Abstract
This article draws on Incident Reporting System data from the National Offender Management Service over a ten-year period (2004–2014) and limited, small-scale interviews with four custodial managers. Pat Carlen's work (2008) on imaginary penalities provides the theoretical framework for an assessment of the reporting, recording and initial response to sexual assaults in prisons in England and Wales. The article argues that the recording of sexual assaults became part of a response to new management systems that emphasised compliance, process and audit rather than realising safety in custody. Although the data shows substantial levels of initial activity among staff it is, in essence, practice without prospect. The article suggests that outcomes generally for sexual assaults in prisons in England and Wales are uncertain. Incident reporting has become a bureaucratic process ‘or paper shadow’, which Goffman described as showing ‘what has been done by whom, what is to be done, and who last had responsibility for it’ (Goffman 1961: 73).
Introduction
The reporting, recording and initial response to sexual assaults in prisons in England and Wales has been a neglected area of academic enquiry (Wilkinson and Fleming 2021). Although there has continued to be a focus on criminal justice responses to rape and sexual assaults committed in the community, this interest has not extended to those incarcerated in prisons in England and Wales. This article draws on National Offender Management Service (NOMS) data covering all reported sexual assault cases in adult men's prisons in England and Wales for the period 2004–2014. The data includes all prisoner-on-prisoner incidents involving adult males, substantiated or unsubstantiated, classified as being a ‘sexual assault’, recorded on the Incident Recording System (IRS) and managed during the data sample period by NOMS. The IRS data and its analysis are supplemented for context with small-scale interviews with four custodial management staff.
In exploring prison officers’ recorded data, the article draws on the work of Pat Carlen (2008) and other theorists (Bennett, 2016; Crewe, 2009; Feeley and Simon, 1992; Garland, 2001; Goffman, 1961; Liebling, 2000). The article assesses the applicability of Carlen's imaginary penalities theory to explain activity that epitomises the initial response to reports of sexual assault in prisons, but results in few formal outcomes and may collectively be described as practice without prospect. The IRS data, where recorded, refers to outcomes as incidents having been referred to police, subject to a police investigation, prosecution pending, subject to prison internal disciplinary processes and an ‘unknown’ category. This article discusses outcomes in the same context. In discussing the reliability of the outcomes data, the Ministry of Justice (MoJ) (2015) makes clear that outcomes, that is those that result in adjudications or prosecutions, are uncertain and difficult to identify in the data as collected. As such, they are not included as official statistics in the Safety and Custody bulletins because the data is ‘not deemed to be of sufficient quality to be published as Official Statistics’. The article demonstrates that the recording of sexual assaults in the period under review became part of an audit system reporting on order rather than necessarily delivering its commitment to safety in custody. We draw on Bennett's work (2016) to demonstrate that the new managerialist ideas of the 1990s compounded compliance with systems and audit rather than specifically responding to prisoners’ needs.
The article is in five parts. Following some brief contextual information, the first part provides the theoretical framework established by Carlen (2008), Goffman (1961) and others for an assessment of the reporting, recording and initial response to sexual assaults in prisons in England and Wales. The following section details NOMS data and methods covering all reported sexual assault cases in adult men's prisons in England and Wales for the period 2004–2014. The third section provides analysis of the NOMS data in terms of the outcomes recorded. The fourth section provides perceptions from four custodial managers (CMs) about their initial responses to reported sexual assaults in prisons and their attitudes towards the general process. Primarily, these interviews were used to assist in the researchers’ understanding of the reporting and recording process. Quotes used in this article should be considered in that context. The final section discusses the applicability of Carlen's theory to an explanation of high levels of recorded activity in response to reports of sexual assault in contrast with the stated outcomes.
The prison context
The rehabilitative ideal that characterised penal policy until the late 1970s gave way to a notion that rehabilitation did ‘not work’ and the retributivist idea that those who committed crime should be punished and prison should be seen as a deterrent. The punitive approach to crime developed over a decade with the 1990s delivering new and increased powers for prison staff and new logics associated with the administration of prisons and their inmates. Critical reports such as the Woolf Report in 1991 and Learmont's in 1995 combined with neo-liberal ideas of new managerialism that emphasised target-setting, compliance and order (Bennett, 2016: 7–10; Learmount, 1995; Liebling, 2000; Woolf, Rt Hon. Lord Justice, 1991). All of these contributed to the drive for change and penal reform. New ways of recording, reporting and exercising power shaped the role of prison officers as they sought to balance their relationship with prisoners and their obligation to ‘assist with the day-to-day management in the complex environment of the prison’ (Liebling, 2000: 338). In their critical review of what they coined the ‘new penology’, Feeley and Simon (1992) identified a marked shift away from reform and rehabilitation as organisational goals were replaced with a focus on managerial processes and the risk management of offenders. In 2003, the Carter Report (2003) recommended stronger offender management, greater use of competition in prisons and, importantly, the establishment of NOMS. By 2004, prisoner numbers were increasing and sentences lengthening and by 2010, the prison population had exceeded 85,000 (Bennett, 2016). However, prison officer numbers were relatively stagnant at approximately 48,000 across England and Wales. By 2006, however, staff numbers were in sharp decline, falling to fewer than 25,000 as prison populations rose (Arnold et al., 2007), placing the prison system under considerable strain. As Anne Owers, ex-Chief Inspector of Prisons (2001–2010) noted (2007), prison managers were expected to do more with less and found themselves responsible for areas they had no knowledge of, or control over ‘in the face of an ever-rising population’.
Although there has been literature on sexual assault in prisons, primarily from the United States, that literature, pre-2003, has not concerned itself with responses to sexual assault in prisons, but rather the challenges associated with distinguishing between consensual relationships and coercion (Eigenberg, 2000a, 2000b; Hensley and Tewksbury, 2005; Stevens, 2013). In addition, it describes the institutional (or prison-based) characteristics (Jenness et al., 2007; Toch, 1992) associated with sexual victimisation (such as overcrowding, cell-sharing and the nature of the prison regime itself) as well as the individual characteristics of victims as ‘targets’ and perpetrators of sexual assault (Banbury, 2004; Banbury et al., 2016; Eigenberg, 1992; Hensley, Koscheski and Tewksbury, 2005; Lockwood, 1980). A more recent body of research from the United States, conducted following the implementation of the Prison Rape Elimination Act 2003, designed to prevent, detect and respond to prison rape, has focused on the processes in place to support reporting and eliminate offending. Fowler et al.'s research (2010) indicated that new prisoners were more likely to self-report sexual victimisation; perhaps because of newly implemented reporting initiatives, or because they were less impacted by traditional views about ‘snitching’. Smith and Dunton (2022) concurred that prisoners in the southern states of the United States supported institutional prison mechanisms that reduced risks of prison rape and endorsed the view that reporting rape was widely supported among prisoners.
Theory
In her study of an Australian women's prison, Carlen (2008) used the term ‘imaginary penalities’ to describe what she saw as the unsuccessful prison-based policies and rehabilitation practice in which ‘agents’ of the prison (or staff) continued to invest time and money. This, Carlen argued, despite collectively and consciously knowing that their efforts were futile because the rehabilitation/reduced offending objectives on which their funding was dependant were unachievable.
Carlen (2008) suggested that the prison rehabilitation ideals were ‘imaginary’ because they were conceived on ‘imaginary prisoners’, the large proportion of whom received short custodial sentences. Short sentences were widely known by staff to inhibit completion of in-prison ‘imaginary’ rehabilitation programmes (Carlen, 2008). Despite this, penalities continued to be applied by agents of the criminal justice system because judges and magistrates remained unaware of their shortcomings and accepted the prison governance narratives that they worked (Carlen, 2008). The ‘imaginary’ status of the prisoners and their associated rehabilitation programmes was underpinned by ‘imaginary back-up’ (or support services) in ‘imaginary communities’ (without structural inequalities), neither of which existed in the real world.
Carlen (2008) spoke of ‘the imaginary’ to describe a ‘mode of knowing’ that was exercised by prison staff to suppress ‘other’ knowledge in order to make sense of the ‘anomic contradictions’ and requirements of the strategy and governance of the prison. In identifying the gap between the reality of the social conditions in which the prison practised rehabilitation and the ideology (or ‘the imaginary’), Carlen (2008) argued that the prison continued to operate by suppression of ‘the other’ knowledge. In her example, ‘the other’ was the reality that there were no effective support or rehabilitation services outside the prison and therefore released prisoners were vulnerable to recidivism (Carlen, 2008). Suppression of knowledge of ‘the other’ allowed staff to continue to deliver (and refer prisoners to) rehabilitation services as if the rehabilitation programmes fully realised their outcomes.
This observation is not new. The failure to focus specifically on rehabilitation was also a central theme in Goffman's Asylums (1961:73) in which the contradiction between what an institution was supposed to do, and what it actually did, ‘formed the basic context of the staff's daily activity’. Goffman (1961) recognised the contradictions for staff working in ‘total institutions’ and acknowledged that some staff joined the institutions as professionals, expecting to practise according to their ‘calling’ and instead discovered that they were exploited to ‘add professional sanction’ to the system. The imagined or intended focus of professionals, Goffman (1961: 88) argued, was invalidated by the need for staff to ensure that ‘inmates’ were obedient while ‘giving the impression that humane standards are being maintained and the rational goals of the institution realized’. Garland went further (2001) and argued that the pretence that rehabilitation programmes have an impact on prisoner outcomes had now given way to an acceptance that the primary role of prisons in England and Wales was protection of the public from the offenders therein. Carlen's concept of ‘imaginary penalities’ has been applied in this article to demonstrate the nature of much of the staff-audited activity in the period under review, which effectively amounts to practice without prospect.
For Carlen (2008), the requirement for prisons to audit and publish their achievements, rather than admitting to policy failures, was central to maintaining the legitimacy of short custodial sentencing in Australian prisons. Carlen's (2008) case study suggested that although prison staff had agency to influence minor adjustments to the rehabilitation programme outcome measures, they retained no power over influencing the overall rehabilitative ethos and objectives of the prison or the wider system. Therefore, staff continued to work towards accreditation, evaluation and personal appraisals in full receipt of the knowledge that goals were broadly unrealisable. The theoretical framework of ‘imaginary penalities’ as the outcome of ‘risk-crazed models of governance’ emphasised the role of positivistic social accounting to generate the misapprehension of policy realisation, rather than acknowledging failure (Carlen, 2008). Here, Carlen echoes Feeley and Simon's (1992) description of the new penology, which asserted that the focus of prisons was the management of prisoner populations based on categorisation and risk assessment.
Carlen's (2008) unique contribution to the debate was her assertion that problems with operationalising crime and risk-reduction programmes were well known and accepted by prison staff. The concept of ‘imaginary’ usefully reinforces the state of consciousness of the prison staff, explaining their continuation of ineffectual rehabilitation programmes as if these objectives were fully realisable. Imaginary penalities contended that staff and the prison governance met auditing requirements by editing performance measures to ensure that they recorded the ‘effectiveness’ of programmes. Auditing activity was carried out in the full knowledge that it supported the ‘official story that prison works’ therefore increasing commitment to that system and inflating prison populations (Carlen, 2008).
Carlen's (2008) focus on the role of evaluation in emphasising the value of interventions, observed that the prison staff in her study, ‘informally modified’ the measurable outcomes to reflect units of staff activity rather than the unrealisable goals set to reduce crime. It is this notion of continuing activity in the face of unrealisable goals and specific outcomes that is the focus of this article. In her case study, Carlen noted that changing measurable ‘outcomes’ to achievable units ensured that audits reported positive evaluations, even though the broader outcomes of the interventions were unrealisable. Carlen's (2008) examples of crime reduction measurements, such as recidivism, were replaced with assessments of how programmes were run, satisfaction levels and less-stringent measures such as the length of time between imprisonments. These modified outcomes were presented by prison staff as positive statements about the effectiveness of the intervention.
The erosion of discretion from agents of the current criminal justice system (Carlen, 2008; Garland, 2001; Hall, 2016; Tombs, 2008) and suppression of knowledge about the failures of the system, enabled prison staff to continue the day-to-day activities of the prison as if prisoners could realise rehabilitation on their release (Bennett, 2016; Carlen, 2008). The ways in which prisoners undertake their sentence and negotiate their path through the prison system and their sentence has been described by Crewe (2011) as a modern embodiment of the ‘pains of imprisonment’, originally described by Sykes (1958). Although conditions in prisons have improved from earlier iterations of incarceration (Matthews, 2005; Owers, 2007), Crewe (2011) argued that current approaches to imprisonment and rehabilitation have limited engagement in programmes by prisoners, partially because their own performance, progress and potential for release may remain hidden from them.
The impact on prisoners’ perceptions of aspects of prison life, such as fairness and openness (Crewe, 2009; 2011), may also be a contributing factor in determining their willingness to report (or not) sexual assaults. Application of the concept of the ‘imaginary’ to initial responses to reports of sexual assaults provides a useful framework with which to examine the approaches adopted by prisons and their staff in England and Wales for the period 2004 to 2014. The rationale for prisoner reporting of sexual assaults, including recording them in centralised systems such as the IRS, needs to be examined more fully in the context of managerialism. Jamie Bennett's work (2016) provides this context.
Bennett's empirical work on the contemporary working lives of CMs 1 considers their work in the wider political, economic and social context within which they worked. He cites the managerialist era of the 1990s as an influential factor in ‘moulding [managers] into more corporately aligned, self-regulating and economically rational actors’ (2016: 10). Bennett notes the ‘proliferation of the technologies and techniques of target-setting’ and the ‘introduction of key performance targets and indicators, audits and ratings systems’. Such processes were seen ‘as the solution to a range of problems in prisons … increasing management control and direction turning good intentions into reality’ (2016: 7). In her capacity as Her Majesty's Chief Inspector of Prisons, Owers (2007: 10) referred to the ‘creative accounting’ in prisons in England and Wales to meet ‘acceptable targets’ and argues such practices ‘serve to disguise, rather than tackle, the real problems prisons are facing’.
Although Bennett acknowledged that there ‘was a perceived positive effect of key performance targets’, in his view they ultimately caused distortion: The first way in which this happened was that they caused myopia where managers focused on the attainment of targets without considering the longer-term impact. They also focused on these measures to the exclusion of other factors. (Bennett, 2016: 82)
Data and methods
Data
The IRS dataset comprises administrative data introduced by the government in the late 1980s, for the purposes of recording and managing information relating to deaths in custody, self-harm and assaults. The IRS has a number of fields including some about the demographic characteristics of those involved, plus a text-based description of the incident and a field classified as ‘outcomes’.
To date, published analysis based on the IRS sexual assault data has been limited to a MoJ (2015) management bulletin on sexual assaults, which provided some trend analysis of all sexual assault incidents 2 reported between 2002 and 2013 and a more recent analysis (Sondhi et al., 2018) of incident descriptions (IDs) that adopted a similar methodology. The management bulletin (MoJ, 2015) included a breakdown of the proportion of incidents occurring on each day of the week, the proportion resulting in minor and serious injuries and, notably, referrals made to the police and incidents classified as being subject to an adjudication.
The IRS is the source for official statistics published in the form of MoJ Safety in Custody Statistics bulletins from 2010. The bulletins report the trends of deaths in custody, self-harm and assaults (on staff and prisoners) in prisons in England and Wales. The MoJ (2018: 5) specified that ‘improvements in centrally held data now mean that there are consistent datasets for deaths (from 1978), self-harm (from 2004) and assaults (from 2002) from which to determine trends’. Safety in Custody Statistics bulletins include information about serious assaults, which are defined by the MoJ (2020: 6) as: … those which fall into one or more of the following categories: a sexual assault, requires detention in an outside hospital as an in-patient; requires medical treatment for concussion or internal injuries; or incurs any of the following injuries: a fracture, scald or burn, stabbing, crushing, extensive or multiple bruising, black eye, broken nose, lost or broken tooth, cuts requiring suturing, bites, temporary or permanent blindness.
As a sub-section of serious assaults, sexual assaults are defined as: A wide range of incidents from rape to inappropriate touching. Reported sexual assaults will also include incidents where there are attempts to retrieve drugs or other prohibited items that may be hidden on the victim. Whether an incident is deemed to be a sexual assault is determined by the perception of the victim. Any assault incident reported as a sexual assault is classified as a serious assault (MoJ, 2015: 2).
Methods
A series of data files were provided by NOMS in October 2014. The files included all reported and recorded prisoner-on-prisoner incidents between 2004 and 2014 from prisons in England and Wales that involved adult male prisoners, aged 18 or over. 3 Over time, the data was sifted, reconstructed and organised into two new datasets, both of which provided insights into the level and nature of reported and recorded sexual assaults. Dataset 1 developed an assessment of the individual and institutional characteristics of reporting and recording of sexual offences in adult male prisons in England and Wales. This dataset constitutes the way in which NOMS recorded the incident and its outcome. It allowed for descriptive statistical analysis of ‘incidents’ (N = 844) and ‘involvements’ (N = 2032; a grouping which includes victims, perpetrators and ‘others’). The full sample of 844 adult male prisoner-on-prisoner sexual assaults involved a total of 2032 prisoners in incidents that occurred between 2004/2005 and 2013/2014 in prisons in England and Wales. Of these, 1116 prisoners were recorded as ‘perpetrators’, 826 identified as ‘victims’ and 90 as being ‘other’, i.e. present during the reported incident, noted as witnesses or with an undetermined role in the reported incident.
Following further coding work, the creation of Dataset 2 provides additional insight into the 844 incidents with analysis of qualitative data fields/prisoner officer generated IDs. This work was required to interpret the emerging trends and patterns of incidents and involvements. Coding was applied to each ID field, which is a compulsory text description in which prison staff summarise what has been reported. This summary was either (but more rarely) written directly by the prison officer who received the initial report or by the CM who managed the prison response to the report. The latter was more likely, given that the IDs often provided an overview of several hours of activity that would have been managed by a CM. Dataset 2 constitutes the interpretation of the data. 4
Interviews were carried out to assist with the descriptive statistics and analysis of the datasets to provide some context. Interviews were conducted with four prison-service employed CMs with specific responsibility for oversight of reported sexual assaults and for recording them in the IRS. CMs are operational prison staff with line management responsibilities. Acting as ‘agents’ (Carlen 2008), CMs are a key figure in the maintenance of managerialism and the systems that underpin it (Bennett 2016) such as incident management documentation and logging onto the IRS to record and monitor all safety in custody incidents, including sexual assaults.
Interviewees were asked for their insights and experiences of the initial responses by prison and police services to adult male prisoners’ reports of prisoner-on-prisoner sexual assaults in prisons in England and Wales. The CMs worked in in the same English Category B local prison and interviews were carried out in 2019. The prison from which interviewees were identified is a busy, Victorian-built establishment with a capacity for approximately 700 prisoners. It draws prisoners directly from local courts, prisoners on remand awaiting court hearings or sentencing, prisoners categorised as Category B and serving their sentence, those housed in its Category C resettlement wing and from a small unit for young offenders from age 18. Owers (2007) described such prisons as, ‘pressurized and overcrowded’ and more like ‘transit camps’, where prisoners were less likely to finish modular courses, training and/or short-drug treatment programmes despite the efforts of many governors.
The interviews were used primarily as a method of checking interpretation of the IRS data. Although it is acknowledged that the ‘findings’ from these interviews cannot be used to draw conclusions, their value is significant in providing perceptions of elements of prison life; for example, the processes associated with reporting and recording sexual assaults and the practical problems of responding to prison-based crime. Mason's (1996) principles for checking the validity and reliability of qualitative research provided a framework for using the interviews to check the interpretation of the data. Using a number of methods to respond to the same set of research questions has been suggested as a preferred research design to improve the prospect of validity, particularly when using datasets (Mason, 1996).
Results
Analysis of the NOMS data
Dataset 2 revealed that most of the 844 incidents (75%) were simply described in the IRS as having been either ‘referred to the police’ (12%) or ‘subject to a police investigation’ (63%), with no additional explanation of the investigation or any indication of further outcomes. Analysis of the IRS data showed that there was an emphasis on the activities recorded when sexual assaults were reported by prisoners (e.g. sending victims to healthcare; bagging evidence and segregating those involved). However, medium to longer term outcomes of incidents remained obscured, not only by failures to update the recording system, but also because outcome labels were imprecise. IDs often listed activities as outcomes, such as making referrals to police. Analysis of Dataset 1 showed that in just over 100 of the 844 incidents, a ‘prosecution pending’ was recorded and a further 78 (less than 10%) were recorded as having been subject to the prison's internal disciplinary inquiry process. It is possible that some of these cases recorded as ‘referred to the police’ were later dealt with by adjudication. It may also be the case that incidents labelled as ‘prosecution pending’ were subject to attrition and did not reach court. There is no way of confirming either of these potential outcomes through the data.
In contrast to the confusion over outcomes recorded in the IRS, analysis revealed substantial prison staff activity in response to reports made by prisoners. An example from 2005 following a report of a sexual assault illustrates this point: He [the victim] was moved to the care suite and assailant to the segregation unit. Both prisoners’ clothing was bagged for police evidence. Incident reported to staffs [sic] police [the following morning] … Cell was sealed and key placed in security key safe. Police interviewed victim at 1400 h … and he decided that he no longer wished to make a complaint. Police therefore would not proceed. Offender V believed to have confided to Offender W that he had been raped by his cell mate Offender P the previous night. Offender W disclosed this to unit staff. Offender P moved to CSU. And Offender V moved to care suite. Police have been informed. All clothing bagged and tagged. Cell secured and locked off. Police Log No xxx.
Although IDs such as these, recorded by staff, listed safeguarding and evidence preservation actions, they frequently focused on listing staff actions to the detriment of detail or the nature of the sexual assault itself. Even with detailed coding, the IRS data revealed that almost one-third of incidents did not include sufficient information to identify the type of sexual assault that had taken place and these cases remained labelled as ‘unspecified sexual assaults’. Systemic deficiencies in accuracy and information about the precise nature of each sexual assault persist in the recording regimen where reports fail to consistently differentiate between types of sexual assault. The following from 2008 is an example of the ID failing to provide key information about the nature of the sexual assault or the alleged offence reported, instead referring to ‘sexual favours’ and defaulting to the focus on response activities:
Today Prisoner V alleged that another Prisoner P has been pressurising him into giving him sexual favours. This information was given to staff during an AB interview today. It appears that V has been getting tobacco from P. To repay the debt P has pressurised V into giving him sexual favours. The alleged assaults have taken place over a number of weeks with the last assault happening days ago. V has said he does not wish to make a complaint to the police. He has been located to the health care centre. P has been located in the segregation unit pending an investigation. This has been reported to the police. [National Operations Unit] have also been informed of this incident via telephone.
{Update} The police have interviewed V today. V refuses to engage with the police. The police therefore are not pursuing the matter any further.
This conscious listing of staff responses suggests an awareness of the need to reflect positive activity in the absence of real outcomes. This theme was explored further in the in-depth interviews with the CMs.
Interviews
Lack of outcomes in-prison sexual assaults may also be driven by the contexts in which investigations take place. Issues faced by prison staff relating to the complexity of conducting initial enquiries and developing criminal investigations in institutions, which strongly rely on the efficient running of a regime (Crewe, 2009; Goffman, 1961; Liebling et al., 2012; Sykes, 1958), cannot be underestimated. Any interruptions to the prison regime present the potential for significant delays.
When explaining some of the difficulties of carrying out initial criminal investigations in prisons, one experienced CM used a physical assault to illustrate the pressures in a prison environment: Time is a major problem here because, I remember once, we had an actual assault take place in the servery where a bloke was punched and had his eye socket and jaw broken and there was blood everywhere, literally. We managed to get him off to hospital and before we could do anything, the cleaners had started to clean all the blood up. CM #03
When describing the actions that should be taken following a report of a sexual assault, one CM captured the ‘craft’ (Fleming and Rhodes, 2018) associated with responding to such incidents and the complexities of formalising a set process: We have a duty of care; we have got to be very careful. Very sneaky if you like, about how we deal with it, so the perpetrator doesn’t get to know, while they’re still in this jail that they’ve been snitched on … [The] victim, doesn’t want him to know because … [now they are] … his target or known as being a liar or a snitch… [So] you have all of those impact factors that kind of make it hard to do protocol. It doesn’t work, you could write it down, but it doesn’t work like that. CM #05
The challenges of realising outcomes in prison-based sexual assaults were well-articulated by all four CMs. When asked about the prospect of criminal justice-based outcomes, one CM noted the intricacies associated with case building. CM #02 demonstrated that he was in possession of conscious knowledge of the unrealisable prospects of achieving formal criminal justice-based outcomes in sexual assault allegations. The factors outlined in the quote below echo Carlen's (2008) assertions that staff were conscious that outcomes were unrealisable and yet carried on delivering activities, defined by prison managers and policies: From when it happens to getting it there, the prisoner, the victim, can change his mind. He goes in and out of ‘do I want to go forward with this, do I not want to go forward with this?’ Historically, we are not very good with dealing with evidence, in terms of gathering it, securing it, keeping it all together. We're very good at losing things, you know, things go a bit AWOL. I don’t think you get any convictions. I don’t actually know the stats … I suppose a lot of it would be dealt with in-house? CM #02
One CM, with 29 years’ experience in the prison service, when asked about criminal justice outcomes in reported sexual assaults cases, noted that: I haven’t seen [criminal justice outcome] one in 29 years … but I’m optimistic. No, I haven’t. I have not seen or been involved … personally in one. I may have given information that has dealt with one, but I’ve never been informed of a positive outcome. So, I’m not saying it hasn’t happened. CM #04
The day-to-day pressure of delivering the prison regime was described by another CM who captured the multifaceted role of prison officers, resonating with Crawley and Crawley's (2012) observation that the role of prison officers incorporated security and use of force beside the creation of social order in an intimate and primarily ‘domestic’ sphere: … generally, as a prison officer, we deliver the regime to the prisoners. We unlock them, give them their meds, their association, their telephone calls, facilitate their visits. We are like authoritarians when that is needed because they are not being compliant. We are like social workers to them because we have to listen to their issues. We are kind of everything from the authoritarian to the care worker … and everything in between. We have to adapt, if you like, to all these different spectrums of characters that come into the system. CM #05
CM #05 resonated with Goffman's (1961) observations that staff not only carried out the rationalised requirements of the institution, but that these often involved contradictory tasks such as, giving the ‘impression’ of humane standards while forcing obedience with the system.
Lack of time was identified by all four CMs as a reason for not seeking information about the case progression of sexual assaults. Staff obtained updates on incidents only when they proactively showed an interest by asking the CM responsible for police liaison. It is possible (and speculative) that staff lack of interest and knowledge about the outcomes of incidents was due to a tacit acceptance that outcomes were unrealisable and an expectation that little beyond initial actions would have taken place. Written or submitted updates to incidents do not appear to have been a routine aspect of the management of the IRS. Generally, once initial activity was completed and a record entered into databases, including the IRS, CMs mostly regarded their role as being complete and the IDs of staff activity support this contention. Apart from ongoing safeguarding or health issues, they viewed responsibility for the sexual assault as residing with the police once a referral had been made to them. Outstanding initial response activities, such as identifying witnesses, being a witness of first complaint or seeking intelligence through prisoner contacts and networks, were either not recorded on the IRS, were passed onto the police (‘somehow’) or did not happen. Given that 794 of the 844 incidents were reported directly to a prison officer and, of these, 746 were made directly by the victim, there are indications that opportunities may have been lost or not recorded. In a non-prison environment, information ascertained by the person to whom the complaint was initially made, or the ‘witness of first complaint’, form an important aspect of a criminal case because they can comment, not only on the description of the incident, but also on the demeanour of the complainant. In prisons, officers are likely to be witnesses of first complaint when allegations are made. However, only 49 of the 844 IDs referred to a victim's demeanour when making the report.
Lack of updates in the IRS suggest that referral to the police, which frequently resulted in no further action, effectively ‘closed’ the incident on the IRS. One CM, when asked if he found out about the progress or outcomes of sexual assault allegations said: No, not really. Unless I have gone out of my way to ask, I suppose. Everyone is busy so there's no time to go back and check out something. CM #02
Recording of ‘outcomes’ as being the ‘conclusion’ of cases in the IRS, has masked a structural unwillingness to address the core question of whether sexual assault investigations in prisons mostly result in no real outcome. Resources and efforts, including detailed accounts of procedures and practice, were outlined by prison staff who also acknowledged the lack of formal outcomes. For Carlen (2008: 19), a distinctive element of imaginary penalities was the focus on procedures rather than outcomes, which she argued were defined by having ‘a teleological and closed focus on inputs instead of a concern about outputs and effective measurement of objectives’. This article suggests the reporting data collected by the prison service in England and Wales itself predominantly focuses on activities rather than outcomes, much of which can be described as practice without prospect.
Conclusion
One of the most striking findings from the analysis of the quantitative data were the challenges in determining ‘outcomes’ from reported sexual assaults. The commonly accepted and continued practice of recording activities and processes (such as ‘referrals made to the police’) as outcomes raised concerns about recording standards and practices (Bennett, 2016). A low number of incidents resulted in formal outcomes such as adjudications and pending prosecutions. However, the final outcomes remained unrecorded in the IRS. The outcome codes used in the IRS (‘referred to the police’, ‘subject to a police investigation’, ‘adjudication’, ‘prosecution pending’ and ‘no recorded outcome’) reveal uncertainty about the real outcomes and perhaps a lack of candour about recording standards required for measuring realisable outcomes. Problems in IRS data recording were acknowledged indirectly by the MoJ's (2015: 2) relatively recent and ad hoc analysis of sexual assaults in prisons, which pointed out that beyond counts of incidents, the remaining IRS data was ‘not deemed to be of sufficient quality to be published as National Statistics’. Acknowledging the systemic weaknesses of the IRS, analysis of Dataset 1 revealed that real outcomes were rare. Analysis of IDs from Dataset 2 also suggested that outcomes recorded as having been ‘referred to the police’ or ‘subject to a police investigation’ were often recorded in the same system (in the IDs) as having been dropped or processed as ‘no further action’. IDs recorded a wide range of reasons for cases being discontinued, which included the victim's wish to withdraw the complaint or because there was a lack of evidence.
Diffusion of outcomes in-prison sexual assaults may also be driven by the complex contexts in which investigations must take place, not dissimilar to Carlen's (2008) acknowledgement of the hopelessness of realising desistance from ‘imaginary programmes’ which were essentially structured to fail. In the absence of stated outcomes, IDs written by prison staff and inputted to the IRS by CMs extend our understanding of activities or imagined outcomes that were recorded in their absence. These short descriptions provided evidence of staff responses to sexual assaults rather than providing good quality insights into the precise nature of the assault itself. The detail of these imagined outcomes revealed that they mostly consisted of temporary fixes, which combined, represent a range of practices without prospect. Returning to Carlen (2008), a pragmatic interpretation of the data analysis suggests that sexual assaults, and their unrealisable outcomes, exist as an open secret among prison staff and governance hierarchies. Prison staff evidenced their commitment to caring for prisoners by conducting a range of activities in response to sexual assaults, all perhaps in the absence of visible strategic support, clear guidance and relevant training. The recording processes and the IRS itself, are reminiscent of Goffman's (1961: 73) observation that work in total institutions was ‘followed by a paper shadow showing what has been done by whom, what is to be done, and who last had responsibility for it’. Information has been recorded without regard for ‘outcomes’, and instead, persistent efforts have been made to present a culture in which imaginary value is placed on recording imaginary outcomes. Aside from the administrative data analysis, opportunities that allow prisoners to speak candidly about their experiences have been lost (Stevens, 2019).
Goffman's (1961) concept that staff activity took place to endorse institutional aims is impossible to confirm on the basis of the data examined from the IRS. The IRS data was recorded for management purposes rather than for research and does not provide such information. Although the data suggests a trend of high levels of recorded staff activity (e.g. the preservation of potential forensic evidence), it is not possible to extrapolate the reasons for this continued practice without confidence in outcomes. In acknowledging the difficulty in identifying clear outcomes from the data, it should also be recognised that outcomes of any kind may not necessarily support and protect victims from various forms of harm.
As Chief Inspector of Prisons, Anne Owers reminded us that prisons were ‘hidden from the public gaze’. In 2007, she concluded that prisons were also ‘places that can and do easily become self-referential, lacking the external checks and balances that make institutions ask difficult questions, rather than revert to a default setting of institutional convenience’. This current research indicates that the recording and reporting of sexual assaults in England and Wales and the administrative tasks associated with that activity remain a predominantly ‘institutional priority’. We cannot know what wider approaches are taken in individual prisons to advance prisoner safety. This data does not report on that. Until we refocus the public gaze on prisons and importantly, the sexual assault of prisoners, much of the work done by prison staff in this area may continue to result in practice without prospect.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
