Data were obtained from the National Crime Survey, which examines crime victimization aspects for a national sample, for 1973-76. Only respondents 65 years-old and older were included in this analysis. A multivariate model was used to examine the risk of personal victimization among the elderly. The wide variation in the risk of personal victimization among the elderly is impressive. Accordingly, caution must be used when interpreting age-specific victimization rates. The low observed rate of victimization for the elderly may result from the high level of fear of crime that causes the elderly to curtail activities substantially. The real risk of victimization, therefore, should be computed by controlling for situational exposure to victimization. Victimization was found to be a person-to-person activity, with victims often being alone at the time of the offense. The development of more extensive social contacts by the elderly would help eliminate victimization related to isolation. One-third of the 1,069 elderly victims were attacked in the course of the victimization, with the predominant injuries being bruises, 'black eyes,' and cuts. Nearly half required medical attention; most of these were covered by some type of medical insurance. The data suggest that nonviolent methods, such as reasoning with an offender, were more likely to be effective in preventing injury. The seriousness of the crime was the only variable that showed a positive effect on the likelihood of crime reporting. Tabular data and 36 bibliographic entries are provided.