A review of the literature suggests three likely sources of discrepant results across studies: (1) differing operationalizations of racial inequality, (2) differing samples, and (3) differing specifications of the regression models. Analyses of data for 199 Standard Metropolitan Statistical Areas indicate that a variety of metrics (differences, ratios, log differences, and log ratios) yield quite similar results. Sample size also does not appear to be a major issue. However, operationalization of racial inequality in terms of socioeconomic status (SES) rather than income can lead to very different conclusions. For certain models, SES-based measures are clearly superior to those based on income. Further, slight modifications of the regression models can render the effects of racial inequality in SES nonsignificant. These results call for a skeptical assessment of previous evidence indicating a positive relationship between racial inequality and rates of violent crime. 4 tables, 9 footnotes, and 25 references. (Author abstract modified)