The relationship between subjective expected utility (SEU) for intercourse and sexual behavior was examined in a 3-year panel study of adolescents. Linear structural equation models were used to evaluate the causal priority of these variables for 225 junior high school students. Panel data allowed this examination, which is not possible in more common cross-sectional studies. It was found that SEU had a significant relationship to subsequent sexual behavior in each of the 1-year intervals, but intercourse was significantly related to subsequent SEU in only the first interval. This pattern was explained as a result of the process of development of utility during adolescent initial experience with sexual behaviors.