This article presents findings from a study investigating the effects of a middle school counseling intervention, using a specially designed curriculum, on participating students' attitudes, self-efficacy, and performance in mathematics. Hierarchical linear modeling was used to test for main effects between condition and sex and condition x sex interaction effects on all outcome measures across 3 time points: pre-intervention, post-intervention, and follow-up. At pre-intervention, boys were significantly higher than girls on all variables except for performance in math. Models for growth indicated that students in the experimental group improved at a significant rate compared to girls in the control group on motivation, value, enjoyment, and confidence in mathematics. Implications of the findings are therefore particularly relevant for girls, who may benefit from interventions designed to increase their attitudes and self-efficacy for mathematics.