Significance: During the adolescent years physical activity participation begins to decrease, often tracking into adulthood (Malina, 2001) and leading to obesity and other metabolic risks (American Obesity Association, 2005; President's Council on Physical Fitness and Sport [PCPFS], 2002). The theory of planned behavior explores variables that affect human behavior such as physical activity (Ajzen, & Fishbein,1980). Three beliefs (control, behavioral and normative) impact one's intent to execute an action such as physical activity behaviors. Physical education can act as a moderator of physical activity (Castelli & Erwin, in press); however, it remains unclear how specific instructional models may impact these behaviors. The purpose of this study is to examine physical activity beliefs across different instructional strategies. Design: The Personalized System of Instruction (PSI; Metzler, 2000), provides self-paced instruction permitting individualized decision making which may lead to greater feelings of control and subsequently improved physical activity engagement. From a Midwestern middle school physical education program, 149 participants (Mage = 12.8 ; SD = 1.111 ; 60 males) were randomly assigned to engage in a seven week PSI intervention entitled C.H.O.I.C.E. (Creating Healthy Opportunities in a Choice Environment) or a control group of direct instruction. Fidelity of the instruction models was confirmed through rubrics and observations (Metzler, 2000). C.H.O.I.C.E. participants used media focusing on health-related fitness content and goal setting, while the control group participated in health-related fitness and direct instruction. All participants completed pre and post physical activity beliefs surveys (Saunders et al, 1997), previous day physical activity recalls (Weston, Petosa, & Pate, 1997), and a physical education survey. Ten participants were also selected for follow-up interviews. Results: The pre test data analysis revealed that the groups had similar beliefs. Post test results revealed no statistically significant differences in post beliefs between the PSI (Mean = 58.80; SD = 9.775) and control group (Mean = 55.97; SD = 9.747; p <.01); however, there were positive trends toward increased control, attitudes, and norms, as all of the PSI groups had higher post beliefs than the direct instruction group. An ANOVA revealed eighth grade participants displayed a statistically significant (p < .01) increase from pre to post in overall physical activity beliefs. Though overall beliefs were not statistically significant for all grades, it is noted that PSI may have the potential to modify perceptions of planned behaviors such as physical activity with a longer intervention.