Rape investigation and attrition in acquaintance, domestic violence and historical rape cases

This paper looks at the progression of rape cases through the criminal justice system, from report to court, exploring the different attrition trajectories for cases that can be characterized as involving acquaintance, intimate domestic violence, and historical child sexual abuse contexts. Using police data from three police forces in England covering 87 cases, interviews relating to 15 victim/survivors and interviews with criminal justice professional, the paper explores investigative processes and victim engagement across rape cases and their different trajectories through the criminal justice system. The legal and extra-legal factors that have been identified in the previous literature as increasing attrition, such as relationship between victim and offender, vulnerability of victim, and evidential issues, were all seen to play some part in the attrition and trajectories of the cases discussed here, but were relevant to different degrees depending on whether the cases involved acquaintance, intimate domestic violence, or historical child sexual abuse. If we are to more effectively deal with the “justice gap” that exists in rape cases, an important aspect is to understand the differences between these groups of cases and the particular circumstance and needs of the victims in these different contexts.

included in categories involving alleged offender known to victim or excluded altogether as involving under-16s (Spohn & Tellis, 2012;Lea, Lanvers, & Shaw, 2003). The article explores and compares these separate types of cases in order to increase our understanding of attrition.
The issue of attrition in rape cases, or "drop-out" from the criminal justice system, has been highlighted by researchers as a concern for many years. Since the 1970s, it has been apparent that rape cases are less likely to progress from reporting to prosecution and conviction than other types of offenses (e.g., Wright, 1984;Grace, Lloyd, & Smith, 1992;Gregory & Lees, 1996;Kelly, Lovett, & Regan, 2005), and attempts have been made to reduce attrition, in particular through changing police attitudes and practice. For instance, police circulars (25/1983 and 69/1986) were issued during the 1980s to shift practice regarding police treatment of victims and the criming of rape cases, although without much impact at that time (Lea et al., 2003). Since the mid-2000s, a more concerted process of government review regarding criminal justice approaches to sexual offenses has involved a series of critical reports and inspections, again highlighting the large attrition in rape cases and the need for a more "victim-focused" approach (HMIC/HMCPSI, 2012). In the past decade, there have been efforts by the police, the Crown Prosecution Service (CPS) and the courts to improve their responses to the investigation, prosecution, and conviction of rape offenses through training, better recording and provision of information, support and anonymity for victims, and monitoring of files. A Victims' Code and government support for rape crisis centers has been part of this attempt to improve victim/survivor treatment and outcomes. The Stern (2010) review into rape cases in England and Wales suggested that ISVAs are the most effective, cost-effective, and affordable example of a reform to a system, making an enormous difference to how victims feel about what is happening to them as they process through the criminal justice system. Limited evidence from CPS data suggests that ISVAs may reduce the number of retractions by victims (Brown et al., 2010). Hohl and Stanko (2015) also found that support, by a sexual assault referral center, halved the odds of withdrawal by victims.
Rape in adults is a non-consensual offense, involving intentional penetration of the vagina, mouth, or anus of another person with his penis, where that person does not consent to the penetration and where the alleged perpetrator does not reasonably believe that the person consented (see 2003 Sexual Offences Act). Rape and other sexual offenses against children, involves ostensible consent as those under the age of legal consent (16 years), cannot be seen to consent regardless of whether it appears that they gave consent. Absence of consent can be also assumed in certain situations involving adults, such as violence or threat of violence, where the victim is unconscious or has a disability that limits capacity to consent. Rape is a gender-related crime where only men can carry out the offense, although both women and men, adults, or children can be raped. The rape cases being looked at in the current research mainly come under these legal definitions of consent, within the 2003 Sexual Offences Act. In some of the historical child sexual abuse cases, from before 2003, earlier legislation was applied (Sexual Offences Act 1956; Indecency with Children Act 1960).
Home Office figures have, since 2003, shown an increase nationally in the number of rapes recorded as crimes and an increase in the actual number of convictions. While only 15% of victims report serious sexual offenses to the police (MoJ, HO & ONS, 2013), there appears to be increasing willingness to report such crimes, although increases also reflect improvements in recording (ONS, 2015). In the 3 years to the beginning of 2012, there was a 26% increase in the number of rapes recorded by the police (HMIC/HMCPSI, 2012). Fifty-eight percent of rape prosecutions in 2008 and 2009 resulted in a conviction for rape or another offense (Stern, 2010), increasing to 63% in 2012-2013. In 2013/2014, the sexual offenses recorded by the police in England and Wales rose by 32%. Increase in reporting was also evident in the location of the current research. Nonetheless, there has been a continual decrease in the overall proportion of such reported crimes resulting in conviction, leading a succession of inspections, including Without Consent (HMIC/HMCPSI, 2007) and Forging the Links (HMIC/HMCPSI, 2012), to conclude that the justice gap for victims of rape has been widening. Stern suggests that policy has generally been viewed as positive and adequate, but the justice gap and high attrition rate have resulted from failures 'in the implementation', with possible shortcomings in the working of police and CPS (Stern, 2010: 9). Others have pointed to problems regarding the courts (Smith & Skinner, 2012). Jordan points out that the problems may indeed be deeper and possibly structural: 'Reviews come and reviews go and women are still raped' (Jordan, 2011, p. 245) and cautions against "accepting the rhetoric of reform while the underlying realities remain little changed" (page 234).
Previous studies have highlighted two main areas related to attrition: the points in the criminal justice process where attrition is likely and the factors that make some cases more subject to attrition. There is general agreement that most attrition occurs at the policing stage, where decisions are being made whether an offense appears to have been committed and if the case should be referred to the CPS for charging. The next attrition point occurs where the CPS decide whether or not to prosecute; and the final point of attrition is at the court stage if the outcome is not a conviction (Gregory & Lees, 1996;McMillan, 2010;Spohn & Tellis, 2012). Both legal and extra-legal factors have been identified as increasing attrition, such as rape myths, relationship between victim and suspect, vulnerability of victim, and evidential issues. Factors affecting attrition have been identified as the age of the victim and degree of violence used (Harris & Grace, 1999), vulnerabilities such as mental health or learning disabilities (Kelly et al., 2005;Hester, 2015) and evidential issues such as previous false allegation by the victim, apparent discrepancies in victim statements, and evidence or police opinion casting doubt on the allegation (Hester, 2015;Hohl & Stanko, 2015, p325). Stern (2010) highlighted that sections of the public may hold the perception that women often make false allegations of rape, and this may also affect the way rape complaints are dealt with by police, prosecutors, and juries. Lea et al. (2003) in their study of attrition in South West England found that a high proportion of cases, one in five, were deemed false allegations by the police. Yet a report for the Director of Public Prosecution (Levitt & CPSEDU, 2013), covering all CPS areas, indicates that less than 1% of rape allegations (38 out of 5,651 prosecutions) were for making false allegations of rape or rape and domestic violence.
Rape myths and stereotypes have been highlighted, since the 1970s, as an important feature in attrition and continue to be evident to some degree (Gregory & Lees, 1996;Hohl & Stanko, 2015). Such myths and stereotypes are often linked to ideas about what is deemed appropriate female behavior, drinking, and clothing. This includes the myth that "real rape" is carried out by a stranger, involves force, and active resistance. As a consequence, rape by a stranger has in the past been more likely to result in arrest and conviction than those involving acquaintances. However, Hohl and Stanko (2015) found that in the context of the London Metropolitan police, stranger cases are now highly likely to suffer attrition because the alleged offender cannot readily be identified.
In actuality, most rapes are by men known to the victim/survivor, which may involve acquaintances, intimate partners, or family members. While the relationship between the victim/survivor and suspect (stranger vs. nonstranger) has emerged as a feature in studies of attrition, Spohn and Tellis (2012), in their overview of criminal justice responses to rape, suggest there is a particular dearth of research into "intimate" contexts for rape, where the alleged offender is a partner or ex-partner (and see McOrmond-Plummer, Easteal, & Levy-Peck, 2014). British crime survey data has indicated that intimate partner rape constitutes possibly half of all rapes, and the proportion of this type of rape reported to the police has been increasing (Kelly et al., 2005;Brown et al., 2010). However, in research on attrition, rape from intimate partners has often been hidden within the category of "known to victim" without much explicit focus. Gregory and Lees (1996) did disaggregate intimate partner rape (involving ex-husband, ex-boyfriend, ex-cohabitee), which constituted 17% of the rapes in their London police sample, and found it was the category with the greatest attrition: two thirds of the cases were not deemed crimes by the police, and none proceeded to prosecution or conviction. Harris and Grace (1999) found that rapes involving intimate partners were most likely to end in no further action, while acquaintance rapes were most likely not to be deemed crimes in the first place. Lea et al. (2003), in their research in South West England, talk about "intimate rapes" (page 593), appearing to define this as having been in an intimate relationship with the accused but without detail or specific exploration of this group of cases.
In recent years, the "Saville effect" (following the disclosure by numerous victims of sexual assault and rape by Jimmy Saville) has markedly increased the number of cases of historical child sexual abuse being reported to the police. Kelly et al. (2005), writing before this recent phenomenon, found that cases involving rape of under-16s were more likely to result in prosecution, but it is not clear how many of these were historical cases brought by adult survivors. Studies of attrition more generally have not tended to include much discussion of historical child sexual abuse cases or have included them within the wider category of alleged offender known to victim.
Previous thematic analysis of rape data by the authors indicated differences in attrition between acquaintance, intimate partner, and historical rape cases (Hester, 2015). Given the gaps in the literature and our own findings, we decided for the current article to disaggregate rape cases involving acquaintance, intimate partner, or historical child sexual abuse in order to examine in more detail possible differences in their attrition trajectories. Given the policy focus on support to victims as a means of mitigating attrition, we included analysis of police referral data and contextualized the different types of cases using victim/survivor interview data.

| METHOD
The research involved quantitative and qualitative analysis of data held by the police and CPS on 87 rape cases, interviews with criminal justice professionals, and interviews with a small sample of victim/survivors who had accessed the criminal justice system (Hester, 2015;Hester & Lilley, 2015). The research covers the period of 2010-2015.

| Police data
Data on rape cases reported to three police force areas during the months of May and November 2010 were accessed in the respective police headquarters and anonymized by removing all names and other identifiers. In two of the police force areas, access was provided to the police database, working alongside a police staff member, and it was possible to examine all cases from when they were recorded on the initial police log, through crime reports and information on outcomes held on the police national computer. In the third area, only police summaries of cases were made available; some were not readily available, and it was deemed too costly by the police for the files to be recalled, although a member of staff was available to answer more detailed questions about cases. In all three areas, only closed cases were examined, and as the research progressed, it was possible to include cases where outcomes had initially been pending. Where cases had proceeded to prosecution, the CPS nationally gave permission for case files to be accessed.
CPS files were made available in the regional CPS headquarters, and anonymized data were compiled. This resulted in police data on 87 rape cases, with additional CPS and court-related data on 17 of the 87 cases that proceeded to court. Individual cases took up to 23 months to progress through the criminal justice system and were tracked over that time.
The analysis involved coding of the incident features and narratives, to identify institutional and other factors as well as general themes and further categories. Quantitative data regarding demographics of victim and accused (e.g., gender and relationship status), vulnerabilities (e.g., mental health and drunk), relationship between victim and accused (e.g., acquaintance, ex/partner, family member, stranger, and multiple perpetrators), criminal justice progression and outcomes (e.g., deemed a crime, arrest, charges, and convictions), and referrals (a range of sexual violence and other agencies) for the 87 cases were loaded on to an SPSS database for analysis, and Pearson's chi-square and logistic regressions were used where possible to test significance of associations. In this paper, we report on the logistic regressions for outcomes (crimed, arrest, charged, conviction), features pertaining to the cases (forensic examination at SARC, DNA evidence, victim withdrawal), and vulnerabilities (mental health issues, drunk victim), comparing the three case types (acquaintance, domestic violence, historical; Table 1). A limitation was that the quality of the police data did not allow further items to be analyzed using regression. It was possible to provide indicative data on police referral of victims/survivors to other services (Table 2), but referrals may not always have been recorded. Qualitative police data were coded thematically, and the data relating to individual cases were added to qualitative framework grids (Ritchie & Lewis, 2003) so that cases could be analyzed more fully, and case "stories" compiled.

| Interviews
In-depth interviews were carried out with criminal justice professionals in the three police force areas, including three police officers in charge of rape investigation and a crown prosecutor with rape remit, to ascertain their views on different types of cases and attrition/progression.
In-depth, face to face interviews were also conducted with a convenience sample of 15 rape victims/survivors, two of whom had family members present. All were clients of specialist sexual violence services in one of the force areas, and all had accessed the criminal justice system. The aim was to explore victim/survivor experiences and engagement with criminal justice agencies and specialist sexual violence services. Victims/survivors were identified and contacted via a letter from specialist sexual violence services to ask if they would be willing to participate in the study via an interview. Where consent was given, the services provided the research team with contact details, and a researcher contacted the ex−/clients directly by telephone to explain the research and arrange an interview. Of the 15 victim/survivors interviewed, nine had experienced childhood sexual abuse, three rape in a domestically abusive relationship, and another three experienced acquaintance rape/indecent assault. All but two cases ended up in court, and 11 of these 13 cases resulted in conviction. All of the victims/survivors in the 13 court cases were supported by ISVAs. The overemphasis on historical cases in the sample is a limitation, as is the location of the sample in only one of the police force areas. The interview data does, however, provide detail regarding victim/survivor experience of the criminal justice system, attrition, and the relevance of wider support.
Interviews lasted between 1 to 2 hours. All interviews were digitally recorded (with consent) and transcribed verbatim prior to analysis. Analysis involved reading and re-reading transcripts to identify themes, using grids to record summaries from the thematic analysis and linked quotes from participants. The main themes included type of case (historical child sexual abuse, domestic violence, acquaintance); disclosure; police involvement; CPS and court involvement; support from ISVA, independent domestic violence advisor (IDVA), and/or other services/agencies; impacts; and any other important features (author).
The research was given ethical approval from the University of Bristol Ethics Committee and CPS nationally.

| RESULTS
In what follows, we begin by providing a brief overview of the quantitative findings from the 87 police file cases and a brief overview of the interview data. We then go on to examine in more detail the findings from the police data and victim interviews regarding the acquaintance, domestic violence, and historical rape cases.

| Quantitative findings
The 87 rape cases reported to the police involved 98 victims (94 women, four men) and 97 accused (all men). Ages of victims/survivors at the time of reporting ranged from 16 to 57 years and for the accused from 17 to 85 years. In two cases, the accused were described as group of "Asian men", and Black French and Nigerian, but ethnicity was generally missing in the police record.
Of the 87 rape cases reported to the police, between half and two thirds were initially deemed crimes by the police (53/87, 61%), but following further investigation, nine of these were reclassified by the police as "no crime" (9/53, 17%). Forty-five cases resulted in arrest (44/87, 51%), although only about half of these were referred to the CPS for charging decisions (23/87, 26%). The CPS consequently applied charges in three quarters of these (17/23, 74%; Table 1), amounting to about one in five (19%) for the whole sample. One of the historical domestic abuse rape cases dropped out at this stage seemingly as the victim/survivor did not want to attend court, leaving 16 cases that proceeded to trial in a Crown Court. All the accused initially denied and pleaded not guilty to sexual and other offenses, resulting in victim/survivors facing questioning by defense barristers in court, and all but four of the accused continued to deny they had done anything wrong. Ten cases (10/87, 11%; Table 1) resulted in conviction. and least likely to do so in acquaintance rapes (4/17, 23.5%). Historical cases were significantly more likely to result in conviction (6/10, 60.0%, p < .05).
With regard to case investigation and progression factors, victims in acquaintance rapes were significantly more likely to be referred to a SARC for forensic examination (p < .05), and those in historical cases were significantly less likely to withdraw from the case (p < .05). Only in a few acquaintance cases was use of DNA evidence mentioned (Table 1).
With regard to vulnerability, victims/survivors of acquaintance rapes were significantly more likely to be recorded as drunk (p < .005). Victims of historical child sexual abuse were significantly more likely to be recorded as having mental health problems (p < .005; Table 1).

| Interviews
Interviews with the police indicated that they saw the three types of rape cases, acquaintance, domestic violence, and historical, as quite different types, and approached investigation of them in different ways with regard to the use of forensic information, evidence from victim records, use of other evidence such as closed-circuit television (CCTV), and witness statements. Acquaintance cases were described as often reported close to the time of the event thus enabling collection of forensic material (DNA and body fluids) and also tended to be located in public places where there might be CCTV evidence and possible witnesses. Rape by intimate partners was deemed a more hidden crime, usually taking place in private settings and without other witnesses, and victim statements were thus very important in such cases possibly combined with forensic medical evidence. For historical child sexual abuse cases, evidence from the victim's school and medical records and interviews with a range of familial and extra-familial witnesses were seen as important, as was the possibility of evidence from further victims/survivors.
The most important issue for the 15 rape victims/survivors interviewed was that they were believed by the police (and other services). Some had had bad experiences, especially some years ago, of disclosing rapes to the police and their allegations not being taken seriously or without the accused being prosecuted. Generally, the support needs of the victims/survivors interviewed tended to differ when they disclosed the abuse, if/when the abuse was reported to the police, if the case proceeded to court, and after any involvement with the criminal justice system (CJS). Their needs were also in some ways linked to "type" of sexual violence experiences, with experience of historical childhood sexual abuse, or rape in a domestic violence relationship requiring quite different types of support. Specialist services were seen as providing the changing mixture of targeted support as and when the victims/survivors' needs changed, such as increase in counseling support when they were feeling more depressed/suicidal, and ISVAs or IDVAs to support them through the often drawn out criminal justice process (for more detail, see Hester & Lilley, 2015).

| Case types
3.3.1 | Acquaintance rapes More than a third of the 87 cases involved acquaintance rapes, where the victims/survivors had known the accused to some extent, perhaps as a friend or acquaintance or had met them in the past 24 hr, but where the accused were not partners or family members. Two were rapes by strangers. Echoing the police interviews, acquaintance cases were usually very recent cases, and it was therefore possible to carry out forensic tests (half were recorded as attending a SARC for forensic and other investigation). As indicated earlier, these cases had the smallest number resulting in charges, and there was only one conviction (and two acquittals). Reflecting the particular lack of progression through the CJS of these cases, nearly a third of victims/survivors were instead referred by the police to counseling or mental health services.
As indicated earlier, victims in acquaintance cases were significantly more likely to be recorded by the police as drunk, and many of the acquaintance cases involved victims who had been drinking and did not always know whether anything onto ward had taken place. Half were with perpetrators known less than 24 hr, and these were slightly less likely to be seen as a crime, although more initially resulted in arrest. The police were often careful to investigate (via CCTV, forensics, interviews) whether an incident where the victim was too drunk to know what had taken place had actually taken place and/or been consensual. One instance involved a very drunk woman who thought she had been raped by two men she met in a bar. The police arrested and bailed both men, but in the subsequent police investigation, CCTV evidence and interviews with others who had been present indicated that the events were not as the woman remembered. Following discussion about the lack of evidence, the woman retracted the complaint.
In one of the force areas in particular, the police were seen to pursue a number of avenues to investigate whether there was evidence of rape. For instance, in one case, a woman reported that she had been raped after drinking a drug spiked drink and was referred to the SARC for forensic medical examination. The police report indicates that the forensic results did not show any sexual activity and that they consequently saw no evidential value in also examining urine/blood for drugs. The suspect was interviewed but denied vaginal intercourse and also denied using or giving drugs to the victim. The police decided not to progress the investigation further as sexual intercourse could not be proved and changed their initial criming of the case to "no crime". The police note however that the initial complaint was not malicious in any way, as it was a case of the apparent victim not being sure what had happened.
In other instances, and force areas, the police appeared reluctant to investigate where the apparent victims were drunk and knew the accused, even if questions might be asked regarding the vulnerability of a very drunk victim and possible lack of consent. For instance, a woman contacted the police to say that a man she had known for a number of years had stayed at her house after they had been drinking. She woke to find him having sex with her without her consent. The police removed clothes from the scene but did not pursue forensics. She later provided a statement that she was not reporting any offenses, but had rung the police for advice as she would not have had sex if she had not been drunk. The police decided not to crime the case.
One case where the rape of a very drunk victim resulted in conviction involved a stranger rape. This instance took place at a party, where the teenage victim became extremely drunk and was raped while asleep by a man whom she had not met before and had not talked to. Police investigation revealed that he had seemingly targeted the victim because he had seen her go "upstairs" and because she was very drunk. Her friends realized what had happened and urged her to report to the police. Although the victim could not remember anything or give a consistent story to the police, her friends provided statements. Vaginal swabs and underwear samples provided evidence of DNA from the suspect. Two suspects were initially arrested, and both denied any sexual contact with the victim.
However, the DNA evidence eventually allowed the police to identify the main suspect, and he pleaded guilty to rape in court. This was the only case in our entire sample where the rape of a very drunk victim resulted in conviction.
In one of the two cases that the police recorded as a "false allegation," two young women said that six Asian men had taken them to a derelict flat and that they had been raped by one or two men each. The police began their investigation, obtained medical evidence from both women and interviews. However, the women had placed information on Facebook and in emails that indicated that the sex had been consensual and no crime had occurred. As the case involved young women (aged 17 years) and sex with multiple partners, questions may perhaps be asked as to whether the women were subject to a wider context of exploitation and sexual grooming. Following the current research, the police carried out further checks to ensure that the case was indeed a fabrication. One of the stranger rapes was also deemed by the police to be a fabrication, where, following some investigation and arrest of the suspect, the victim indicated that she had made up a classic rape story (stranger rape in alleyway) so that her partner would not realize that she had had sex with another man.
Interviews with victims/survivors indicated the complex nature of some of the acquaintance cases and the importance of specialized support alongside the criminal justice intervention to mitigate attrition. Clare and Kay, for instance, were both raped as teenagers by male acquaintances who lived in the vicinity. They talked about the way the police linked them up with sexual violence and specialist support services and how important this was for their cases to proceed successfully through the criminal justice system. Kay was threatened by the accused's family when she reported the rape to the police. The police supported her through the criminal justice process, and a women's support service provided "target hardening" (braces on the doors) to allay her fears of retaliation. When Clare reported to the police, she was also threatened, by the mates of the accused, and had to move house. The police referred Clare to a specialist sexual violence service, which she found very useful in providing practical advice with housing and other support. An ISVA also helped her through the court process, which she found essential: "I don't think it would 'ave ended up going to Court without them. Or I'd a' gone to Court but I'd a' probably ended up being an absolute wreck" (Clare).

| Intimate partner domestic violence rape
The domestic violence rapes were by current partners or ex-partners and included two "historical cases" where the rapes took place a number of years prior to reporting to the police. As reflected in the risk assessment protocols used by the police when attending domestic abuse situations (www.dashriskchecklist.co.uk), domestic violence rapes were deemed some of the most serious cases by the police and resulted in the greatest proportion of arrests. However, these were also most likely to be withdrawn in the early stages of the criminal justice process due to fear and threats faced by the victims from the perpetrators. Thus, as indicated earlier, only one in five resulted in charges, with five cases proceeding to court of which three resulted in conviction. In one of the cases resulting in acquittal, the judge applied a restraining order for life on the accused. The cases referred to SARCs were for forensic examination and also to obtain further support. Reflecting that these were high-risk domestic abuse cases, one in five was referred to MARAC and a couple of the victims/survivors referred to refuges where they wanted to get away from the violent partners. The police referrals to social services were seemingly related to child protection issues.
Of the three cases that made it to court and resulted in conviction, one involved an extremely violent offender who had been abusive to two separate partners, one of them some years ago. The police compiled corroborative evidence from the two unrelated victims, witness statements from various relatives, and the women's new partners as well as evidence from historical medical records. While there was no recent forensic evidence, the range of statements and documentary evidence led the CPS to charge the accused with anal and vaginal rapes and also a series of violence against the person offenses. The accused denied any of the charges, but eventually pleaded guilty during the trial to assault and threats to kill. He was convicted for those offenses as well as for vaginal and anal rape and sentenced to 6 years in prison. He was also placed on the sexual offenses register and had a 5-year restraining order imposed. However, this was the only domestic violence rape case that resulted in conviction for sexual violence offenses. In the two other cases with convictions, the sexual offenses were dropped during the trials, and only violence against the person offenses were successfully prosecuted.
In one of the domestic violence rape case, the police were keen for the case to proceed to charges for sexual offenses as well as violence against the person offenses. However, the victim did not think she was strong enough to go through a court case in relation to the sexual offenses. The police obtained a withdrawal statement from her regarding the rape allegations, but still wanted to pursue assault charges. The case was considered by the CPS, but they advised no further action as without the victim attending court to talk about the rape allegations they deemed there was insufficient evidence to provide a realistic prospect of conviction. The CPS thought a defense lawyer would bring up the fact that she had dropped the rape and very likely use it against her and deny her a fair trial.
Consequently the victim, being asked what she wanted to do following the CPS's advice, said she would rather not go ahead with any of it. Previous criminal justice system involvement was used by the police as indication both of vulnerability and of potential credibility as a witness if the case was to proceed. This was especially obvious in domestic violence rape cases.
For instance, in one such case the victim reported that she was raped a week ago by her ex-partner and had been so previously. The police had a number of incidents of rape of the victim by the same perpetrator recorded, but also of rape by another ex-partner where she had withdrawn her complaint. Following an interview with the victim, the police decided to discuss the case with the CPS 'due to history of victim' and the fact that she was vulnerable to further abuse.
The victim withdrew her complaint and provided a withdrawal statement. The case was not recorded as a crime.
Victims/survivors whom we interviewed who had experienced domestic violence rape found that a mixture of support from both ISVAs and IDVAs as well as other support was important, with specialist sexual violence support particularly important during the criminal justice process if the abuser was being tried for sexual offenses, and then domestic violence support in the longer term. For instance, Imelda was scared to tell the police about the domestic abuse and anal rape she suffered from her husband due to his extreme violence and threats. When he smashed up the house she rang her parents, who reported it to the police. The police referred Imelda to the SARC, and she was put in contact with a specialist sexual violence service. Support from a sexual violence counselor at that service enabled her to eventually engage with the police: Cos, they gave me like a few days to think about what I wanted to do and I met with [the sexual violence service] who talked through like all the different options and supporting me in making the decision to go and do the video interview… (Imelda)

| Historical child sexual abuse
The third largest group involved historical child sexual abuse, where the rape took place at least a year ago, but where the victim/survivor was an adult by the time of reporting. As indicated earlier, historical cases were (significantly) the most likely to result in charge and conviction. The case file data indicated that this was often because there was more than one victim, because victims had disclosed similar accounts to a variety of people at different times, and because medical or social services records provided supporting evidence.
As indicated, mental health issues were a significant feature in the historical cases. Reflecting the particularly detrimental impacts of child sexual abuse, about one in five of the victims/survivors in these cases were referred by the police to counseling or mental health services, some were recorded as being put in touch with ISVAs outside SARCs and some to adult services, or social services where they were teenagers. Only a quarter of the victims/survivors in historical cases were recorded as attending SARCs, reflecting the lack of recent forensic evidence available, and this was more likely a route to other support.
Individuals recorded by the police as having mental health problems tended not to have their cases progressed to charge, although the police conferred with the CPS in some instances. In one of these cases, a woman was reported to have been repeatedly raped by her brother-in-law some years previously. The police interview took place with a care coordinator present due to the victim suffering from post-traumatic stress disorder, depression, and psychosis. The police recorded that the victim made "clear disclosures" of historical rapes, and the accused was arrested. However, the CPS deemed that no further action was possible due to inconsistencies between the victim's account and accounts by other family members, and were concerned that a defense counsel would be able to capitalize on the different accounts and that family members had heard differential degrees of disclosure by the victim. While the CPS did not see a realistic possibility of conviction, the case file notes that they were "aware that there can be understandable reasons why immediate disclosures were not made." It was apparent from case files that the police did not always agree with the CPS decisions. In one instance, where a male victim reported historical child sexual abuse, the CPS said that there was not enough evidence to proceed.
However, the police queried this decision, and the file was resubmitted. Nonetheless, the case resulted in no further action, and the police subsequently recorded that it was undetected.
In one case that did result in charges, involving historical rape by the victims' brother-in-law, the CPS file indicated that the jury could not decide on a verdict, and the case was considered for retrial. However, despite support from an ISVA and the police, the victim did not want to go ahead and was concerned that she had not been believed. She did not want to go through it all again. The police spoke with her and consequently recommended that the case be withdrawn.
The victims/survivors we interviewed who reported historical child sexual abuse had mixed experiences of the police investigations and criminal justice involvement. For instance, Laura, who experienced abuse as a child from a family member, found the police involvement quite negative, as, following initial interview, they did not keep her informed about what was happening during the 2 years the case was in the criminal justice process. The CPS eventually decided to take no further action, but the police did not tell Laura why that was the case. The length of the process ripped her family apart and led to Laura being bullied at school. Given the high level of mental health problems for this group of victims/survivors, support was particularly crucial to ensure engagement with and progression of cases through the criminal justice system. Victims/survivors often required support from specialist sexual violence services before they were able to report to the police. One of the male victims/survivors we interviewed had experienced sexual abuse in his early teens. He had tried to report to the police at the time, but did not get a positive response. He eventually came across his abuser again 30 years after the abuse and reported to the police again, this time with positive response, although the renewed focus on his experiences intensified his mental health problems. The police referred him immediately to the specialist sexual violence service, which provided counseling that he thought was brilliant. He was eventually helped through the court process by the ISVA based in the service.

| DISCUSSION
The research echoes previous findings (McMillan, 2010;Spohn & Tellis, 2012), that most attrition occurs at the policing stage, and it was apparent that the police would pursue a range of evidence to ascertain whether or not a crime had been committed. There were also instances where the police considered the evidence strong enough to proceed but where attrition occurred at the charging stage. CPS decisions to take no further action took into account the nature of the evidence available and also presumptions of how the victim might appear to a jury in court (victim credibility), which meant that vulnerability factors such as mental health issues or previous history of abuse tended to increase attrition (see Kelly et al., 2005). The rate from charge to conviction was the same as the national figures reported in the Stern Review (2010), and higher than the 3-year average reported by the Ministry of Justice, Home Office & Office for National Statistics (2013), although none of the national figures reflect individual case progression.
The research echoes previous studies in finding that factors contributing to an increase in attrition included vulnerability of victims, evidential issues, and the relationship between victim and suspect (Kelly et al., 2005;Hohl & Stanko, 2015). Victim/survivors with mental health issues or whom the police deemed drunk were least likely to have their cases proceed. Compared to earlier studies, however, there was some indication of the police taking more seriously allegations in contexts with a vulnerable victim. For instance, attempting to persuade the CPS to charge precisely because a victim was vulnerable and also pursuing evidence in contexts with a drunk victim on the basis that such an individual could by default not have consented. The police appeared to pursue a range of evidence to establish whether a crime had occurred, including CCTV and forensics as well as witness statements, with attrition resulting in some cases as there was not the requisite evidence for a crime to have occurred. However, resource constraints also appeared to affect the type and range of evidence collected in some instances with possible impacts on attrition.
The research echoes the findings for the Director of Public Prosecution that very few allegations are false (Levitt & CPSEDU, 2013), and compared to the findings by Lea et al. (2003) that one in five cases were deemed false by the police, in the current research there was evidence of the police taking care not to label cases as false.
The relationship between the victim and suspect was a key feature in the pattern of attrition. The type of relationship, whether the suspect was an acquaintance, was an intimate partner, or had abused the individual as under-16 in the past, led to different attrition trajectories. Echoing trends in recent studies, there were very few stranger rapes in the current study where the victims/survivors had not known the accused before and the largest category was acquaintance rapes (Lea et al., 2003;Kelly, 2010). Acquaintance rapes had the largest rate of attrition, with proportionately fewest arrests, charges, and few convictions, and included significantly more of the victims who had been drinking and were drunk. Echoing Harris and Grace's (1999) study, acquaintance rapes were most likely not to be deemed crimes in the first place. While Gregory and Lees (1996) found less than one in five cases involved an intimate partner or expartner, in the current research, more than a quarter of cases involved rape by intimate partners. This reflects the greater public awareness and policy developments since the 1990s regarding domestic violence and increase in reporting of domestic violence to the police (Woodhouse & Dempsey, 2016). In Gregory and Lees' study, intimate partner rapes had the greatest attrition with fewest cases deemed crimes by the police and none proceeding to prosecution or conviction. By contrast, the current research found that such rapes were most likely to be crimed and result in arrest, indicating the seriousness with which they are now treated, although these cases had proportionately fewest convictions as victims were highly likely to withdraw due to fear and threats from the suspect. The current research not only echoes Kelly et al. (2005) in finding that cases involving rape of under-16s were more likely to result in prosecution but also goes beyond that study making clear that the cases were historical and brought by adult survivors. Despite significantly higher levels of mental health issues for victims/survivors in historical cases, this group of cases resulted in the highest rates of charge and conviction, and support for victims from sexual violence services is likely to have played a key factor on this.

| CONCLUSION
While a number of studies have previously examined progression of individual rape cases from reporting to court outcomes in the UK, most were prior to the enactment of the Sexual Offences Act 2003 in 2004 (e.g., Gregory & Lees, 1996;Lea et al., 2003;Kelly et al., 2005), and none were concerned with the same location as the research reported in this paper (Metropolitan Police, 2007;McMillan, 2010;Hohl & Stanko, 2015;and see HMIC/HMCPSI, 2007).
However, previous research has tended not to disaggregate cases involving acquaintance, domestic violence, or historical rape. By separating these cases and also examining contextual features regarding support to victim/survivors, the current research provides new data that take further our understanding of differences between these cases and attrition.
By combining quantitative and qualitative data from analysis of criminal justice data with CJS professional and victim/survivor interview data, the research found that different types of rape cases, involving rape by acquaintances and stranger, in the context of intimate domestic violence, as historical child sexual abuse, or by family members, tend to have different attrition trajectories and involve different investigation factors. If we are to understand and possibly tackle attrition in rape cases, that is to deal with the "justice gap," it is important that we understand how these different types of cases progress and the particular circumstances and needs of the victims and contexts for the rapes that they involve. The legal and extra-legal factors that have been identified in the previous literature as increasing attrition, in particular relationship between victim and offender, vulnerability of victim, and evidential issues were all seen to play some part in the attrition and trajectories of the cases discussed here, but may be deemed to be relevant to different degrees depending on whether the cases involved acquaintance, intimate domestic violence, or historical child sexual abuse.
Although domestic violence cases were most likely to be seen as crimes, historical child sexual abuse cases were more likely to result in conviction. Acquaintance rape was least likely to result in arrest. Investigation of acquaintance rape tended to focus to a greater extent on forensic examination, including medical examination of victims, use of DNA samples from clothing, and investigation of CCTV coverage. Alcohol use was especially prevalent in these cases, and the police were at times reluctant, or unable, to investigate fully the lack of consent by a very drunk victim. Intimate domestic violence rapes were often carried out in contexts of extreme threats and other violence against the victims, who tended to be fearful of the repercussions of police and criminal justice involvement, and as a consequence withdrew their statements and support for the cases to proceed. Successful domestic violence rape cases may rely on mixtures of ISVA and IDVA support combined with other specialist domestic violence advice and support.
Investigation of historical child sexual abuse cases, where there is no possibility of recent forensic evidence, focused on a wider range of witness statements from family and others, and documentary evidence from school and medical records. Mental health impacts were especially prevalent for historical child sexual abuse victims, and successful prosecution in such cases appeared to rely heavily on victims being supported by specialist sexual violence services and counselors alongside the criminal justice response.
The Stern (2010) review into rape cases in England and Wales suggested that ISVAs are the most effective, costeffective, and affordable example of a reform to a system, making an enormous difference to how victims feel about what is happening to them as they process through the criminal justice system. While the current research did not assess effectiveness of ISVAs, it does provide some indication that the increasingly victim-focused approach proposed via government policy (Payne, 2009), and in police inspections and guidance, is being reflected to some extent in police-initiated support for victims/survivors. The police were linking into multi-agency networks, referring victims/ survivors to a range of specialist sexual violence support as well as to other social care and medical services support depending on the type of case, and these activities were deemed crucial by the victims/survivors concerned. The research thus indicates the importance of exploring different types of rape cases (acquaintance, domestic violence, and historical), but also that their progression through the criminal justice system needs to be considered within the wider context of the support provided to victim/survivors.