Banner extension SCoPiC Banner
Network Members
  Policy and Prevention
Collaborative Studies
Contacts & Links

The Sheffield Pathways Out of Crime Study (SPOOCS)

Principal Investigator: Professor Sir Anthony Bottoms

Institution: University of Sheffield


Overview and Objectives

The Sheffield Pathways out of Crime Study SPOOCS) is a longitudinal desistance study of more than 100 persistent young adult offenders which used quantitative and qualitative methods to further two interlinking research objectives:

  • to obtain a fuller understanding of the processes of reduction in, or desistance from, offending in early adulthood
  • to analyse life patterns of persistent young adult offenders to identify factors which differentiate those seeking to desist

Non-occasional young adult offenders were followed for four years from their early 20s, the apex of the age-crime curve. SPOOCS is the first UK study to employ this design and one of the few existing desistance studies which has collected data on both official and self-reported offending.

In the context of existing criminological literature on desistance, SPOOCS is distinctive in analysing the early stages of desistance in a sample of mostly persistent offenders, and this special focus emphasises both the precariousness and the sense of struggle involved in the process of desistance. The researchers have developed a tentative model of desistance which aims to capture these features. In doing so, it emphases the dimension of agency, the offender’s response to the various social situations he encounters, as well as the varying social capital available to offenders as they seek to change their lifestyle.

SPOOCS was particularly interested in the relationship between community and individual factors in the explanation of offending, desistance and crime-free gaps, and builds on a long tradition of work at Sheffield University on the study of crime in its urban context.

back to top


Study D

The sample consisted of 113 persistent young adult male offenders identified using Police National Computer Data. Participants were selected using four criteria; the first two were chosen for the overall research design, the second two for reasons of practicality. All SPOOCS participants

  • were born between 1982 an 1984 (at first interview the mean age was 20 years 9 months)
  • had been convicted of one of more “standard list offences” (a Home Office category which includes offences such as theft, burglary and violent but not most motoring offences)
  • had recent contact with the criminal justice system
  • had an address in Sheffield at the time of their last conviction

Because of the Data Protection act, the Probation Services approached eligible persons to ask them to take part. Because approaches made by letter to offenders in custody produced a more positive response than approaches made in person to offenders being supervised in the community, 82% of the first research interviews took place in custody. As a direct result of this the achieved sample had a significantly more criminally active profile than originally anticipated. In additions, given low numbers of females in custody, it proved impossible to recruit more than a handful of female offenders, so it was decide to opt for a male only sample.

SPOOCS collected data from two primary sources:

  • official criminal records, obtained from the PNC
  • a series of one-to-one 90 minute interviews with each participant

Participants were interviewed on four separate occasions at 9-12 months. Retention rates were high, with 78% taking part in the fourth interview.

  • 113 participants completed the first interviews
  • 98 participants completed the second interview (87% of the original sample) after an average interval of 13 months
  • 88 participants completed the third interview (78% of the original sample) after an average of 11 months
  • 88 participants also completed the fourth interview after an average of 11 months

In some cases, contact was missed at one of the interviews but re-established later. Thus, although the response rate for both the third and fourth interviews was 78%, a total of 97 participants (86% of the original sample) completed either the third or the fourth interview. Overall, the interview recontact rates can be regarded as very satisfactory for this kind of study.

During the research interviews participants were asked about

  • where they lived
  • how they spent their time
  • their relationships with their families, girlfriends and mates
  • criminality and victimisation
  • drug and alcohol use
  • views of the criminal justice system
  • hopes and plans for the future

The interviews also include a number of formal scales and checklists (such as Beck’s Hopelessness Scale) which are well validated from previous research.

back to top


Select Findings

  • Rates of criminality

Even when first contacted, the sample has high rates of criminality: on average, by age 21 they had experienced eight court appearances involving a conviction for a standard list offence (for example, theft, burglary and/or violence).  By age 21 the sample had 909 stand list conviction occasions, and by the end of data gathering (August 2007) had acquired a further 263. About half also had a significant drugs problem. As expected, there were many social problems in their past histories, such as exclusion from school and significant levels of unemployment.

back to top

  • Persistence

During the three of more years of follow-up, 80% of the sample members were reconvicted for a standard list offence and, in the self-reported criminality section of the interviews, some of those who were not reconvicted admitted to further offending. 30% of the men admitted 150 or more offences in the previous year (adjusting for time at liberty); the offences admitted were mostly property offences such as burglary, theft and taking car, but there was also a fair amount of violence, including robberies. Thus most of the sample did not completely desist from crime. However, even at the time of the first interview over half the sample expressed a definite wish to stop offending, and this proportion subsequently increased. Moreover, when offending levels during the period of the research were compared with those in the year before the first interview, there was definite evidence that the average frequency of offending had significantly reduced. Comparing self-reported criminal at the first and fourth interview, there was also definite evidence of polarisation in the sample: the self-reported criminality of the majority had reduced, but there was a minority who retained very high offending levels.

Offending Group Reconviction Scores (OGRS), derived from a Home Office reoffending prediction instrument which uses age and measures of previous criminality (e.g., total number of conviction occasions, age at first conviction, and number of custodial sentences before age 21) to predict the likelihood of future recidivism, were calculated for the sample at the time of the first interview and correctly predicted the 80% reconviction rate. However, OGRS was not the strongest predictor of post-first-interview frequency of offending, whether measured by official convictions of self-report. This suggests that past criminality does not necessarily determine offending levels in the early 20s, and that new factors emerge to help alter old patterns of behaviour. Qualitative data from this study suggest that a very important feature in this regard is the onset of adulthood and the realization that the advent of new adult roles (e.g., father, breadwinner) might require a change of lifestyle.

back to top

  • Drug use

It is often assumed that persistent criminality is associated with drugs. According to the men’s self-reports, that was not the case in this sample. 47% were classified as ‘drug problem’ offenders, the great majority of whom admitted to drug dependency. Thus, at the first interview, the sample was equally divided between drugs problem offenders and others.

back to top

  • Desistance

Regression analyses of official and self-reported criminality showed that one of the strongest predictors of reductions of or persistence in offending was the respondent’s own assessment of the number of “obstacles to desistance” that he faced: that is, his assessment of how likely it was he would encounter difficulties such as continued drug use, lack of work, being tempted by opportunities for easy money, or the felt need for excitement through the commission of crime. Other predictors of post-first-interview desistance included objective measures of employment and drug situations, the relative absence of criminal friends, and a measure of “empathy”. These variables strongly emphasize the immediate pressures of day-to-day living for many of the men, and the way in which pressures towards criminality can arise even for those who have in principle decided to stop offending. The “empathy” measure, which seemed to increase in importance over time, is also interesting as it suggests a developing awareness, among those beginning to desist, of the need to take account of others’ needs and feelings in social contexts.

back to top

  • Social lives

One important but often overlooked source of social capital available to offenders which was identified by this study is parental support, even for offenders in their early 20s. At the time of the first interview just over half were (perhaps surprisingly) living with a parent or parents). A fifth were living with a girlfriend or partner, and a further half either were currently or had recently been in a steady relationship with a girl. Since a great majority of both parents and girlfriends disapproved of offending, these social ties produce a pull towards desistance. Set against these influences, however, most respondents said their mates were very important to tem and they often trusted them deeply – but two-thirds of the sample said that at least ¾ of their friends had criminal records. Thus at the apex of the age-crime curve, there are often competing influences in the lives of persistent offenders, as between parents, girlfriends and mates.

Most offenders felt they received only limited support from the probation service in their attempts to desist. Most reported that interviews with probation officers were short, and that the discussion tended to be about ‘general stuff’ and ‘how I am managing’. Only 1/3 of offenders considered the probation supervision was ‘very’ or ‘fairly’ useful to them. Given that most offenders did want to desist, and that this is also the wish of the probation service, it would seem that there is scope for reassessing probation work with offenders of this kind, to see whether relationships that are seen as more useful to the men can be built.

back to top

  • Obstacles to desistance

The sample displayed many of the familiar social handicaps of persistent offender populations. For example, about half the men had been excluded from school for at least a month during their school career; 86% had left school with no qualifications (thought some had acquired some since, mostly in young offender institutions); and nearly 60% had had no job of any kind (including casual jobs) during the year before the interview. Thus, if members of the sample wished to ‘go straight’, their lack of qualifications and work experience would be disadvantageous in the employment market. Such problems would also inevitably be magnified by their criminal records.

Statistical analysis suggested six ‘obstacles’ from Burnett’s ‘obstacles’ checklist interfered robustly with the intention to desist:

  • Need for excitement or to relieve boredom
  • Lack of money
  • Taking drugs
  • Opportunity to make easy money
  • Lack of work
  • Where I live

back to top

  • Intentions to desist

At the first interview, respondents were asked which of the following three statements was closest to their present view:

  • ‘I have made a definite decision to try to stop’
  • ‘I would like to stop but I’m not sure if I can’
  • ‘I am unlikely to stop’

In alter interviews a fourth category was added:

  • ‘I have stopped offending’

At the first interview, 56% said they had made a definite decision to stop; thereafter, the proportion saying they definitely intended to stop, or had stopped, increased in each interview. Moreover, although individuals’ desistence intentions sometimes fluctuated from interview to interview, an expressed intention to stop did predict subsequent lower self-reported offending. On this bases, and the more qualitative evidence from participant interviews, it seems that most expressions of ‘intent to desist’ were genuine. A significant reasons for the high incidence of intentions to desist appears to be the onset of adulthood: there was much comment in the research interviews about ‘growing up now’, ‘becoming more responsible’, and so on, often linked to a triggering event such as the importance of a relationship or the birth of a child.

‘Empathy’ was measured by a short scale which focussed on issues of mutual responsibility and awareness of others’ feelings in social interactions. At first interview, there was no significant relationship between intention to desist and empathy, but there was such a relationship at both interviews 2 and 4 (empathy was not measured at interview 3). This suggests that empathy is an emerging concept as offenders take a step towards desistance.

Parental attachment was also related to the intention to desist, suggesting that it may constitute an important source of social support and social capital as men seek to move towards desistance. The Probation Service typically does no attempt to contact parents in the case of offenders over 18, but in doing so they may be ignoring a potentially important support in the quest for desistance.

Perceived ‘obstacles to going straight’, having a drug problem and empathy predicted self-reported offending. The appearance of empathy here seems to confirm its importance to desistance. Dealing with drug problems is often a ital step in helping to pave the way towards the desistance. Overall, there was a reduction from interview 1 to interview 4 in self-reported problematic drug use. This was highly skewed, being particularly evident in those saying they had stopped offending at interviews 3 and 4.

A further important result shows that OGRS scores were still related to self-reported offending at interview 4, but that this relationship was weaker than a number of variables relating to the more immediate situation, such as drugs and empathy. This was also true for official criminality. These results show that fresh factors in offenders’ lives help to shape desistance as well as the continuing influence of historic variables. OGRS correctly predicted the aggregate simple reconviction rate in this sample, but findings suggest it is not the best predictor of frequency of offending (official or self-reported). This has potential practical implications and suggests the Probation Service should perhaps consider whether other checklists, such as intention to desist and offender self-perception of obstacles to desistance, might be more helpful.

back to top

  • Conclusions

90 of the 113 men (79.6%) were reconvicted for at least one standard offence in the period from their first interview to 31 August 2007, a period (on average) of over three years. Self-report data also showed that some official ‘desisters’ did in fact commit further offences. Is this, then, despite the strong evidence of ‘intentions to desist’, in reality a non-desisting sample? Much depends on how desistance is defined. If it means complete cessation, then most of the men did not desist. Considering frequency of offending, however, a different picture emerges. The mean number of standard list offences per year at risk was 8.0 in the year before the first interview, but only 3.8 during the period of the research. The results are essentially the same for self-reported offending: only a few men stopped completely by interview 4, but there was a marked reduction in average criminality levels. There was an interesting polarisation between interviews 1 and 4, leading to a U-shaped distribution; a significant minority maintained very high criminality level, but the majority reported reductions. The criminal careers literature has not concentrated on the early 20s and this differentiation and polarisation at this age in a new and important finding.

Perhaps not surprisingly for such a highly offending sample, this demonstrates steps towards complete desistance, in other words, early stages of desistance, rather than complete cessation.

back to top

  • A tentative model of desistance

SPOOCS researchers have forwarded a tentative model of ‘steps towards desistance’. Desistance in a gradual and often a stumbling process, with ample opportunity for relapse.

Two features influence all aspects of the progress of individuals:

  • Pre-programmed potential (e.g., social and criminal history)
  • Social capital (positive and negative and changes over time)

These links typically feed in at different points in time and can create opportunities to try out a non-offending lifestyle, to reinforce an alternative lifestyle, or to drag people back to offending. The more formal actions of criminal justice practitioners can also constitute facilitators or knockbacks in this complex process.

The general desistance proves has 5 stages:

  • Current offending is influence by a triggering event…
  • …which leads to the decision to try to change
  • …which leads the offender to begin think about himself differently
  • …which leads the offender to take action towards desistance
  • …which requires maintenance; the offender looks for reinforcers, but may encounter obstacles
    • a failure to maintain desistance in the face of obstacles may lead to relapse and a return to the beginning of the cycle
    • successful maintenance and reinforcement in the face of obstacles may result in adoption of a crime-free identity as a non-offender (i.e., desistance)

This model contains both a representation of the individual’s dynamic, if often hesitance, pathway and representations of the continuing influence of the past, and of social capital in the present. These are precisely the dual individual/social emphases of the SCoPiC Network; thus SPOOCS has demonstrated their importance in the study of desistance among young adults. 

As a longitudinal study focussed on persistent offenders and utilising quantitative and qualitative research methods, SPOOCS makes a distinctive contribution to the study of desistance. A rich dataset has been assembled and will repay much further analysis. The researchers now plan to embark on an active programme of dissemination, both of the results and their obvious potential relevance for policy and practice.

back to top


Future directions and implications for policy and prevention

At the 2007 SCoPiC Conference, SPOOCS researchers propose a ‘Second Start’ policy initiative, aimed at making the Probation Service a key agent for facilitating change towards desistance in this age-group. These proposals were repeated at a follow-up conference for UK policymakers arranged by Professor Paul Wiles, Government Chief Social Scientist, and received favourable attention.

SPOOCS researchers have also scheduled presentations to the National Offender Management Service (NOMS) and the Cabinet Office in connection with their current review of youth and young adult crime policy.

Future research priorities for SPOOCS research are:

  • To continue analyses of the quantitative and qualitative data
  • To search PNC data on individual offenders to confirm trends towards desistance (or persistence)
  • To carry out an additional interview with participants who are still contactable
  • To organize an international symposium for those working on desistance to compare results and theoretical explanations from both ‘early-stage’ and ‘later-stage’ desistance analyses, attempt a synthesis, and suggest new steps
  • To formulate in detail, with practitioners and policy makers, a ‘Second Start’ programme, implement it, and evaluate its effectiveness

back to top            Back to Studies page