In a true experiment, perceived controllability of rape was manipulated thorugh a pamphlet rating paradigm (N=60). Women who saw materials that presented rape as an uncontrollable event reported greater anxiety on the street and in their homes and greater intention to perform precautionary behaviors than women who read materials that presented rape as somewhat more controllable. In a quasi-experiement, women (58) were classified as viewing rape as controllable or uncontrollable based on their descriptions of the 'typical rape.' Women who viewed rape as an uncontrollable event reported higher risk of rape and, unlike respondents in study 1, reported less use of precautionary behaviors on the street than women who viewed rape as a somewhat controllable event. In study 3, women (40) were classified as perceiving high or low controllability based on pretest responses. As in study 2, women who reported lower control perceptions also reported more fear and less precautionary behavior. As predicted, the high-control women reported significantly higher levels of fear after being exposed to the low-control message used in study 1. These women, however, did not increase their intended self-protective behavior. These results are discussed in terms of reactance and learned helplessness. 4 tables, 20 references. (Author abstract modified)