Forensic Psychology

by Graham M Davies

  Perhaps more surprisingly, even when the police are not available to arrest wrong-doers – because they are on strike – there is no convincing evidence of upward surges in crime. There was a “rash of crimes” following a one-day police strike in Montreal in 1969 (Clark, 1977), but apart from this instance other studies of industrial action by police do not support what has been called the “thin blue line” hypothesis. There were public riots during a longer strike in Boston in 1919, but analysis suggests these events were about to happen in any case, and would have done whether or not the police had gone on strike (White, 1988). Pfuhl (1983) analysed crime-rate data for police strikes in 11 US cities during the 1970s. Strikes ranged in duration from three to 30 days. While there were increases in rates of some crimes in some locations, in the majority there were no changes at all. Similarly, following a police strike in Finland in 1976 no upsurge in crime or other social upheaval occurred (Makinen, & Takala, 1980). Thus, our predilections notwithstanding, there is no overall pattern supportive of police presence as a major deterrent against crime.

  A third type of evidence that calls into question the assumption of a general deterrent effect of sentencing emerges from studies of the death penalty. The usage of this most serious of sanctions is a highly emotive issue regardless of its actual outcomes. For persons who oppose it and who press for it to be abolished, “any discussion of its effectiveness as a deterrent is irrelevant” (Hood & Hoyle, 2015, p. 8). But the rationale for applying capital punishment is very substantially based on its expected deterrent effects. As of April 2014, there were 39 countries where the death penalty remained in force and where it had been used at some point in the preceding 10 years. According to Hood and Hoyle (2015) who conducted a global survey of its use on behalf of the United Nations, the case for retaining it “usually rests not only on retributive sentiments but also on assumptions about its unique deterrent effects as compared with alternative lesser punishments” (2015, p. 8). However, international comparisons have failed to find that the availability of the death penalty as a sentencing option has any clear suppressant effect on rates of the most serious crimes such as homicide. Hood and Hoyle (2015) allocated different countries (or their constituent member states) to discrete categories according to their pattern of usage of capital punishment over a more than 50-year period. Some countries or states retained the use of the sentence throughout that epoch. Others had either not used it throughout the same period, or at some point had formally abolished its use. Of specific interest, the United States had an interlude when execution was not used (1967–1976, as a result of a moratorium and subsequent rulings by the Supreme Court), following which its use was resumed in 37 of the 50 states.

  Examination of patterns of homicide and other capital crimes under these different jurisdictions yields no evidence that capital punishment is associated with reductions in their rate of occurrence. Even in Texas, “by far the most active death penalty state” (Sorensen, Wrinkle, Brewer, & Marquart, 1999, p. 483), an analysis of crime statistics for the period 1984–1997 found no relationship between the execution rate and the murder rate. While there are econometric studies where apparent deterrent effects have been reported (Yang & Lester, 2008), this emerges using just one type of methodology and such studies cannot demonstrate a causal link between executions and homicide rates. In addition, the model of rational decision-making (derived from price theory in economics) that underpins such analyses bears little relation to what is known about the causation of crime.

  Possibly the most thorough analysis of data comparing American states, across the period 1977–2006 (Kovandzic, Vieraitis, & Boots, 2009) found no evidence of a deterrent effect. In a present-day “tale of two cities”, Zimring, Fagan, and Johnston (2010) undertook a direct comparison for the period 1973–2007 between two closely comparable urban areas, Singapore and Hong Kong, with completely opposite policies on capital punishment. The former’s execution rate rose steeply from 1991–1992 onwards and at one point was the highest in the world, though thereafter it steadily declined. From 1966 onwards Hong Kong had no executions and the death penalty was abolished in 1993. Despite this very marked contrast, Zimring and his colleagues found the two cities had remarkably similar rates of homicide throughout the entire period studied.

  There are other sound reasons for suggesting that the presence of criminal sanctions is not necessarily the fundamental force that holds back a tide of lawbreaking on a daily basis. Firstly from an “instrumental” standpoint, the majority of individuals obtained reward and satisfaction from their everyday involvement in routine, socially accepted activities in the domains of family, work and leisure. These entail deep and significant personal investment for virtually everyone, and whilst the patterning of this may be uneven and may sometimes present individuals with formidable challenges, in behavioural terms such activities are the source of a steady flow of positive reinforcement. Arguably, it is when such reinforcement schedules are substantially weakened (for example through inconsistent parenting, family breakdown, school failure, unemployment, or specific aversive experiences) that individuals are susceptible to following other routes towards achieving goals. Several major criminological theories (e.g. strain, differential association, and social learning theories) are founded on these types of observations (McGuire, 2004).

  Secondly, there are clear “normative” factors influencing individuals’ self-regulation of their own conduct in society. Tyler (2006) has adduced strong evidence to suggest that the everyday law-abiding behaviour of most people is a function of their endorsement of the law and perception of its operation as legitimate, rather than their fear of detection and punishment. Tyler conducted a series of structured interviews with a random sample of 1,575 respondents in Chicago, 804 of whom were re-interviewed one year later. Participants were asked about their experience and perceptions of the law, the extent of any contacts with the police or courts, and various factors potentially linked to their behaviour. The key influence underpinning most individuals’ adherence to lawful behaviour was their commitment to doing so, arising partly from personal morality and partly from beliefs in the appropriateness of laws. Thus where the law is perceived to have legitimacy, that is, where its rationale is considered sound and its operation perceived as fair, most citizens regularly and willingly act in accordance with it. Where individuals disagree with laws or experience their enactment as unjust, their compliance is weakened; some laws may then be difficult, even impossible to enforce, despite the availability of punitive sanctions.

  With reference to other penal philosophies it is possible to collect data that can subject them to outcome tests. This applies, for example, to incapacitation. Criminologists have constructed elaborate models for the evaluation of its crime-preventive effects. Doing so entails a type of modelling process, for example projecting from data on the numbers of dwellings broken into by the average burglar in a year, and estimating the number of crimes that would be prevented by sentencing larger numbers of house burglars to imprisonment. While the logic of this appears straightforward, the effects of incapacitation are much weaker than we might have supposed, and it proves a proportionately very expensive method of reducing crime. For example, applying such a model to England and Wales, Tarling (1993) concluded that even if the prison population were to be increased by 25%, the net effect on the rate of crime would be a drop of just 1%. Similar conclusions were reached following time-series analysis of crime trends and imprisonment rates in California (Zimring & Hawkins, 1994).

  17.4.1 The Effects of Criminal Sanctions

  “Crime and punishment” are inextricably linked in our culture and within everyday discourse and public expectations. Notwithstanding the range of penal philosophies reviewed above, the presumed individual deterrent function of the sentence is arguably pivotal in perceptions of the process as a whole. Although sentencing policy documents do not often give it priority, “there is a widespread assumption that the resulting punishment will deter the convict
ed offender from re-offending, or potential offenders from offending” (Easton & Piper, 2012, p. 106, italics in original). Given the central place allotted to punitive sanctions in the legal systems of the world, does the available evidence justify these ubiquitously held expectations that individuals will be deterred from future offending by the experience of official punishment? Several types of evidence indicate that these expectations are very badly misplaced.

  If punitive sanctions reliably achieved individual or specific deterrence, it would be not unreasonable to expect the following. We might expect that there would be a relationship between the experience of punishment and what individuals say in relation to their behaviour as a result. We might expect there to be an association between the severity of sentences and recidivism outcomes, when other variables are held equal, with heavier penalties leading to larger effects. We might anticipate that in specially designed studies where some offenders are dealt with more severely, they would go on to commit fewer crimes. In a pattern that might at first appear counter-intuitive, none of these expectations is supported by the evidence that has been collected to date.

  The first type of evidence is derived from what individuals say has had an impact on them and has influenced them to change their behaviour (where that occurred) after being arrested for offences such as theft or substance abuse. Klemke (1982) interviewed young people some time after their first arrest to explore levels of subsequent involvement in offending. Individuals who had desisted from offending were very unlikely to attribute this to any deterrent effect of their contacts with the criminal justice system. Of course, some researchers would regard self-report data of this kind as a rather weak form of evidence, given variables such as impression management, fear of arrest, and other factors potentially at work.

  A second type of evidence is likely to be regarded as more robust, being based on officially recorded statistics: patterns of reconviction following different types of sentences, usually involving very large samples. These data are of particular interest as they allow comparisons to be made between actual offence rates and those that would be expected, based on predictions made by using a specially designed scale to assess risk of reoffending, derived from information on individuals’ previous criminal histories.

  The Offender Group Reconviction Scale (now in Version 3, OGRS-3) was developed for this purpose. But when Lloyd, Mair, and Hough (1994) used it to compare the impact of different types of sentence (e.g. imprisonment versus community penalties), they found that rates of recidivism two years later were virtually identical. Patterns of subsequent criminality appeared wholly unaffected by the type of punishment imposed. There was no evidence that the more severe sanction had any suppressant or punitive effect. Similarly, in a large-scale review for the Solicitor General of Canada, integrating data from 23 studies, and with an aggregate sample of 68,248, Gendreau, Goggin, and Cullen (1999) found no clear relationship between the lengths of prison sentences and rates of subsequent recidivism, once other differences between samples were taken into account.

  But perhaps the strongest kind of test of the deterrence hypothesis comes from specially designed experiments in which the intervention consists of “enhanced” (i.e. harsher) punishments whilst the comparison sample receives “business as usual” in the criminal justice system. There have been numerous studies of this type, and they more or less uniformly show that the net effects of dealing with offenders in this way are either absent (the effect size is zero) or actually negative; that is, people “got worse” and committed more crimes (Gendreau, Goggin, Cullen, & Paparozzi, 2001; and see McGuire, 2004). A particularly stark illustration of this comes from evaluation of the “three strikes and you’re out” policy implemented in some American states from 1994 onwards. In California, subsequent analyses found effects that were the reverse of what had been predicted. Those counties that used the law the least experienced larger declines in violent crimes than those counties that used it most (Males, 2011).

  If we accept this evidence, and conclude that deterrent effects are very weak, simply non-existent or actually harmful, we may be left with what looks like a paradox. Most people probably assume that the unpleasantness of being punished for doing something is likely to decrease our chances of doing it. Behavioural psychology has demonstrated the power of reinforcement and punishment in shaping patterns of human behaviour. But research in that field has also shown that for punishment to work efficiently, certain conditions have to be met. For example, in order to be effective, it should be virtually certain to happen, and should follow swiftly on the problem behaviour to be reduced. These circumstances are rarely if ever met in the criminal justice system, and other evidence suggests that individuals on the verge of committing crimes are unlikely to be thinking in any detailed way about the possible negative consequences of their actions (McGuire, 2004).

  17.5 REDUCING OFFENDING BEHAVIOUR

  Fortunately there are other approaches to addressing the problem of offender recidivism at the individual level that have emerged as more effective than official sanctions. This field of activity is sometimes denoted as tertiary prevention (Gendreau & Andrews, 1990) as it focuses on adjudicated offenders, those who have been convicted of crimes and duly sentenced.3 More colloquially, drawing on the title of a journal paper by Martinson (1974), which famously drew negative conclusions regarding early evidence on offender rehabilitation, it is often referred to simply as “what works”.

  The discussion of this within social science is often very hard to separate from moral, political and philosophical dimensions of the same issue. For many people, their opinions concerning how society should deal with offenders are influenced by their broader beliefs on other social questions. In journalistic terms this is sometimes called the “law and order” debate, and it can be very difficult to distinguish the relevant empirical evidence from other aspects of it, and decide how that evidence could be used realistically to inform the administration of justice. There has been continuing controversy over whether it is possible to change the behaviour of those who have been persistently involved in crime.

  However, the systematic study of criminal sentencing and of the treatment of offenders took a sizeable stride forward some years ago when several review studies were published based on the application of statistical integration or meta-analysis of research findings. Since its initial use in the field of criminal justice in 1985, this approach has been employed in a number of reviews of treatment-outcome studies with offenders (Wilson, 2001). The pace of development was such that by mid-2012, a total of more than 100 meta-analyses of tertiary-level offender treatment had been published (McGuire, 2013); several more have appeared since then. While much research remains to be done, the cumulative findings of this work provide some firm indications concerning which approaches to working with offenders are most likely to yield intended outcomes in terms of lowering criminal recidivism.

  17.5.1 The Evidence Base

  The majority of the reviews conducted in this field, and the bulk of the primary research on which they are based, originate from North America, although data from many countries are encompassed within them. Several have been conducted in Europe, and these have been reviewed separately, as a validation or generalisability test of the American results (Redondo, Sánchez-Meca, & Garrido, 2002). Most are published in the English language, but in the largest review so far conducted (Pearson, Lipton, & Cleland, 1997), contacts were made with 14 non-English speaking countries and more than 300 reports were obtained in languages other than English.

  Not surprisingly, given that most crimes are committed by men, the overwhelming majority of the primary studies deals with male offenders. In one of the largest meta-analyses, carried out by Lipsey (1992, 1995), only 3% of published studies focused exclusively on samples of female offenders. With regard to age, roughly two-thirds of the reviews focus on interventions with adolescent or young adult offenders in the age range 14 to 21 years. This covers the peak age for delinquency in
most countries. The remaining studies are either concerned exclusively with adults, or include offenders across a wide range of ages. Concerning ethnicity, while many studies provide data on the proportions of offenders from different ethnic groups, the pattern of this is variable and it is not consistently recorded. However, given the over-representation of minority communities under criminal justice jurisdiction in many countries, these findings are based on populations containing a broad mixture in terms of ethnicity.

  Research in this area involves experts from many disciplines. Some reviewers have coded the professional backgrounds of those who undertook the primary studies. For example, Lipsey and Wilson (1998) reviewed 200 reports on the outcomes of work with serious and violent young offenders. Amongst these studies, they found that 29% of the senior authors were from psychology and 19% from criminology. For a further 28% of the studies, no main discipline was identifiable. Other backgrounds represented to varying degrees included sociology, education, psychiatry, political science and social work. But clearly, psychologists have made a notable contribution to this work.

  17.5.2 Findings of the Reviews

  The first meta-analysis published in this field was an evaluation of the impact of institutionally-based interventions for juvenile offenders. Garrett (1985) surveyed 111 studies conducted in the period 1960–1984, describing residential treatment programmes. Altogether, 13,055 individuals with a mean age of 15.8 years were participants in these studies. Since then, several meta-analyses have been conducted covering a wide range of interventions with young offenders, including community as well as residential settings. Other reviews have focused on specific types of offence (driving while intoxicated, violence, sexual assault, substance abuse). Several have focused on different types of punitive sanctions, such as intermediate punishment and “scared straight” interventions, correctional boot camps, and outdoor-pursuit, wilderness-challenge schemes.