While the argument that mass media exert direct causal effects on perceptions of crime possesses intuitive appeal, it is not strongly supported by empirical research. As part of a larger survey to assess a mass media crime prevention program, this study analyzed responses to questionnaire items on the number of newspapers read and hours spent watching television, fear of crime, perception of city safety, and attention to crime problems. Respondents were over 18 years old, chosen according to both stratified sampling procedures and random dialing, and interviewed by telephone. Control variables included sex, age, socioeconomic status, community size, victim experience, and attitudes toward the media. The analyses failed to support the hypothesis of direct media influence. Moreover, responses to a question 'have you heard or read anything that makes you fear for your personal safety or the safety of your property in this neighborhood' showed no correlation with media consumption. These findings may indicate that public perceptions of crime are diffuse phenomena that reflect general anxieties concerning the social and political environment and thus are difficult to measure through surveys. In this framework, the role of the mass media as disseminators of crime content becomes less salient and informal communications, such as gossip and rumor, appear more important. Tables, 4 footnotes, and 35 references are provided.