The objective of the study was to improve physical activity (PA) surveillance of the Healthy People 2010 Objective 22:10 (i.e., 50% of the lesson time engaged in PA) by establishing a pedometer steps/min guideline to quantify time engaged in PA during physical education. A sample of 180 middle school students had their PA measured via pedometry (steps/min) and behavioral observation (PA time). Factorial analyses of variance were used to examine PA differences. Linear and logistic regression, decision accuracy, and receiver-operating-characteristic (ROC) statistics were used to test steps/min cut points against the 50% PA recommendation. PA differences were not found (p greater than 0.01). Steps/min was a significant (p [less than or equal to] 0.01) predictor of PA time, and the binary outcome of meeting or not meeting the PA recommendation. A steps/min interval of 82-88 was an accurate indicator of the 50% PA recommendation. The ROC statistic was 0.97 (p [less than or equal to] 0.01), suggesting steps/min was an excellent discriminator of the binary outcome. Pedometer steps/min is a valid, objective, and practical approach for surveillance of physical education PA, a key physical education and public health outcome. (Contains 1 figure and 4 tables.)