With the development of cheaper software and more efficiently designed keypad and receiver units, the popularity of classroom response systems (CRS) has been steadily increasing since the mid- 1990's. In health education, CRS has the potential to radically enhance the dynamics of instruction, interaction, assessment, and comprehension of concepts. CRS use in the context of health education can: (a) allow students to anonymously submit sensitive personal information during in-class assessment of health related behavior; (b) encourage discussions on sensitive or personal issues based on polling data; (c) provide instantaneous comparative analysis of health behavior (individual or group) using class demographics; and (d) foster group level critical thinking using a combination of immediate feedback and appropriate guidance. Potential notwithstanding, any new CRS must be validated and examined for efficiency, applicability, and ease of adaptation. The purpose of this study was to evaluate the validity of a CRS as a data collection instrument when administering questions selected from the Youth Risk Behavior Survey (YRBS) in a college health education classroom. A convenience sample of 170 participants from 6 introductory health and wellness classes participated in the study. Participants completed a 42 item paper-based questionnaire of which 37 items were selected from the 2001 YRBS. After completing the paper-based survey, participants transferred their answers using a CRS. Answers were analyzed to determine the extent of criterion-referenced validity. To establish criterion-referenced validity (i.e., classification accuracy) evidence, the contingency coefficient was calculated between the CRS (TurningPoint, 2006) and the paper-based survey. The responses from the paper-based questionnaire served as the criterion measure. The proportion of agreement between the two measures was calculated for each variable (i.e., a total of 37 contingency coefficients) and the mean and standard deviation of those contingency coefficients were presented. The average of 37 contingency coefficients was .975 with a standard deviation of .017. Considering that a value of the contingency coefficient close to 1 is most desirable, the average contingency coefficient of .975 indicates a high accuracy of classifications between the two measures. The range of 37 contingency coefficients was .060 with maximum and minimum values of 1 and .940, respectively, which shows a little variation of validity estimates among the 37 variables. Thus, this CRS is a valid alternative to paper-based assessment when dealing with closed ended questions.Keyword(s): assessment, health education college/univ, measurement/evaluation