The seven crimes were in three general categories: violence with theft, violence without theft, and forcible rape (does not fit into either of the other categories exclusively). The sample includes all instances of these offenses in police records from April 1, 1975 to March 31, 1977, for which there are descriptions of both victim and offender characteristics (as reported by the victim or by witnesses). The study yielded a matched sample of about 7,400 victims of violent crimes and their offenders, allowing for an analysis of correlations between victim and offender characteristics. The data support the hypothesis that nontheft violence is more likely than personal theft to arise out of interactions among peers. The importance of the role of similar victim and offender lifestyles seems especially evident with nontheft violence, while violence with theft ties in better with the exposure component of the Hindelang, et al model. The general patterns are that different aged offenders do not discriminate by the sex or race of the potential victim. Neither do offenders of different races disproportionately attack certain ages or sexes. On the other hand, female offenders are slightly more inclined than male offenders to victimize younger and nonwhite victims. Most of the exceptions to these patterns involve weaponless thefts. For these crimes, blacks are especially likely to be attacked by young offenders, and young victims are especially likely to be attacked by females. These relationships need to be examined within sex, age, race, and crime-specific categories before more refined conclusions can be reached. Tabular data and 10 references are provided.