This study investigated the effects on student learning and attitudes of a physical education instructor appearing overweight versus lean, and including role modeling of exercise behavior versus "Do as I say" (only) modes of instruction. High school students (N=107) were randomly assigned to watch one of four videotapes of the same instructor teaching health-related fitness concepts using the same script. In two tapes the instructor wore a "fat suit" so as to appear overfat. In one of those tapes he role modeled exercise participation to emphasize its importance for health—while he merely instructed in "Do as I say" mode in the other tape. Two more tapes did the same except that the instructor appeared lean. After viewing the instructional tapes, the students’ learning was tested, and they completed questionnaire items tapping their attitudes to the instructor and his effectiveness. A manipulation check showed that the fat suit convincingly made the instructor appear fat versus lean in the videos (F [1, 104]=21.48, p < .001; ES=.82 SD). Subsequent 2 x 2 (appearance x role modeling) ANOVA and MANOVA analyses indicated that role modeling exercise behavior (versus "Do as I say" only) had a positive effect on the learning and attitudes of the participants, but there were no effects for fat vs. lean appearance. Students who viewed the role modeling tapes scored higher on the knowledge test (F [1, 103]=11.01, p < .001; ES=.62 SD), and they reported more positive attitudes (F [3, 100]=2.70, p < .05) than the students who viewed the "Do as I say" tapes. Specifically, students liked the role modeling instructor better, (p < .01, ES=.54 SD), believed him to be more of a role model who actually practiced what he preached (p=.05, ES=.39 SD), and more highly endorsed an item "regarding whether physical education instructors should be good role models" (F [1, 102]=5.56, p < .05, ES=.45 SD). Unlike previous studies (e.g., Melville & Maddalozzo, 1988), this experiment showed no negative effects on cognitive learning or attitudes from the instructor appearing to be overfat. In contrast, these data suggest that, irrespective of their appearance, instructors can more effectively influence health-related fitness learning and attitudes by including role modeling of desired exercise behavior than by merely using a "Do as I say" mode of instruction.Keyword(s): exercise/fitness, high school issues, national issues