In this study, the relative fits of three different factor-structure models of adolescent reckless behavior were examined using the Reckless Behavior Questionnaire (RBQ) with individual samples of college and high school students. Both one- and two-factor models were found to be satisfactory representations of the RBQ with both samples. In order to test the construct validity of the one- and two-factor models, relations between instruments generally associated with reckless behavior were examined by gender. Using the two-factor model, gender differences were found for both the college and high school samples; thus, it was determined to be the more parsimonious fit of the data given previous research supporting gender differences. Findings are discussed in terms of current conceptualizations of factor patterns of adolescent problem behavior and implications for future investigations.