Abstract Following Eysenck (1967), Gray (1973) and Wakefield (1979), it was hypothesized that the arithmetic achievement of extraverts would be better than introverts in those classrooms in which teacher-presented rewards predominated, and the arithmetic achievement of introverts would be better than extraverts in those classrooms in which teacher-presented punishment predominated. Five classrooms were rank-ordered according to their ratios of teacher-presented reward to teacher-presented punishment. A total of 101 fourth and fifth grade students from the five classrooms were then administered the Junior Eysenck Personality Questionnaire (JEPQ) and an arithmetic pretest. Subsequent to 40 school days, the students were administered the arithmetic posttest. Partialling out arithmetic pretest scores, lie scale scores from the JEPQ, and total arithmetic raw scores from the Comprehensive Test of Basic Skills (CTBS), regression analysis of arithmetic posttest scores with the interaction of extraversion scores from the JEPQ and classroom rank-order of the ratio of teacher-presented reward to teacher-presented punishment yielded significant results supporting the hypothesis. Additional analysis indicated the presence of non-hypothesized interactions of the Psychoticism and Neuroticism scales of the JEPQ with reward-punishment and extraversion.