3 x 2 x 2 design. Subjects who earned a reward during the treatment session for maintaining Baseline 1 output levels, significantly outperformed never-rewarded controls during Baseline 2. Control subjects initially highest on motivation measures declined significantly from Baseline 1 to Baseline 2; comparable subjects in the reward groups showed no change. In addition, reward-group subjects that were initially lowest in motivation significantly increased their output and time on task across sessions. (Author/RL)