This hypothesis was tested by comparing the results of two national Canadian surveys: the 1991 post-censal Aboriginal People's Survey, which involved a sample of over 18,000 respondents, and the 1993 Canadian General Social Survey, which involved over 10,000 respondents. Contingency tables (cross-tabs) and multivariate logistic regression were used to assess differences in reported fear levels between Aboriginal and non-Aboriginal cases. Although Aboriginal rates of violent victimization were higher, there were no appreciable differences in fear levels. In some situations, non-Aboriginal Canadians were even more likely to report fear of crime. This relationship held even in controlled analysis for urban-based Aboriginals and non-Aboriginals. Assault did not substantially increase the fear levels of either group. Low-income earners were more afraid of victimization than high-income earners. Although other factors mediated some of these effects, policymakers should focus on the large number of urban-based, impoverished Aboriginals who live with a heightened fear of crime. 3 tables and 30 references