Expectancy-value theory asserts that intrinsic motivation is enhanced by values that one perceives in a behavior and is hindered by cost perceived associated with the behavior. Perceived values and cost play a significant longitudinal role in motivation in classrooms and in sports (Jacob et al., 2002). The purpose of this study was to examine the expectancy-value components (attainment, interest, utility values and cost; Eccels, 1987) in a mandatory university physical education setting and as related to self-initiated physical activity. The data were collected individually from a random sample of 386 university students in China's key (highest ranking) universities where students are constantly under pressure to demonstrate the highest academic achievement. Self initiated physical activity was measured using Bouchard Three-Day Physical Activity Survey (Bouchard, 1987). The participants were taught to log in their activities each quarter hour every day for 3 days. The expectancy-values were measured using a modified version of Expectancy-Value Inventory (Eccles & Wigfield, 1995). Both instruments were translated into Chinese from English and were validated by bi-lingual Chinese-American scholars (n = 6) in physical education. Construct validation analysis showed high integrity of the expectancy-value structure with high factor loadings (.64 - .99) for each value category. Constant comparison analysis on the open-ended survey data of cost revealed that 45% students thought the physical education curriculum did not meet their needs, which was de-motivating. Decisions on continuing physical education were primarily based on health benefits (63% of students) and learning knowledge/skills (33%). Regression showed that daily self-initiated (non-physical education) physical activity was motivated by the Attainment Value (important to health, R2 = .20), while the decision to take physical education (Yes or No) was determined by the Interest Value (B = .80, logistic regression coefficient) and the Utility Value (B = .70). Interestingly, students' motivation and their likes/dislikes in physical education classes were not associated with their self-initiated daily physical activities (×2 = .003, p = .956). These results suggest that for students in these Chinese universities, motivation for physical education and motivation for self-initiated physical activity derives from different sources: attainment value plays a part in motivation for self-initiated physical activity, while interest and utility values of the content are the primary motivators for physical education. The findings imply that motivation specificity (Bong, 2001) should be considered when developing strategies to motivate college students to participate in physical education classes or to develop an active lifestyle.Keyword(s): NA