'Precaution' was scored as the number of self-help measures used among eight possible measures, and nine vulnerability measures were controlled. Regression analyses allowed the effects of precaution and prior victimization to be estimated independently of the effects they share with one another and with the vulnerability factors. Analyses indicate that when examined individually, none of the precautionary measures were associated with reduced frequencies of victimization in the following year. Concurrent measures of precaution and fear were correlated, indicating either that high fear leads to precaution or that precaution reinforces the fear. These findings suggest that the promotion of self-help precautionary measures independent of other crime prevention tactics is insufficient to reduce crime and the fear of crime. Alternative strategies which focus less on the security of individual households and more on social and physical aspects of the community should be explored. 4 tables, 53 references, and appended study instruments. (Author summary modified)