An interrupted time-series design was used with switching replications. No consistent effect of television's introduction was observed for violent crimes, burglary, or auto theft. However, the introduction of television was consistently associated with increases in larceny, irrespective of whether television was introduced in 1951 or 1955 and irrespective of whether state- or city-level data were examined. Analyses of the early content of television indicate that the advertising of consumption goods was high, that upper-class and middle-class life-styles were overwhelmingly portrayed, and that larceny was portrayed much less often than crimes of violence in crime shows. The effect of television on larceny was tentatively attributed, therefore, to factors associated with viewing high levels of consumption-perhaps relative deprivation and frustration-rather than to factors associated with the social learning of larceny through viewing it on television. (Publisher abstract)