Physical education programs have the potential to promote physical activity levels among school age students (Wallhead & Buckworth, 2004). However, research has documented poor student motivation for active participation in physical education. To understand students' motivation toward physical education, the social cognitive paradigm of achievement motivation has been extensively utilized in the past decade. Within this paradigm, three factors are considered as important for motivation, namely, self-efficacy (Bandura, 1997), task values (Wigfield & Eccles, 2000), and outcome expectancy (Dawson et al., 2001). With respect to outcome expectancy, Dawson et al. (2001) identified two dimensions: outcome likelihood and outcome values. To date, however, few studies have been conducted that combine these factors to study student achievement in physical education. Therefore, the first aim of this study was to explore the relationships among students' self-efficacy, task values, outcome likelihood, outcome values, intention for future participation, and effort/persistence in physical education. The second aim was to examine the potential gender and grade differences among these variables. Participants were 252 sixth to eighth graders (118 boys, 134 girls) from a suburban public school. During the last week of instruction students completed questionnaires assessing their self-efficacy, task values, outcome likelihood, outcome values, and intention for future participation in physical education. Physical education teachers reported their students' effort/persistence with an assessment instrument used in a study by Ferrer-Caja and Weiss (2000). Correlation analyses yielded that self-efficacy, outcome likelihood and intention were positively related to effort/persistence (r = .13 - .20), and all the other relationships were significantly positive (r = .28 - .78). Multiple regression analyses revealed that task values and self-efficacy were positive predictors for intention, accounting for 29.09% and 3.62% of the variance respectively. For effort/persistence, however, self-efficacy emerged as the only predictor, explaining 4.51% of the variance. A 2 * 3 (gender* grade) MANOVA yielded significant main effects for gender (Wilks' lambda = .91, F_6,235 = 3.74, p < .01), and grade (Wilks' lambda = .87, F_12,470 = 2.85, p < .01). When compared to girls, boys reported significantly higher self-efficacy (p < .05) and intention (p < .01). It was also found that sixth graders scored higher than eighth graders on task values (p < .05) and effort/persistence (p < .01). The results suggested that self-efficacy and task values were two important contributors for students' motivation, and gender and grade differences should be taken into account when addressing students' motivation.