Shared Beginnings, Divergent Lives: Delinquent Boys to Age 70
By John H. Laub and Robert J. Sampson
Harvard University Press, 2003. 331 pp. $49.95
This criminological monograph poses, as its point of departure, a question criminologists ought to ask but usually do not: what happens to delinquents over the long haul? Do some of them eventually desist from crime and others continue until they die in old age? The reason this question is rarely asked is that, to address it, requires a longitudinal research design that granting agencies will rarely fund and expenditures of years of effort on the part of the researchers themselves.
In addition to posing this important theoretical question, Laub and Sampson attempted to answer it in Shared Beginnings, Divergent Lives, an imaginative longitudinal study. They located and capitalized on a classic data set originally collected by Sheldon and Eleanor Glueck to study the causes of delinquency. The Gluecks were interested in why the 500 delinquent males they found in Massachusetts reformatories had become delinquent when 500 youngsters carefully matched by age, socioeconomic status, ethnicity, and IQ did not. The Glueck research team interviewed members of the sample not only at an average age of 14 but also at 25 and 32. During their lifetimes the Gluecks did not permit other researchers access to their data files. After they died, however, the data passed first into the files of the Harvard Law School Library and then into the files of the Murray Research Center at Harvard. Laub and Sampson realized that, if they could get hold of these data and reinterview the surviving research subjects—born between 1924 and 1932, and now men of about 70—they might be able to throw light on why some of these men stopped being criminals while still fairly young and others continued criminal behavior into old age.
Laub and Sampson not only tracked down and interviewed 52 of these men, 40 face-to-face and 12 by telephone, but they also mined the Glueck files, which provided voluminous data on the research subjects until the age of 32: data on arrests and incarcerations, military service, marriages and parenthood, drinking and drug use, and work histories. Thus, they had records to check against the information yielded by the geriatric interviews. They reached several conclusions: a) All criminals eventually desist, some early and some much later; b) Desistance does not come suddenly, as an epiphany or a resolution to change. It results from a gradual process in the course of which the man makes commitments that crowd out criminal behavior patterns; c) The major commitments, or “turning points,” that have this capacity to replace criminal patterns of behavior are a successful marriage, a satisfying job, and military service; d) Laub and Sampson specifically reject the notion that personality characteristics or experiences in childhood or adolescence imprint criminality in some offenders. They infer from their qualitative and quantitative data that these turning points cannot be predicted in advance because meeting the right woman, finding the right job, or fitting into a satisfying military assignment may depend on luck and personal agency. Thus, Laub and Sampson avoid a deterministic approach to causation; they emphasize the choices that their research subjects actively made.
Laub and Sampson reached these theoretical conclusions in the course of carrying out a challenging research project. Readers may find their theory convincing, as I do, and applaud a major intellectual contribution to thinking about crime. However, unconvinced skeptics concerned with their analysis of the process of desisting from criminal behavior can point to the following methodological weaknesses in the empirical study:
1) The sample of serious delinquents. As Laub and Sampson recognize, they did not have an opportunity to select the sample of serious delinquents. All 500 of the Glueck delinquents were white males from the Boston area, born between 1924 and 1932, who had been incarcerated in two Massachusetts reformatories. Laub and Sampson have to assume that this non-random sample of serious delinquents desisted from crime for essentially the same reasons that some quite different sample would have. One could argue that female or rural or Hispanic offenders would follow a different trajectory of desistance.
2) Response bias. Laub and Sampson interviewed only 52 out of the 500 in the original sample. Some had died; some had disappeared; 53 had unlisted phone numbers and did not respond to mailed requests to participate; 9 were too ill to participate; 27 were talked to on the telephone but despite an offer of $25 for their time refused to be interviewed. Laub and Sampson had to hope that the unknown differences between those they interviewed and those they could not interview were not relevant to the desistance process.
3) Causal order. Laub and Sampson demonstrated a convincing correlation between desistance from crime and three “turning points”: a successful marriage, a satisfying job, and military service. What they could not demonstrate is whether some antecedent characteristic or circumstances in the lives of these men predisposed them to choose to marry and to work at making their marriages succeed, to take a legitimate job, or to join the Armed Services. A skeptic might claim that marriage, jobs, and military service are intervening variables that may reinforce a desire on the part of their respondents to desist from crime but do not have causal efficacy.
Laub and Sampson deserve enormous credit for revitalizing an old idea: maturing out of crime. Whereas other scholars seemed to regard aging by itself as the engine driving desistance, Laub and Sampson point to three plausible turning points in the lives of serious offenders. Offenders who embrace marital and parental roles, work roles, or even military roles gravitate toward a law-abiding life style. Such roles give them a stake in conformity. But it is a pity that Laub and Sampson did not point to other longitudinal studies that might support their theory by overcoming some of the methodological weaknesses of their research. For instance, they do not refer to Project Metropolitan, a cohort of 7,719 boys and 7,398 girls born in 1953 and registered as living in Stockholm on November 1, 1963. True, Project Metropolitan collected crime data on these persons only until 1984 (when cohort members were 31). But partly due to the comprehensiveness of Swedish record keeping and partly to the extraordinary diligence of the director of Project Metropolitan, Professor Carl-Gunnar Janson (2000), a desistance study of such a cohort could confirm (or disprove) some of the details of the process that Laub and Sampson hypothesize. Perhaps they wish to pass the torch to others.
Janson, Carl-Gunnar, ed. 2000. Seven Swedish Longitudinal Studies. Stockholm:
Swedish Council for Planning and Coordinating of Research.
Jackson Toby is a visiting fellow at AEI.