American Homicide (review)

Given the regional, ethnic, social, and cultural diversity of the country and the sparseness of records prior to the twentieth century, many have concluded that a statistically based history of homicide in America was impossible. But, armed with “complete or near-complete records from scores of counties”(xi), Roth has written a history of homicide from colonial times to the present. He concludes that homicide rates among nonrelated adults correlate with confidence in the stability and legitimacy of government and “fellow feeling arising from racial, religious or political solidarity” (18).

The evidence and thesis both raise some serious questions. First, can “scores” of counties actually provide sufficient data to reach conclusions about four centuries of American history? Roth refers readers to a website for further information about his database and the formulas that he uses to compensate for missing data (and for an explanation to reviewers who “misunderstand” his work). However, rather than being a statistical problem amenable to mathematical manipulation, the deliberate decisions involved in determining whether and how homicides are reported are an essential aspect of the history of homicide.

The book includes voluminous citations but fails to address the fact that a careful examination of the data in local studies demonstrates that the variables associated with homicide resist any simple correlations. Roth acknowledges the myriad causes and explanations that have been suggested in previous works but rejects them all, pointing out that “the best predictor of increases and declines in America’s homicide rate has been the percentage of new counties named for national heroes—an indirect measure of how Americans felt about their nations and one another” (22). The lack of immediate connection between the two would seem to suggest that other factors might be more useful.

Roth defends his argument in a survey so sweeping that virtually every page contains enough overstatements, generalities, and outright inaccuracies to make specialists in the region or period under discussion cringe. All too often, the source cited for a particularly startling claim says nothing of the kind.

In addition to the problems of evidence and method, Roth treats homicide as simply a matter of rates regardless of context. For example, in discussing the colonial period, Roth notes, “Once the Natives were not numerous enough to defend themselves, Natives and Europeans stopped killing each other”(47). He also suggests that slavery led to a decline in homicide (70–73). Despite the indecorous implication that genocide can lead to a decline in homicide rates, Roth concludes with a banal, if naïve, prescription: “If [politicians] recognize the role that emotions and beliefs play in homicide, and the importance of legitimate government and national unity, they may be able to act in more constructive ways”(473). Regardless of motive, politicians acting in more constructive ways does sound promising.