Imprisonment is the Criminal Justice System’s primary response to offending (Beijersbergen, Dirkzwager & Nieuwbeerta, 2016), but with many UK, European and American prisons nearing full capacity, considerable debate surrounds the effectiveness of prisons (Mears, Cochran, Siennick & Bales, 2012). Imprisonment is extremely expensive (Nagin, Cullen & Jonson, 2009) and as approximately 45% of inmates are reconvicted within 1-year post-release (McCarthy & Brunton-Smith, 2018), it seems critical to examine current prison models to evaluate whether alternatives may be more efficacious (Walmsley, 2016). The utilitarian approach justifies punishment through imprisonment by the intended positive impact prison could have on future criminality (Gideon & Thomas, 2018).
The focus is making society safer (Rodrigues, Sacau, Oliveira & Gonalves, 2018), by preventing crime (Mears & Barnes, 2010) through deterrence, incapacitation and rehabilitation (Kleck, Server, Li & Gertz, 2005). Considering this in relation to psychological research, this paper explores whether it is right to assume prisons ‘work’. Reoffending rates, arguably the most important measure of success (Spivak & Sharp, 2008), inform the conclusion that prisons fail to deter (Andrews & Bonta, 2010), have minimal incapacitation effects (Sweeten & Apel, 2007), fail to facilitate rehabilitation (Schafer, 2018) and therefore, do not work.
A key rationale underlying prison sentencing is deterrence theory (Cullen, Jonson & Nagin, 2011). Based on rational choice theory, it suggests that by locking criminals away, crime will be reduced through teaching offenders (specific deterrence) and society (general deterrence) that ‘crime does not pay’ (Sapouna, Bisset & Colong, 2011), as imprisonment is a costly penalty (Lipsey & Cullen, 2007). Deterrence theory predicts fear of imprisonment will reduce offending (Petrosino, Turpin-Petrosino, Hollis-Peel & Lavenberg, 2013) by out-weighing potential benefits (Becker, 1968).
From an empirical perspective, specific deterrence is questionable due to high recidivism rates; often reaching 60% within 3-years (Cochran, Mears, Bales & Stewart, 2014). Likewise, general deterrence lacks credible support (Paternoster, 2010).
Specific deterrence implies imprisonment incurs a variable price, most costly when sentences are longer, and conditions harsher (Listwan, Sullivan, Agnew, Cullen & Cilvin, 2013). Initially, this seems credible given evidence that prisoners serving longer sentences (more than 1-year) are less likely to re-offend (Ministry of Justice [MoJ], 2016). However, this is correlational and McCarthy and Brunton-Smith (2018) argue most likely explained by natural resistance associated with ageing. Chen and Shapiro (2007) further challenge specific deterrence by demonstrating ‘harsher’ prison conditions (maximum- versus minimum-security) equate to an increased likelihood of re-arrest post-release. Gaes and Camp (2009) replicated this, randomly assigning prisoners to the maximum- or minimum-security prisons. Officials were unaware of prisoner status to avoid bias and after controlling extraneous variables, researchers quantified harsher prison assignment increasing/increases reoffending by 31.3%.
Specific deterrence advocates prison will reduce future criminality more than non-custodial sentences as it is costlier (Cid, 2009). However, numerous matching-based studies, considering all types of crime and offenders, continually demonstrate community-based interventions reduce reoffending more than prison (Cochran, Mears & Bales, 2014; Marsh & Fox, 2008). Jonson’s (2010) meta-analysis of 57 studies found the reoffending risk for imprisoned offenders increased 14% compared to those with community sanctions, and even when considering only highest-quality studies, this was still significant at 5%. Similarly, Smith, Goggin and Gendreau (2002) found an 11% increase. Despite lack of research into gender differences, and while women are more likely to receive non-custodial sanctions (Sapouna et al., 2011), Mears, Cochran and Bales (2012) found prison’s criminogenic effect present for both genders.
Surveys of offenders reveal they are more likely to dread intensive and lengthy community-based punishments than shorter prison terms (see Moore, May & Wood 2008; Petersilia & Deschenes, 1994). Additionally, Nagin et al. (2009) reported some offenders claim prison is more pleasant than anticipated. This potentially reduces the perceived loss of imprisonment, subsequently increasing the likelihood of reoffending. This is supported by Raaijmakers, Loughran, Keijser, Nieuwbeerta and Dirkzwager (2017), who found offenders viewed prison as ‘relatively trivial’ after previous imprisonments. Experimental research by Deschenes, Turner and Petersilia (1995) demonstrated when randomly assigned intensive probation supervision, some individuals opted instead to serve their sentence in prison. Though dated, this nevertheless contradicts specific deterrence; non-custodial sanctions are perceived as more onerous than prison (Wood & May, 2003).
Reoffending is difficult to measure as it is affected by many different factors which should be controlled, such as; socioeconomic status, post-release supervision, imprisonment length, crime severity (Esperian, 2010). Additionally, studies may be based on varying data; criminal prosecution, new convictions, or re-imprisonment (Israel & Chui, 2006), making comparisons problematic. Due to these limitations, recidivism figures should be interpreted with appropriate critique (Tomek & Rozum, 2018). Based on comparative analysis, Bales and Piquero (2012) concluded increased recidivism risk for custodial compared to community sentences is present regardless of research methodology. Furthermore, Spohn and Holleran (2002) and Spohn (2007) found time to reoffending was shorter for ex-prisoners, even after controlling background characteristics and prior criminal record. Consequently, findings appear robust and dispute specific deterrence theory (Gaes & Camps, 2009). Recidivism measures are further limited by failure to encompass hidden crime, meaning they are underestimates and prisons are even less effective than research suggests (Tomek & Rozum, 2018).
The notion of ‘resetting bias’ has been proposed, opposing deterrence explanations (Pogarsky & Piquero, 2003). Also known as gambler’s fallacy, this infers offenders believe risk of apprehension is lowered following imprisonment (Pogarsky, 2007), encouraging reoffending (Sitren & Applegate, 2006). This potentially explains high recidivism rates (Cid, 2009). Overall, comparison of custodial versus non-custodial sentences has persisted as means of assessing deterrence (Villettaz, Gilli?ron & Killias, 2015). Research across various methodologies consistently indicates criminogenic effects of imprisonment (Caudy, Tillyer & Tillyer, 2018). Based on best available evidence, for all offenders, prisons do not have specific deterrent effects (Schafer, 2018), undermining utilitarian beliefs, this implies prisons do not work as intended.
Prison physically isolates criminals from communities, preventing risk of further offending during imprisonment (incapacitation) (Bhati & Piquero, 2008). Yet some prisoners continue to offend in prison (Drury & DeLisi, 2011). Worral and Morris (2011) found custody level significantly and positively relates to misconduct, which in turn predicts earlier reconvictions (Huebener, Varano & Brynum, 2007). This may justify claims that prisons are ‘schools for crime’ – crime may be due to association with other criminals (Jones, 2014). Prison fosters cultural values supporting crime, transmitted through daily interactions (Butler & Maruna, 2016). Consequently, prisons are social learning environments in which criminal orientations are reinforced and illegal involvement deepened (Cullent et al., 2011). Consistent with social learning theory, prison sentences are likely to intensify commitment to criminality (Akers, 2017). Evidently, prisons do not have true incapacitation effects as crime persists in prison (Wood, Williams & James, 2010). This further weakens utilitarian arguments, if prisons do not reduce crime during imprisonment, the assumption they work via incapacitation is unfounded.
Labelling theory offers explanation for prisons apparent criminogenic effect (Cullen & Cullen, 1978), arguing prisons have unanticipated effects of increasing criminal propensity through strain, association and stigmatisation (Hagan, 1973). When admitted to prison, a label is assigned, signifying attribution of role-appropriate behaviour (Pakes & Winstone, 2011). Becker (1963) described how labels work well for those maintaining social rules but adversely for those breaking them. Prisoners accepting and internalising their institutionally given ‘deviant’ label, act consistently with the self-conception, increasing reoffending (Kituse & Lemert, 1968). Chiricos, Barrick, Bales and Bontrager (2007) examined a Florida law allowing judges sentencing offenders to probation to withhold formal guilty verdict, with arrest record disappearing on successful completion of probation. Essentially enabling a natural experiment whereby some offenders were labelled, and others were not. Based on a large sample of 95,919 men and women over 2-years, researchers discovered formally labelled offenders were considerably more likely to recidivate. Unfortunately, this study did not control for potentially confounding variables. It also seems likely that offenders who failed to complete probation (therefore receiving official labels), would be more likely to recidivate regardless of labelling, so results are potentially biased.
Contrasting deterrence theory, labelling theory claims prison sentences generate more recidivism than non-custodial sanctions (Cid, 2009). Cid (2009) conducted an 8-year follow-up study comparing recidivism rates between former prisoners and offenders who served suspended prison sentences. Utilising a representative sample of 483 offenders and controlling risk factors predictive of reoffending, logistic regression techniques examined predictive power of sanction on reconviction rates. Analyses revealed offenders given suspended sentences had lower reconviction risk than those with full prison sentences. This provides empirical support for labelling theory, indicating alternatives to prison are more effective at reducing recidivism. Having taken place in Barcelona, cultural differences limit generalisability. However, this provides beneficial longitudinal data, more insightful than short-term estimates; common but crude measures of crime (Frost & Clear, 2012). If labels associated with prison sentences increase reoffending, support cannot be given in favour of prisons.
Drawing on general strain theory (Agnew, 1992, 2001, 2006), Listwan et al. (2013) suggested certain ‘strains’ imposed by prisons on those labelled ‘prisoners’, increases likelihood of recidivism. Strains are physically or psychologically distressing conditions (Raphael, 2011), which may induce negative emotional states (anger or frustration) (Burton, Lux, Cullen & Miller, 2018). Crime may become a coping strategy, allowing individuals to escape from or reduce strain, obtain revenge against the source or alleviate negative emotions (Blevins, Listwan, Jonson & Cullen, 2010). Listwan et al. (2013) implied out of 1,613 recently released inmates from Ohio, those who experienced higher levels of strain in prison were more likely to recidivate. Though, results were not significant and must be interpreted cautiously.
An alternative coping mechanism to avoid strain in prison may be to join gangs for protection (Frost & Clear, 2012). Becker (1963) focused on how deviant labels can further embed individuals into deviant social groups, representing sources of support in which crime is accepted. Gangs are a pervasive problem in prisons (Gaes, Wallace, Gilman, Klein-Saffran & Suppa, 2002), endorsing violence (Listwan et al., 2013) and facilitating crime involvement (Griffin & Hepburn, 2006). Furthermore, gang members’ prospects of reoffending are significantly worse than general inmate population (Pyrooz, Decker & Fleisher, 2011), evident from Dooley, Seals and Skarbek’s (2014) experiment in which prison gang membership resulted in 6% increase in recidivism. Prison gang membership signifies a commitment to criminal lifestyle (Pyrooz et al., 2011). Association with inmates with longer criminal histories and more serious offences may encourage recidivism (Bayer, Hjalmarsson, & Pozen, 2009) by expanding knowledge of criminal conduct, inspiring more sophisticated crimes post-release (Duwe & Clark, 2014). If strains inflicted by prison increase reoffending and encourage gang membership, which further promotes criminality, then prisons cannot be deemed successful for crime prevention.
Nagin et al. (2009) describe labelling as criminogenic because stigmatisation instils strained identity. Prisons are particularly consequential, providing lengthy opportunities for acceptance and reinforcement of criminal identity. Re-entry into society following imprisonment is also criminogenic because the societal reaction is pivotal to denying opportunities (McCarthy & Brunton-Smith, 2018). Ex-prisoners face housing and job discrimination, challenges re-establishing relationships, key factors in reducing reoffending (Pakes & Winstone, 2011). Bernburg, Krohn and Rivera (2006) used Rochester Youth Development Study data to demonstrate formal criminal labelling, leading to greater stigmatisation, increases short- and long-term criminal conduct. This study reveals how the societal reaction to criminal labels can trigger a criminal career (Cullen et al., 2011). Stigmatisation is greater for ex-inmates compared to those serving non-custodial sanctions (McCarthy & Brunton-Smith, 2018). Prison has become a tool for social exclusion (Pakes & Winstone, 2011) and thus it cannot be assumed prisons work, as stigmatisation results in failure to reintegrate ex-prisoners back into society, ultimately making society less safe.
Deterrence theorists conceptualise prisons as a calculated price when deciding to offend (Cid, 2009), but for those portraying prison as criminogenic, such thinking is reductionist (Nagin et al., 2009). According to labelling theory, prison unintentionally increases exposure to crime-inducing influences (Burton et al., 2018), resulting in conversion to ‘hardened’ criminals (Tella & Schargrodsky, 2013). Research documents often-cruel treatment in prison (Haney, 2012), communicating prisoners’ lack of worth and value, serving to degrade and reinforce stigma (McCarthy & Brunton-Smith, 2018). Although intended to prevent crime, evidence-based research by Cullen et al. (2011) proposes sound arguments for prisons criminogenic effect on offenders. The systematic review concluded that imprisonment is a ‘crude strategy’ neglecting underlying causes of crime. Labelling theory, a perspective gaining increasing support (Cullen et al., 2011) clearly illustrates why it is wrong to assume prisons work.
Placing emphasis on rehabilitation offers a better option for dealing with crime (Harding, 2014). Rehabilitation is the focus in Scandinavian countries which typically have much lower imprisonment rates, for example; 71 compared to 148 prisoners per 100,000 population in Norway and UK respectively (Walmsley, 2016). Fazel and Wolf’s (2015) worldwide systematic review found Norway’s 2-year recidivism rates were approximately 20% which compares favourably to the UK’s 59% (MoJ, 2012). Evidence-based interventions consistent with effective treatment principles, adhering to Risk-Need-Responsivity models have been shown to reduce recidivism up to 35% (Andrews & Bonta, 2010). However, rehabilitative work in prisons is extremely difficult because institutions have ingrained attitudes and culture fundamentally punitive, not research-based or conducive to change, and consequently not therapeutic (Andrews & Bonta, 2010). Rehabilitation into society is challenging whilst simultaneously preventing societal interaction (Pakes & Winstone, 2011). Furthermore, overcrowding, now commonplace in prison, damages rehabilitative initiatives by decreasing amount of time individuals can engage with programmes (Long, 2018).
In England, after controlling for individual differences, reoffending rates for offenders on probation in 2007 were 7% lower than prisoners with short-term sentences (MoJ, 2010). Sapouna et al. (2011) suggest this is because offenders in communities have greater access to rehabilitation services. This supports perceptions of rehabilitation within prison being somewhat illusory and confirms ideas that non-custodial sentences are more effective at reducing crime (Mears et al., 2012). Unfortunately, current research suggests prisons most often achieve the opposite of rehabilitation (Pakes & Winstone, 2011), or at best only delay reoffending (Sedhley, Scott, Williams & Derrick, 2010).
Throughout this paper, it has been continually demonstrated that prisons fail to reduce reoffending in comparison to non-custodial sentences (Drago, Galbiata & Vertova, 2011). Moreover, prisons have criminogenic effects on subsequent offending, promoting criminality through strengthening deviant identity (Rodrigues et al., 2018). Reducing reoffending has critical implications for society (Bierie & Mann, 2017), yet, high recidivism rates suggest many offenders are negatively affected by prison (Cullent et al., 2011). Future research must improve recidivism measures and should empirically examine prison at both individual and institutional levels, creating hierarchical models (Harding, 2014). This alongside studies of specific crimes will inform rehabilitation (Gallant, Shery & Nicholson, 2015). Well-designed longitudinal research across cultures demonstrate prisons fail to deter (Andrews & Bonta, 2010), have minimal incapacitation effects (Sweeten & Apel, 2007), and fail to effectively rehabilitate (Schafer, 2018). Prisons do not prevent offending (Pakes & Winstone, 2011), essentially making society less safe (Bales & Piquero, 2012). Prison is an expensive strategy yielding diminishing results, which must be considered when formulating public policy (Petersilia, 2011).
Therefore, prisons appear counterproductive and psychological evidence seriously undermines utilitarian arguments supporting prison (Mears et al., 2012). Considering the research, it seems wrong to assume prisons ‘work’ and if prisons continue the same way, reoffending will never be reduced (Pakes & Winstone, 2011).