Although many cultivation studies claim to measure fear of victimization, the estimates of the probability of victimization and personal safety constitute the measures of cultivation effects. Data from a random telephone survey of 101 respondents from a midwestern State indicate that the fear of violence index and the four traditional cultivation measures were not highly correlated and that television viewing was related modestly to fear of violence, but not crime probability. These results support the necessity for a conceptual distinction between items that measure fear of crime and items that measure probability of criminal victimization. Also, the results of cultivation analysis may vary considerably with the dependent variables. 5 notes and 11 references (Author abstract modified)