Grounded in symbolic interactionism and drawing on data gathered in the 2007 Monitoring the Future Study (n = 2,201), this research examines how self-esteem and perceived intelligence, as well as attitudes and behaviors related to school environments, associate with perceptions of anabolic-androgenic steroids. With perceived risk and approval/disapproval of steroid use as dependent measures, logistic regression models identified explanatory effects for sex, race, newspaper exposure, and peer use of steroids, with males, Whites, those who seldom read the newspaper, and those with friends who had used steroids estimating lower levels of risk and indicating less disapproval. Those who viewed themselves as being more intelligent than their peers as well as those with higher levels of self-esteem estimated higher levels of risk and expressed greater disapproval, whereas those who indicated "hating" school and those who had skipped school the most estimated lower risk and less disapproval. Logit log-linear models identified the most parsimonious representations of variable relationships, demonstrating how the presence of a self-esteem index stands to reduce the centrality of sex and race as explanatory measures. (Contains 4 tables.)