Articles


An Exploration of Third Party Disclosure and Outcomes in Registered Sex Offenders

Published 15/12/2015
Type Article
Author(s) Jackie Craissati, Claire Quartey
Corresponding Authors Jackie Craissati, Consultant Clinical & Forensic Psychologist, Oxleas NHS Trust Claire Quartey, Assistant Psychologist, Oxleas NHS Foundation Trust
DOA
DOI

Disclosure – sharing specific information about an offender with a third party – is an important tool in the range of risk management options available to agencies when managing convicted sex offenders placed on the Sex Offender Register in England and Wales. However, disclosure has not been widely researched, and assumptions that it is effective in reducing the likelihood of further sexual offending cannot be empirically supported as yet. This study aimed to gather descriptive data on the frequency and nature of third party disclosures on Registered Sex Offenders in five London boroughs, with reference to the method of disclosure, the outcome, and subsequent offence failures. One hundred and twenty nine third party disclosures were made over the one year period, most commonly to those in close relationships with the offender. Outcomes were variable in terms of acceptance and rejection, although the method by which disclosure was made had a significant impact on the outcomes. The rate of offender non-compliance with statutory regulations over the research period was relatively high suggesting that this was a group of less compliant offenders. Implications for future research and good practice are discussed.


Introduction
Disclosure is considered to be an important and necessary resource for the effective risk management of sexual offenders. Disclosure, in this context, refers to ‘the sharing of specific information about a MAPPA offender with a third party (not involved in MAPPA) for the purpose of protecting the public’ (MAPPA Guidance, 2012). This should not be confused with information-sharing between agencies.

MAPPA refers to the multi-agency public protection arrangements which were introduced in England and Wales in 2001 under Sections 67 and 68 of the Criminal Justice and Court Services Act (2000) to facilitate the sharing of information among agencies to improve public protection. MAPPA is now covered by the Criminal Justice Act (2003, section 325-327). Police, probation and prison services form the Responsible Authority with a statutory duty to ensure that the risks posed by specified sexual and violent offenders are assessed and managed appropriately, and other agencies – such as health, youth justice and the local authority – have a ‘duty to co-operate’ (Criminal Justice Act, 2003, section 325(3)).

One of the categories of offenders falling within the remit of MAPPA is Registered Sex Offenders (RSOs). These are convicted sex offenders who are subject to notification requirements under the Sexual Offences Act, 2003. Under the terms of the Act, RSOs must notify police of their name, address, all foreign travel, bank and credit card details, as well as information about their passports and other identity documents. RSOs – as with other MAPPA offenders – can be managed at three levels. Level 1 refers to ordinary agency management where the risks posed by the offender can be managed by the agency responsible for the supervision or case management. This involves information sharing but does not require multi-agency meetings. Level 2 is for those offenders whose risk requires active multi-agency management and for whom risk management plans are agreed via a multi-agency meeting. Level 3 is reserved for cases which meet the level 2 criteria but where the management issues require senior representation from the agencies, and there may be a need to commit significant resources at short notice.

The most recent MAPPA guidance (Section 10.2, 2012) sets out principles underpinning the use of disclosure. Disclosure should be lawful, proportionate, accurate and necessary. Proportionate refers to a number of considerations within the guidance, including the potential risk to the offender as a result of disclosure and the consideration and recording of alternatives to disclosure; the provision of information regarding key triggers for offending behaviour or other advice regarding management, not necessarily an offence history; and procedures to support both victims and offenders in case there is a breakdown in the process. Finally, the guidance suggests that ‘it is preferable that the offender knows that disclosure is taking place. On occasion, the offender may make the disclosure himself or herself in the presence of the police or the Offender Manager (probation officer), or may later confirm or verify the content of the disclosure’ (p.56).

Although there have been process evaluations of the MAPPA process (for example, Wood and Kemshall, 2007), there are only two published research studies on the use of third party disclosure under MAPPA. Cann (2007) reported on a survey of the 43 police areas in England, Wales and Northern Ireland. Only 29 of the 40 areas which responded had made disclosures in the preceding six months. The survey revealed that the primary reason for disclosure was child protection, usually precipitated by a change in an offender’s personal relationships, housing or employment situation. Most commonly it was the details of the convicted offence which were disclosed. Both negative and positive consequences were highlighted: potential destabilisation of the offender versus enhanced public protection by limiting opportunities for offenders to access risky situations.

Penny and Craissati (2012) surveyed regular participants in MAPPA level 2 meetings across London with a response rate of 196 questionnaires. Respondents expressed confidence in their knowledge of disclosure law and procedures, but widely divergent views in their response to a range of hypothetical disclosure scenarios. The authors raise implications of the research for practice including improved understanding and guidance around third party disclosure, additional training and governance oversight.

In 2007, the Review of the Protection of Children from Sex Offenders was published, which led to the Child Sex Offender Disclosure Scheme. This scheme was rolled out across England and Wales in three stages and was implemented across the country in 2011. The scheme is primarily the responsibility of the police, who manage a clear access route for the public to register their child protection interest in a named individual and – should there be relevant convictions and a risk posed – to receive a disclosure of pertinent information regarding the individual. A pilot over 36 weeks resulted in 53 enquiries, with 25 child concern reports being raised as a result (Chan, Homes, Murray and Treanor, 2010). Eleven of these reports involved disclosure to a parent regarding the offender’s conviction. The authors note that there was no indication that the scheme had had a negative impact on the behaviour of RSOs in the area, although this was not formally investigated. A related study with similar findings was conducted in four police areas in England and Wales (Kemshall and Wood, 2010). Interestingly, Kemshall, Kelly and Wilkinson’s (2012) reviewed the views of applicants within the Child Sex offender Public Disclosure Scheme. They found that in the few cases where a disclosure was made, the applicants who were not in a professional role remained anxious, and uncertain as to how to make use of the information.

The evidence from the United States, where community notification has been in place from the late 1990s onwards, is equivocal. Sex offender details have been publically available for a number of years; a summary of the evidence (Fitch, 2006) would suggest that there is no empirical evidence that community notification has had a positive impact on sex offender recidivism rates, nor has it resulted in fewer assaults by strangers on children.

There have been some attempts to consider disclosure from the offender perspective. Both Connor (2007) and Hudson (2005) had similar findings. Connor’s qualitative study found that sex offenders were deeply avoidant of situations potentially requiring disclosure, for fear of rejection and retaliation; the offenders had already constructed personal narratives around their offending which avoided stigma – minimising or lying – which disclosure threatened, not least because of a fear of losing control over the disclosure process, and it being used against them.

Traditional approaches to risk management have focused on identifying risk factors for targeted intervention, a strategy identified by Kemshall and Wood (2008) as a “protection strategy” which aims to protect communities through control of risk. This approach is often dominated by the offender adopting avoidance strategies which can lead to impoverished lives and hopelessness, as suggested by the participants in Connor’s (2007) study. There is a growing research literature which suggests that the progression from persistent offending to desistence is the outcome of a complex interaction between subjective and social factors (Farmer, McAlinden and Maruna, 2015); that is, structure (informal social controls such as relationships with intimate partners and employment) and agency (the way in which offenders think about themselves and their lives). For persistent offenders in particular, this entails finding a more positive identity as a nonoffender which includes meaningful engagement in the community, which has a more powerful impact on desistance than traditional risk management methods (Maruna and Farrell, 2004; Farmer, McAlinden and Maruna, 2015). The desistance literature also highlights the importance of building and sustaining hope in the journey from offending to a new non-offending identity, which includes others believing in the offender’s current personal identity rather than his ‘old me’. There has been little published research into desistance from sexual crime. Kruttschnitt, Uggen and Shelton (2000) conducted a retrospective study of 556 sexual offenders to examine whether informal social controls – specifically employment and marriage – predicted desistance from offending, and found that job stability significantly reduced the probability of re-offending but marital status did not. Harris (2014) in a qualitative study of 21 desisting sexual offenders found that the majority of them attributed their desistance to changes in thinking. The most recent qualitative study of apparently desisting sexual offenders (Farmer, McAlinden and Maruna, 2015) found that the offenders themselves attributed their desistance to cognitive factors, weighing up the advantages and disadvantages of repeating their crimes. The authors tentatively suggest that the desisting narratives in their study tended to support a move away from backward looking approaches towards future-focused therapeutic interventions with an emphasis on optimism and hope.

As yet, there is no published evidence available to confirm whether disclosure does indeed protect the public by managing (or perhaps avoiding) risky situations, or whether it inadvertently reduces opportunities for accessing socially valued roles (such as a parent, or skilled worker) which in turn impedes processes of recovery or desistance.

The aim of the study was to build on the limited evidence base in relation to disclosure under MAPPA arrangements, by focusing on RSOs and providing details of the frequency of disclosures and how they are made, the life domains in which they are made, and the outcomes in terms of inclusion or exclusion. There were no specific hypotheses, although it was expected that:

• disclosures would probably be made most commonly in higher risk sex offenders;
• there would probably be fairly high levels of rejection as a result of the disclosure.

Methodology
Participants
The sample included all those individuals convicted of a sexual offence, held on the Sex Offender Register in five boroughs of south east London, and who were subject to a third party disclosure at any time within the preceding 12 months (December 2013 to November 2014). The five boroughs were primarily chosen for their geographical proximity to each other (for each of data collection) and were considered to a reasonable spread of types of London borough; the lead author was known to the police teams in the area which greatly facilitated cooperation with the project.

Data was collected on the type of offence (child victims, adult victims, or non-contact sex offending), the start date of the Sex Offender Registration, the risk level (as measured by the Risk Matrix 2000, Thornton et al., 2003), and whether the offender was subject to a Sex Offences Prevention Order (SOPO). Additional data was collected in a semi-structured interview with the Jigsaw teams (the police teams responsible for participation in MAPPA and maintenance of the SOR), including details on the disclosures, how they were made (by the police first by telephone or visit, or offender first, and then followed up by the police), and the outcomes.

Disclosure domains were divided into the following domains: employment (current or potential), training establishments (such as colleges), relationships (family and friends), social (such as clubs, the church, or housing associations).

Outcomes were defined, in relation to the response of the person/organisation disclosed to, as follows: accepted, partially accepted (subject to conditions such as restricted access or supervised contact), rejected, and withdrawal (of the offender prior to the disclosure in anticipation of a disclosure being made).

Offence failures were defined as any re-arrest for a general, violent or sexual offence, or a recall to prison, between the time the disclosure was made and the data collection point. Therefore, the time at risk for a failure was up to a maximum of one year.

Procedure
Anonymised and non-identifiable data was provided by the Metropolitan Police Public Protection Unit, from the ViSOR7* electronic record system, where the specific indicator ‘disclosure’ was endorsed on the offender record. The ViSOR and PNC (police national computer) numbers enabled the Jigsaw teams to identify the offenders, and discuss disclosure information without identifying individual offenders. The Metropolitan Police lead for Public Protection contacted the head of each of the five Jigsaw teams, advising them of the nature of the research, and the support of the London MAPPA Strategic Management Board (SMB). Individual meetings were then set up with each of the five team leaders, in order to work through the semi-structured interviews, and complete any missing information.

The researchers were given local ethical approval by the National Probation Service (NPS) London Division to access the London probation electronic record system; additionally they were provided with a list of the Registered Sex Offenders held on post-custodial licence or community orders by the probation service in the five boroughs, and who had been in the community for the past 12 months (December 2013 to November 2014). This data was cross referenced and checked with the police data (above), and any additional offenders who had been subject to a probation third party disclosure were added to the sample.

Measures
The Risk Matrix 2000 (Thornton et al., 2003) is routinely completed by Jigsaw teams and/or probation officers from file information on all convicted sex offenders across London. It is an actuarial measure predicting sexual reconviction in sex offenders. It comprises a simple baseline risk classification based on conviction data, adjusted at stage two if two or more aggravating factors are found (for example, male or stranger victims). Two cross-validation studies tested the predictive validity of the scale in a short-term follow up sample of treated sex offenders and a long-term follow up sample of untreated sex offenders. The ROC Area under the Curve was 0.77 and 0.75 for the two samples. ROC provides a measure of predictive accuracy, with 0.50 representing no better than chance, and higher scores reflecting increasing accuracy. The two and sixteen year follow up recidivism rates were 0.9% and 8% low risk, 1.3% and 18.3% medium risk, 5.7% and 40.5% high risk, and 17.2% and 60% very high risk, respectively.

Ethical approval
Support for the project was provided by the Strategic MAPPA Board for London, and ethical approval obtained from the Metropolitan Police and from the Deputy Director of the National Probation Service for London. The consideration of safe and anonymised transfer and storage of sensitive information was an important consideration.

Results
The ViSOR records did not identify all the disclosure cases, and some cases recorded as disclosures in fact pertained to information exchange with other agencies. However, Jigsaw teams clearly knew their caseload of offenders on the SOR, and were able to identify all suitable cases with ease. Cross referencing with the probation service records confirmed the police disclosure in 10 cases, and identified a further 19 cases where the probation officer had made a disclosure, not already recorded in police records. An examination of these cases confirmed that there were no significant differences, and they were incorporated into the main analysis.

In December 2014, there were a total of 1,576 offenders on the Sex Offender Register (SOR) in the five boroughs in the study. There were 129 third party disclosures made for this group of offenders across the year of study, suggesting an approximate 8% disclosure rate within the year. All the participants in the study were male (there were no disclosures in relation to female offenders). Of the offenders on the SOR, 51 (3%) were being managed at MAPPA level 2 at the time of disclosure, and all the rest were managed at level 1, save two managed at level 3. The start of the SOR was unknown in 15 cases, but for the remaining 114 offenders, the average time on the SOR was 4 years 8 months (sd 4 years 2 months, range 1 month to 21 years 10 months). The results are set out in more detail in Table 1.

Table 1: Offence related details in relation to Borough

Disclosures were made about twice as often for those with child victims (63, 49%) as for those with adult victims (29, 23%) or those with non-contact offences (37, 29%). However, the spread of offender types was broadly similar across the five boroughs.

There were significant differences between the boroughs in terms of risk level, as measured by the Risk Matrix 2000 (X2 23,075, df 12, p<.05): Borough B and Borough D made disclosures on a much greater number of very high risk offenders, while Borough A and Borough E made the most disclosures on low risk offenders. There were also significant differences between the boroughs in terms of the MAPPA level at which the offenders (for whom there was a disclosure) were being managed (X2 27.540, df 4, p<0.01): Borough E and Borough D managed nearly all the disclosures at Level 1, while Borough A managed the majority at Level 2. There were no significant differences in relation to risk between offender types.

There were 47 (36%) offenders on Sex Offences Prevention Orders (SOPOs), with a proportionally greater number in Borough A borough although this was not a significant difference (p<.09). Although there were no significant differences in the number of SOPOs by offence type, there were significant differences in relation to risk level (X2 12,987, df 3, p<.01): 11 (52%) low risk, 8 (18%) medium risk, 10 (30%) high risk, and 16 (62%) very high risk offenders were on SOPOs.

For 10 of the 129 disclosures (8%), either the method of disclosure or the outcome was unknown, and these cases were therefore excluded from the following analysis. By far the most disclosures were made in the domain of relationships, and training was the least common domain. Overall the most common method was for the police to telephone the disclosure recipient (22, 19%) or to visit and disclose ((47, 40%). However, in terms of relationships, almost half the disclosures were made by the offender informing the recipient first and this being followed up by the police and confirming the disclosure. Table 2 provides more detailed findings.

Table 2: Method of disclosure and outcome in relation to disclosure domains

Disclosures and outcomes
Employment

There were 20 disclosures to potential or actual employers, two thirds of which resulted in rejection of the offender. Within the sample, only one offender (out of 29) with adult victims had an employment disclosure, and he was accepted. Fifteen (out of 62, 24%) disclosures were in relation to offenders with child victims, of whom 60% were rejected; all six (out of 37) of the non-contact offenders were rejected. Offenders were no more successful in applying to work for charitable organisations (paid and unpaid) than for private companies.

Training
Disclosures were evenly spread across offender types. Twenty-two percent of offenders were rejected by training establishments, either by withdrawing an offer, or by not responding following the disclosure. There was no difference in outcome in relation to how the disclosure was made.

Relationships
Disclosures were evenly spread across offender types. Two thirds of offenders (44, 67%) were accepted by those to whom the disclosure was made, and a further 3 (5%) were accepted with conditions, usually regarding supervised contact with their children. However, 19 (29%) were rejected following the disclosure. Forty of the disclosures were to partners (including ex-partners) or the offender’s children; the rest were made to friends and close family members with children. Offenders were half as likely to be rejected (37% versus 63%) if the offender disclosed first rather than the police (X2 15,395, df, 6, p<.05).

Social
Disclosures were evenly spread across offender types. These disclosures were to situations regarding shared housing, schools, churches and clubs. Overall around half resulted in rejections although the disclosure was significantly more likely to result in acceptance if the offender disclosed first (X2 15,395, df 6, p<.05).

Further analysis of the data to explore the possible relationships between the method of disclosure, disclosure domain, outcome and risk level did not result in any significant findings.

Offence failures
Overall, there were 25 (20%) failures, in terms of re-arrest or recall, following the disclosure, but within the study period. However, these failures were linked to 20 offenders, as one offender failed after two disclosures, and another offender failed after five different disclosures. There was one re-arrest for a sexual offence, and one for a violent offence; there were five general (acquisitive) re-offences, three for failing to reregister/breach of notification requirements, and seven breaches of SOPO. Three further offenders were recalled without any re-arrest for a further offence. There were no significant differences between offender types and failure overall, although there was a significant difference in the way in which offender types failed (X2 21,161, df, 10, p<.05): 75% of those failures with adult victims breached their SOPO, and 44% of failed noncontact offender generally re-offended. Although the numbers were very small, the one violent failure and all three of the recalls had been rejected following disclosures. The most significant association with failure was the risk level (X2 7.685, df 3, p<.05): 2 (10%) of failures were low risk, 3 (8%) medium risk, 8 (24%) high risk, and 7 (28%) of failures were very high risk. Furthermore, 7 (64%) of those on a SOPO failed by breaching their order (X2 11,380, df 5, p<.05).

Offenders were significantly more likely to fail following disclosure to employment if the police made the disclosure first (X2 16,232, df 4, p<.01). There were no significant findings in relation to disclosure method and failure in the other domains. However, those offenders who were rejected by training establishments were significantly more likely to fail (X2 9,000, df 2, p<.05).

Discussion
This paper presents the results of a descriptive study into the frequency and nature of third party disclosures on RSOs in five London boroughs, with reference to the method of disclosure, the outcome, and subsequent offence failures. The overall aim was achieved, providing data which can serve as a bench mark for other MAPPA areas in London and further afield.

There were some methodological limitations to the study. First, the official records were not entirely accurate, in terms of the ViSOR system, and although the police teams clearly had a good grasp of their caseload, some disclosures may have been missed. Furthermore, there seemed to be no routine recording within the probation electronic record of disclosures, and there was an impression that the probation service were only leading on disclosure decisions in a minority of cases (18 out of the 129). It is not clear whether this accurately reflects practice, but may raise questions as to whether probation staff should be leading such decisions where the offender is under their statutory supervision.

The decision to disclose, and the outcome of disclosure may have been related to factors not picked up in this study, such as the non-compliance of the offender, or the inappropriateness of the life domain in which they wished to participate. The study can make no judgment on the appropriateness or otherwise of the decision to disclose. The findings should therefore be interpreted with caution, and require further exploration.

Nevertheless, some interesting differences and associations were found. The boroughs were not identical in their practice, with one borough making twice as many disclosures during the study period as any of the other boroughs, despite no discernible difference in risk levels; and another borough managed the majority of disclosures at MAPPA level 2 as compared to the other boroughs who were making decision at level 1. The MAPPA Guidance (2012) suggests that disclosure decisions should be made at a level 2 meeting, although in reality, the lead agency may consider it more expedient to make decisions and carry out the actions without reference to the multi-agency setting.

The tentative hypothesis that disclosures would be more likely to be made on higher risk RSOs was only partially borne out, with a significant number of low risk RSOs in two boroughs also being subject to disclosure. Similarly, there were significantly more SOPOs (in the disclosure group) in the low and the very high risk category, as compared to the medium and high risk categories. These findings suggest that there is a group of low risk offenders who are subject to stringent risk management arrangements. Without further information, it is not possible to know whether this relates to the particular living circumstances of the offender – perhaps a family setting with available child victims – or different professional practices across boroughs.

Disclosures were made twice as often on offenders with child victims than for those with adult victims or non-contact offences. Again, it is difficult to draw conclusions without knowing the proportion of offender types on the SOR; however, there were no differences in risk level across offender types, and it may therefore reflect a greater focus on those with child victims, mirroring public concerns.

The suggestion that disclosures would result in rejection was also only partially borne out. As might be expected, the greatest number of disclosures were made in the relationships domain, but this was also the category most likely to lead to acceptance. This may be of some reassurance to offenders who fear rejection within relationships as a result of disclosure. However, the protective quality of these relationships was not investigated in this study, and there is the possibility that the acceptance into relationships led, in some cases, to further involvement from social services. Although training was the least common domain, the outcomes were also encouraging with around 80% acceptance. The caveat to this is that those who were rejected from training establishments were significantly more likely to fail than those accepted.

The picture was least encouraging with the employment domain, and to some extent, the social domain, with relatively high levels of rejection. This may represent the appropriate management of high risk situations, but raises the possibility that RSOs are encountering obstacles which may not always be warranted, when seeking the social capital (i.e. the ‘new me’ roles) which is central to a model of desistence from offending. Potential employers, charities, clubs, churches and housing associations may not have any significant investment in tolerating or managing risk in the current climate of public concern.

The rate of offence failure (20%) during the study period seemed to be rather high, although only one offender was re-arrested for a sexual offence. Half of the offence failures related to breaches of statutory requirements (SOPOs and SOR notification). In particular the majority of offenders with adult victims breached their SOPO, and almost half the non-contact offenders were re-arrested for general offences. This raises the possibility that these were particularly non-compliant or antisocial offenders; an interesting question is whether disclosures tended to target less compliant offenders, whether less compliant offenders were more likely to be alienated by the disclosure process, or an element of both.

The disclosure method yielded some interesting findings which may have implications for practice. Although not a universal finding, it would appear that when the offender was able to make the disclosure first – followed up by the police – the acceptance rate was significantly improved, particularly in the relationships and social domains. Furthermore, when the police initiated disclosure in the employment domain, this was significantly associated with subsequent offence failure. This suggests that allowing the offender opportunities to initiate the process of disclosure may improve the offender’s sense of control, and enhance the willingness of the person/organisation to whom the disclosure is made, to be accepting. However, not surprisingly, the most robust association with offence failure was the risk level, which provides some additional reassurance that the RM2000 is a reasonable tool for use by MAPPA agencies.

The implications of this study for future research and practice fall into two broad areas:

• Replicating the investigation with a larger sample across London or in less urban environments, in order to explore further the decisions to disclose, and their positive impact in terms of effective risk management, reduced offence failures and increased community reintegration.
• Investigating the emerging possibility that it might be feasible to improve social integration and acceptance by developing a good practice model for carrying out disclosure – including allowing the offender to make the disclosure first – which could be further evaluated.

Disclosure remains an important and potentially powerful tool for effective risk management. Public protection and the needs of victims should always be the priority. However, this study provides an opportunity to ensure that disclosure decisions are consistent across areas and between agencies, and raises opportunities to enhance good practice models which have positive outcomes for both offenders and victims.

 

Acknowledgements
The authors would like to acknowledge the support of the London MAPPA Strategic Management Board (SMB) in conducting this research, and to thank the Metropolitan Police teams who gave their time to engage with the study.

7* ViSOR is a UK-wide system used to store and share information and intelligence on those individuals who have been identified as posing a risk of serious harm to the public; it is designed to facilitate the work of MAPPA by assisting co-operative working between the three ‘Responsible Authorities’ (police, probation and prison services).

8* Total on Sex Offender Register (SOR) as at December 2014.

References

Cann, J. (2007). Assessing the extent of discretionary disclosure under the Multi-Agency Public Protection Arrangements (MAPPA). Report 13/07. London: Home Office Online.
Chan, V., Homes, A., Murray, L. and Treanor, S. (2010). Evaluation of the Sex Offender Community Disclosure Pilot. Ipsos MORI: Scotland
Connor, J. (2007). Disclosure of Sexual Offending: Offenders’ Perspectives. Doctoral thesis (Clinical Psychology): University of East London.
Farmer, M., McAlinden, A, and Maruna, S. (2015). Understanding desistance from sexual offending: A thematic review of research findings. Probation Journal, 62 (4), 320-335.
Fitch, K. (2006). Megan’s Law: Does it Protect Children? NSPCC.
Harris, D. (2014). Desistance from sexual offending: Findings from 21 life history narratives. Journal of Interpersonal Violence, 29(9), 1554-1578.
Home Office (2003) Criminal Justice Act 1991. London: HMSO.
Hudson, K. (2005). Offending Identities. Sex Offenders’ Perspectives on their Treatment and Management. Devon, UK: Willan.
Kemshall, H., Kelly, G. and Wilkinson, B. (2011-2012) Child Sex Offender Public Disclosure Scheme: The views of applicants using the English pilot disclosure scheme. Journal of Sexual Aggression, 18 (2) 164-178.
Kemshall, H. and Wood, J. (2008). Risk and Public Protection: Responding to Involuntary and ‘Taboo’ Risk. Social Policy and Administration, 42(6), 611-629.
Kemshall, H. and Wood, J. (2010). Child Sex Offender Review Public Disclosure Pilots: A Process Evaluation (2nd edn). Home Office.
Kruttschnitt, C., Uggen, C. and Shelton, K. (2000). Predictors of desistance among sex offenders: The interaction of formal and informal social control. Justice Quarterly, 17(1), 61-87.
Maruna, S. and Farrall, S. (2004). ‘Desistance From Crime: A Theoretical Reformulation’ , Kölner Zeitschrift f ür Soziologie und Sozialpsychologie 43: 171-194
National MAPPA Team, National Offender Management Service (2012). MAPPA Guidance: Version 4. https://www.justice.gov.uk/downloads/offenders/mappa/mappa-guidance-2012-part1.pdf
Penny, C. and Craissati, J. (2012). Decisions on disclosure to third parties made at MAPP meetings: opinions and practice. The Psychiatrist, 36, 379-385.
Thornton, D., Mann, R., Webster, S., Blud, L., Travers, R., Friendship, C. and Erikson, M. (2003). Distinguishing and combining risks for sexual and violent recidivism. In R.A. Prentky, E.S. Janus, & M.C. Seto (Eds.) Sexually coercive behavior: Understanding and management (pp. 225–235). Annals of the New York Academy of Sciences (Vol. 989). New York: New York Academy of Sciences.
Wood, J. and Kemshall, H. (2007) The operation and experience of Multi-Agency Public Protection Arrangements (MAPPA). London: Home Office.