by Donald West
A rating of ‘troublesomeness’ obtained before the boys were ten years old was derived from the combined opinions of the boys’ primary school teachers and of their male classmates. The teachers’ questionnaire included items on work performance, difficult relations with others in the class, disciplinary problems and irregular attendance. Classmates were asked to name four other boys in class who best fitted such descriptions as: “gets into trouble most”; “most daring”; “most honest” etc. Both assessments were used to allocate boys between the worst behaved quarter and the rest. Whether assessed by teachers, peers, or a combination of the two, there was a significantly higher proportion of future juvenile delinquents among the worst quartile on childhood “troublesomeness” (44.6% of 92 troublesomes against 12.7% among the other 319 boys). From the standpoint of causal explanations, it is not very helpful to find that bad behaviour predicts further bad behaviour, except to note that childhood troublesomeness, highly predictive of delinquency, comprised a variety of misconduct other than specifically delinquent type activity. On the other hand, if one is concerned to identify a vulnerable group for purposes of preventative intervention, it is useful to know that simpler observations of classroom behaviours are as effective as more complex evaluations of family backgrounds.
Whilst the link between delinquency and family adversity was indisputable, it was less effective in predicting the fate of individuals than the Gluecks had claimed. Relatively few boys had multiple adversities with such a high probability of becoming delinquents as to justify targeting them with preventative intervention. Of the 63 boys, 15% of the total sample, who had been identified as especially vulnerable though multiple background adversities, a half acquired a conviction record. Even so, had they been picked out for intervention, the effort would have been arguably wasted in half the cases. Prediction exercises assess risk, not the actual destiny of an individual.
Early in the research, information given by the Education Authority about the thirteen secondary schools to which most of our boys were allocated revealed a substantial difference in the proportions of their pupils with delinquency convictions. Dividing the schools into groups of high and low delinquency rate, there were proportionately twice as many boys who had already been rated as troublesome attending the high delinquency schools than among boys who went to low delinquency schools. However, the proportion of troublesome boys who became delinquents was similar in both types of school. The large difference between the schools was a reflection of the nature of their intake. The influence of the differing secondary school environments on our boys’ pre-existing delinquent propensities was minimal, but none of the boys were attending private schools.
We tried to test whether settling down to married life would lessen the likelihood of further convictions. Young men marrying early were matched with others of the same age with similar conviction records before marriage, who remained single for the next two years. There was a slight difference, the married men having fewer convictions than the single men during the two-year follow up. The effect might have been greater but for the tendency of delinquents to marry partners who also had a conviction record and who might exert an opposite influence. In fact, men whose wives had a conviction record did have rather more reconvictions. Moreover, examining fathers of the sample who had acquired a criminal record before marriage, those whose bride also had a conviction record had significantly more re-convictions subsequently. This confirmed the impression that wives exert a negative or positive influence according to their own delinquency potential.
An advantage of the Cambridge study was the availability of measures of behaviour independent of official criminal records, which are necessarily limited to apprehended offenders. During the interviews at age eighteen to nineteen the boys were asked about offences for which they had been convicted. There was fair agreement between their actual and confessed convictions, providing some assurance as to the honesty of the information given at interviews. They were also presented with seven cards each bearing the description of an offence, such as “stealing from parked cars, vans, lorries etc.” They were asked to say whether and how often they had committed each of these offences in the past three years. A high score on delinquency potential, derived from these self-reports, proved significantly predictive of future recidivism. These data also gave an indication of the low probability of being convicted for any one offence: 87% of burglaries and 94% of cases of taking away vehicles that were self-reported had not led to a conviction.
An important finding was that questionable life-style features admitted to at age eighteen to nineteen were particularly frequent among young men with a conviction record. Eleven such characteristics were identified – self-reported aggression; unstable work record; spending leisure time ‘hanging about’; gambling excessively; involved with antisocial groups; admitting to drinking and driving; sexually unrestrained; heavy smoker; prohibited drug user; anti-establishment responses on an attitude questionnaire; wearing tattoos. High scores on most of these items were twice or three times more prevalent in the quarter of the sample who had by then been convicted than among the three quarters so far conviction-free. Of those with a conviction, 67.3%, against only 14.6% of the un-convicted, had as many as four of these characteristics. The findings show that delinquents tend to display a cluster of anti-social behaviours and attitudes that make them troublesome in many other ways than their recorded law breaking.
As applications for repeated research grants were successful, and survey work was continued until the sample was in late middle-age, a massive dossier of information on the progress of each individual accumulated. In particular, a clearer picture of the distribution and continuity of criminal conviction careers emerged. By their twenty-first birthday, of the total sample, 34% had acquired a conviction record; by age fifty this had increased to 42%. The figure may seem high, but it is not so very much greater than the national figure for males of their generation, namely 33%. It follows that the mere possession of a conviction record should not be regarded a cause for permanent distrust. However, being first convicted at an early age was a bad sign. Of the 35 boys first convicted at ages ten to thirteen, 32 (91%) became recidivists
A small minority, 7% of the sample, were chronic offenders, each with at least 10 convictions. This small group accounted for a half of all offences leading to conviction that were recorded against members of the sample. Because the delinquents, and especially the recidivists, so often had convicted fathers and/or siblings, the high concentration of offences among a small minority of problem families was very striking, 6% of the families accounted for a half of all the offences recorded against members of the sample, their parents and offspring.
As later conviction records up to middle age became available it was possible to show that unfavourable features identified in childhood, especially when severe, besides being correlated with youthful delinquency, were significantly predictive of continuing criminality. Virtually all the measures considered unfavourable that were derived from data relating to age ten or earlier (such as low income, large size family, convicted parent, poor parental supervision. low IQ, troublesomeness), were significantly more frequent among the young delinquents classed as ‘desisters’ (i.e.convicted before but not after the twenty-first birthday), and even more frequent among the 16% of the sample classed as ‘persisters’ (i.e. convicted both before and after the twenty-first birthday). Assessments of anti-social lifestyle derived from the eighteen-year-old interviews showed similar trends. Anti-sociality ratings for desisters were significantly worse than those for the un-convicted and the difference was even worse for the persisters.
Some aspects of our statistics were less gloomy. Of the totality of offenders, 29% were convicted once only. Convictions tailed off in later years. At the peak age range fourteen to twenty, an average of 37 members of the sample were convicted each year, compared to only 5 over the age range forty-four to fifty. Excluding those convicted only once, none of the men had any fu
rther conviction more than fifteen years later than their first conviction. Few very serious crimes were recorded against our sample. The commonest offences were burglary and thefts of or from motor vehicles; these were mostly committed at ages ten to twenty-five. Less common offences, such as frauds, assaults, drug offences, and sex offences, occurred at any age.
When interviewed at age forty-eight only one of the ‘desisters’ self-reported any offence, suggesting that this group had truly relinquished offending. The common forms of youthful criminality were shown to be to a large extent a time-limited phenomenon that had become much rarer with increasing age. Interviews at age forty-eight showed that, in addition to the tailing off of convictions, the anti-sociality features of earlier years were also diminishing. Prohibited drug use, drink problems and admissions of delinquent acts were much less often reported, whilst living conditions, employment and cohabitations were more often assessed as satisfactory. A measure of life success, taking account of such data, showed that 95% of both the un-convicted and the desisters were now living reasonably successful lives. Even the minority classed as persisters, included a majority of 65% rated successful. Of the handful of latecomers to crime, convicted for the first time after the twenty-first birthday, 84% achieved a rating of reasonable life success.
To conclude, although the well-known factors of family deprivation and early troublesomeness do indeed strongly predispose to later social problems and delinquency, this course of events is not inevitable and, save for the worst cases, a majority of those affected appear to achieve some social stability eventually.
Results from the Cambridge Study, and implications for social policy drawn from them, are generally supported by findings from other researches in the UK and US. From an extensive analysis of our data, Professor Farrington has concluded that poverty, poor parenting and criminality in the family, together with troublesome and impulsive behaviour and poor school performance, are the most important risk factors for later delinquency and criminality. These risk factors are linked also to the development of a multiplicity of social problems, such as unstable marriages, alcohol and drug abuse, poor employment prospects, welfare dependency and the genesis of delinquent offspring. Risk factors are readily observable before the age of ten, and when multiple and severe serve to identify a minority for whom intervention is worthwhile. Parental training schemes, pre-school intellectual enrichment programmes, and psychological techniques of behaviour modification for children have all been tried with apparent success.
Even small reductions in delinquency potential would have beneficial effects on social harmony far beyond the reduction in common crimes. Treatments that secure some lessening of the seriousness and frequency of an individual’s offending should not be dismissed as failure. Translating research findings, such as those of the Cambridge Study, into social intervention programmes is never simple. Schemes that tackle just one aspect of deprivation, be it educational retardation or poor parental discipline, can only have a very partial effect in view of the large number of other influences involved. The effectiveness of any scheme depends on the effort and resources invested, the ways of enlisting client co-operation and the determination of the practitioners operating the programme. Considering the enormous cost to families and community of the antisocial behaviour of children and young adults, even a modest positive outcome is worthwhile.
Unfortunately, politicians look for immediate results. Proving the cost-effectiveness of interventions is a difficult task. The time gap between intervention in childhood and the beneficial outcome in adolescence and adulthood mean that convincing evaluations take a long time. Obtaining comparable samples of children who have completed an intervention programme and others who have not been given the opportunity to participate is difficult to obtain for logistical and ethical reasons. The Cambridge Study avoided active intervention and observed what might be called the natural course of events. However, one of the potential advantages of long-term follow up cohort studies is that pilot interventions can be applied to sub-samples, leaving the rest of the sample, which is of the same generation and population source, as a valid comparison group.
A word of caution about intervention is in order. The operation of a criminal justice system is essential, but evidence from our study of the effects of its intervention on young offenders, however beneficial in some cases, tends to be followed by an amplification rather than a diminution of offending. Labelling individuals as delinquent personalities, or segregating them in classes regarded as ‘sin bins’, can bring about further alienation and prove counter-productive.
Having been responsible for starting the Cambridge Study I am naturally proud of its success and pleased that cohort studies identifying patterns in criminal careers are prominent in modern criminology. Such research cannot but be helpful in promoting rational schemes for moderating some of the causes of common crimes, but one must not expect too much. The links between early background circumstances and subsequent behaviour are important, but only as a contribution. Their effects are modified by other factors. The hazards of life events, good or bad, are also important. Securing a supportive partner, escaping from delinquent friends, or finding a rewarding job can prove turning points, whereas being caught and imprisoned can further alienate from the mainstream community.
The measures used in social surveys are inevitably imprecise. Measurements of family income depend upon the accuracy of informants. Used as an index of ‘poverty’ they do not match the severity of consequential deprivation, which is in part due to how parents manage the family budget. It would be naive to expect throwing money at dysfunctional families to solve their problems. In short, the measures available to social researchers are of limited accuracy and bear somewhat problematic connections with the theoretical ideas from which they originated. Criminality, defined by the number of convictions, depending as it does on police activity and prosecution policies, is an imperfect measure of the frequency, seriousness and persistence of crime. It has some built-in inefficiency as a prediction variable. Moreover, ‘crime’ covers differing forms of misconduct, dependent on current laws and enforcement policy. It cannot be but a distant reflection of some hypothetical behavioural tendency.
The correlations produced in surveys do not necessarily reveal causal pathways. For example, the correlation between fathers’ and sons’ criminality could be the result of habits of lawlessness being transmitted by bad example. This is unlikely to be the general rule because, in many instances, the father’s convictions were past history before the son was born and criminal fathers were much more often condemnatory than supportive of their sons’ delinquency. Quite possibly there could be a hereditary transmission of, for example, a shared physiological tendency to impulsive or aggressive reactions. So far, genetic research has not yielded observations usefully predictive of the generality of anti-social behaviour.
Criminological surveys can never achieve the accuracy or universal applicability expected of the physical sciences. They describe what happens at a particular time in a particular setting. Human behaviour changes with the ever-changing environment. Wars, plagues, famines, dictatorships and economic revolutions overturn customary crime patterns. Furthermore, general surveys, such as the Cambridge Study, are essentially concerned with the commoner crimes that account for the bulk of the business of the courts. Particular crimes, such as murder, serial sex killings, large scale financial frauds that victimise thousands, organised crime that controls drug and people-trafficking, call for specialised studies. Criminology tackles a plethora of human problems. One research approach, however valuable, cannot produce a universal panacea.
SEXOLOGICAL ADVENTURES
Ventures into Courts and Prisons
Academics are given some liberty to pursue activities beyond their strict employment obligations. This enabled me to accept, now and then, invitations to act as a psychiatric expert. Having published the book Homosexuality, which had three subsequent editions and a final version Homose
xuality Re-visited in 1977, followed by Sexual Crimes and Confrontations in 1987, most of the requests concerned sexual issues. One of the more interesting commissions was to provide expert testimony at a 1978 trial instigated by David Norris, at the time a lecturer at Trinity College, Dublin. He was pleading that, being a homosexual, his constitutional rights had been abused by the criminalisation of buggery and gross indecency adopted from British laws enacted before Irish Independence. He maintained that the state of the law, obliging him to hide his sexual orientation for fear of dismissal from his employment, had caused him great stress. This was a brave attempt by a public figure, later to become a senator, to try to overturn discriminatory laws.
The questions put to me at the High Court hearing about the effects of the 1967 decriminalisation in England did not seem very pertinent to the constitutional technicalities at issue. I was not in a position to quote any systematic research, but was able to say that there was no known evidence that decriminalisation had had an appreciable effect on the prevalence of male homosexuality, despite the fears of “opening the floodgates” that had been expressed in parliamentary debate. It was known from my writings that I was a supporter of decriminalisation, but I explained that I had tried to weigh the known evidence about the effects of homosexuality objectively. The judge intervened to say he was doing the same.
