The purpose of this study was to investigate the effect of cosmetics use on attributions concerning the likelihood of provoking sexual harassment and of being sexually harassed. Subjects were 85 female and 76 male undergraduate volunteers. The study was a 3×2 between-subjects design with three levels of cosmetics use (heavy, moderate, no cosmetics) and two levels of sex of subject (male, female). Each subject viewed one of three colored photographs of a model wearing either heavy, moderate, or no cosmetics, and then indicated how likely the model was to provoke sexual harassment and to be sexually harassed. Data were analyzed using analyses of variance and the Newman-Keuls test. When the model wore heavy cosmetics, she was rated as more likely to provoke sexual harassment than when she wore moderate cosmetics. Similarly, when the model wore moderate cosmetics, she was rated as significantly more likely to provoke sexual harassment than when she was not wearing cosmetics. When the model wore either heavy or moderate cosmetics, she was also rated as more likely to be sexually harassed than when she did not wear cosmetics. In addition, male subjects rated the model as more likely to provoke and to be sexually harassed than did female subjects. Results are discussed in terms of sex role spillover.