JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.The unemployment/crime rate relationship (U-C) has been described recently as "inconsistent, " "insignificant, " and "weak. " Prior assessments of the U-C relationship have used no more than 18 U-C studies, and no more than 7 with 1970s data. In this paper, I review the findings of 63 U-C studies, 40 of which involve data from the 1970s when unemployment rose dramatically. My analysis shows the conditional nature of the U-C relationship. Property crimes, 1970s data, and sub-national levels of aggregation produce consistently positive and frequently significant U-C results. I discuss the implications of these results and argue that it is premature to abandon "this now well-plowed terrain" and suggest potentially fruitful paths for future studies of the U-C relationship.In 1969, civilian unemployment reached a fifteen year low of 3.5 percent. Then, three consecutive recessions pushed unemployment to progressively higher levels in 1971 (5.9 percent), 1975 (8.5 percent) and 1982 (9.7 percent).' By the spring of 1982, unemployment topped 11 percent in six states.2 Hardest hit were teenagers, young adults, minorities, and blue-collar workers. For example, unemployment during 1982 reached 14 percent for 20-24 year old adults, 18 percent for construction workers, 28 percent for auto workers, and 42 percent for black teenagers. By the spring of 1982, fears of economic depression were openly discussed in the national media (Newsweek, 1982; New York Times, 1982; Time, 1982).Even as unemployment climbed to levels unmatched in 40 years, the significance of this trend was being discounted in a variety of ways. For example, government economists periodically redefined the concept of "full employment" to include progressively higher levels of unemployment (Harrington, 1980). Some argued that the "new unemployment" was less harmful because it increasingly involved women, teenagers, and voluntary job leavers (Bulkeley, 1977; Feldstein, 1973;Guzzardi, 1976). Others argued that high unemployment was unavoidable and necessary to combat inflation (Friedman, 1979; Safire, 1982), or would soon be diminished by the growth of jobs in service industries (Business Week, 1981). Still others claimed that rising unemployment had little or no impact on rates of crime (Wilson andCook, 1985 Wilson andHerrnstein, 1985).In this paper, I address the claim that discounts or dismisses the relationship between rates of unemployment and rates of crime (U-C, hereafter). I show that those claims have come to constitute a "consensus of doubt" about the U-C relationship that presumes both empirical and theoretical grounding. I review the findings from 63 studies that report some

Florida law allows judges to withhold adjudication of guilt for individ-Traditional labeling theory explains the potential "escalating" consequences of a criminal or delinquent labeling experience in two ways (Lofland, 1969;Sherman et al., 1992). The first consequence involves a

Perceptual deterrence researchers have used simple cross-sectional correlations between prior behavior and current perceptions to study the effect of legal threats on social control. Such designs are inadequate because they: (1) confuse the causal ordering of perceptions and behavior, and (2) fail to take into account other inhibitory factors in an explicit causal model. In an analysis of panel data, the methodological simplicity of earlier studies is shown to have led researchers to reach erroneous conclusions. Our data suggest that past studies report an experiential effect, not a deterrent effect, and that the effect of perceived sanctions on criminal involvement is minimal once social definitional factors (moral commitment, informal sanctions) are controlled.

Evidence for a relationship between unemployment and imprisonment has been regarded as "elusive" and "conflicting. " Such conclusions have been based primarily on aggregate-level data. Individual-level data have provided only indirect evidence for this relationshb. This research considers prosecution, incarceration, and length of incarceration outcomes for I , 970 criminal defendants arrested in 1982. Multivariate logit and OLS estimates show a signifcant, strong, and independent impact of unemployment on pretrial and postsentencing incarceration. The interaction of race and unemployment shows that the greatest likelihood of incarceration is for unemployed black defendants, especially those who are young males or charged with violent and public order crimes. Theoretical implications for the control and punishment of "dangerous classes" are discussed.The links between labor market marginality and criminal punishment have received extensive and careful assessment over the past decade. The relationship formulated more than 50 years ago by Rusche and Kirchheimer (1939) has been advanced conceptually and empirically by a rapidly growing body of scholarship. That growth was highlighted by the recent (1989) special edition of Contemporary Crises marking the fiftieth anniversary of Rusche and Kirchheimer's seminal contribution. 1 Several authors have broadened the application of Rusche and Kirchheimer to include alternative measures of labor market marginality and state controls (e.g., Inverarity and Grattet, 1989;Laffargue and Godefroy, 1989). Still, a substantial core of recent effort has addressed the relationship between unemployment and imprisonment. Since 1976, more than 40 empirical studies have tested that specific relationship (DeLone, 1990), and that *This paper is a revised version of a paper presented at the American Society of Criminology meetings, November 1988. The authors are grateful to Gordon Waldo for his assistance at an earlier stage in this project and to Leroy Gould, Taiping Ho, Gary Kleck, and anonymous reviewers for their helpful Comments.Rusche and Kirchheimer were concerned to show (1) how different forms of punishment attended different systems of economic production and (2) how the severity of punishments vaned directly with the size of surplus populations. It is the latter issue that has commanded the bulk of recent attention. 1.CRIMINOLOGY VOLUME 29 NUMBER 4 1991 2. Parker and Horowitz (1986:769) assert that "the available evidence suggests a genuine lack of a relationship."We sometimes use the term significant in place of srutisricully signifcunt to simplify expression. We recognize the important differences between substantive and statistical significance and only the latter is implied here. 3. -104.70 d z u NOTE: r-ratio in parentheses.

This paper reports on the first longitudinal study to consider the relationship between perceptions of legal sanctions and self-reported criminality. The longitudinal design helps to address the problem of interpreting causal order that traditionally has troubled deterrence researchers using cross-sectional data. Self-reports of unlawful behavior (petty theft, marijuana use, payment using bad checks) over the past year are correlated both with perceptions of legal sanctions measured a year earlier (Time 1) and with perceptions of sanctions at the time of the self-reports (Time 2), for a random sample of 300 undergraduates at a large state university. The principal advantage of this method is that it affords a comparison of "deterrent" effects (perceived sanctions at Time 1 and subsequent reported behavior) with what are termed "experiential" effects (reported behavior and subsequent perceived sanctions at Time 2). The latter have been consistently treated as deterrent effects by prior researchers using cross-sectional data. For perceived certainty of arrest for generalized oth ers, the observed deterrent effect was weak and in every instance less than the experiential effect. However, perceived certainty of one's own arrest produced modest but consistent deterrent as well as experiential effects. A path analysis suggests that the deterrent effect of perceived certainty (personal or aggregate) is much weaker than once believed and the experiential effect is substantially stronger. The implications of these findings for an interactive model of de terrence are considered.