SA CRIME QUARTERLY No 22 • DECEMBER 2007 13 T he levels of rape and other forms of sexual assault in South Africa have been the subject of international attention and condemnation over the past ten years. The repeatedly cited dictum that refers to South Africa as the ‘rape capital of the world’ is often accompanied by the presentation of South Africa’s (often contested) rape statistics, which fluctuate between 52 500 and 54 000 per year. It has also been suggested that in South Africa, rape has one of the lowest conviction rates of all serious crimes, with research indicating that only about ten per cent of reported rapes receive guilty verdicts (SALC 1999). The Department of Justice and Constitutional Development figures also show that of more than 54 000 cases of rape reported in 1998, fewer than seven per cent were prosecuted. In contrast, South Africa is also known for its commitment to constitutional rights and protections and for engaging in progressive legal reform processes. Accepting that levels of rape ‘are high’, and are arguably an important measure of women’s safety and equality, the focus on reported rape statistics and convictions has the potential to detract policy- makers from the more substantive issues surrounding the treatment of rape victims within the criminal justice and public health care systems. Over the past decade, non-governmental organisations have worked tirelessly in advocating for systemic shifts in how rape cases are treated within these systems. The objective of these advocacy efforts has been to shift the criminal law and encourage the criminal justice process to be more responsive to the needs and experiences of rape survivors. The introduction of appropriate procedural measures to rectify the historically poor treatment of rape survivors has resulted in an exhaustive process to change the law on sexual offences and to ensure concomitant shifts in criminal justice practice in relation to the management of rape cases. The concerted focus on the Sexual Offences Bill (see the next article in this issue), for instance, has been an attempt to improve the treatment of sexual offences cases through changes in the definition of rape as well as the introduction of LOSING GROUND? Making sense of attrition in rape cases South Africa is notorious for being the ‘rape capital of the world.’ The new Sexual Offences Bill introduces a number of legal reforms intended to improve the handling of sexual offences cases. But these new reform efforts will not have the desired impact if laws are not properly implemented or interpreted. It is argued that we need to pay urgent attention to the unacceptable high number of cases that drop out of the system, and ensure that victims are given the tools and support to participate effectively in the legal process. Lillian Artz and Dee Smythe Gender, Health & Justice Research Unit, UCT Lillian.Artz@uct.ac.za Dee.Smythe@uct.ac.za SA CRIME QUARTERLY No 22 • DECEMBER 200714 Attrition rates and conviction rates It is perhaps important at this juncture to call attention to some of the difficulties in establishing and discussing attrition rates. For instance, depending on what stage of the criminal justice process ‘case fallout’ is established, attrition rates might be calculated according to the proportion of cases where there has been a conviction, or the total number of rape cases reported to the police (referred to as the report-to-conviction rate), or the proportion of cases convicted out of the total number of rape cases brought to trial (referred to as the trial-to- conviction rate). The report-to-conviction rate will always be substantially lower than the trial-to- conviction rate. When ‘conviction rates’ are reported in research, it is often not clear whether the rate is reflective of conviction rates of cases reported (a docket opened) or investigated (charges laid) by the police, or whether conviction rates apply only to cases which have been brought to trial. Conviction rates are lower when the statistics include cases that were reported, but not investigated (for example, disposed of at reporting or early on in the investigation stage of the case). They may also be calculated with or without the number of convictions overturned on appeal. In specific relation to rape cases, rape conviction rates may also include rape and attempted rape, or just rape. Thus, specialised sexual offences courts should expect to yield higher conviction rates, due to the specialised nature of the prosecutorial and court practices in these courts, and because the conviction rate is calculated using the trial-to-conviction formula. Ironically, there is no clear indicator – across all offence types and in relation to sexual offences more specifically – of what constitutes a ‘good conviction rate’. Both criminal justice reports and more scholarly research on attrition and conviction rates have sidestepped the setting of concrete thresholds for what constitutes effective prosecution and ‘good’ conviction rates, despite using them as performance indicators. This raises the question of whether an ‘increase’ in conviction rates means that the actual criminal justice process is more effective, or whether legal measures to eliminate inappropriate, insensitive and discriminatory practices within the criminal justice system. The government has signified a commitment to interleave these proposed legal reforms into practice. This is illustrated by the proposed adoption of national health guidelines for the management of rape survivors, national anti-rape strategies as well as the establishment of designated sexual offences courts, multi-service Thuthuzela Care Centres and inter-service level protocols to guide the management and disposition of cases. At its most instrumental, the aim of the shifts in law and practice is to increase reporting and conviction rates in rape cases. Paradoxically, the measures by which criminal justice agencies gauge their success in dealing with cases is not always consonant with these objectives. The most obvious example is the focus on decreasing the levels of reported rape. While laudable, it is not the level of reported rape cases that is at issue. Instead, it is the staggering number of cases that – due to the inadequate management of rape cases – do not make it through the criminal justice system that is of primary and immediate concern. Our research on sexual offences over a number of years has illustrated that at various stages within the criminal justice process, cases simply ‘drop out of the system’ – a phenomenon known as ‘case attrition’. Between 2003 and 2006 we conducted two studies, which together examined the disposition of approximately 1 600 rape cases across six urban police stations. The objectives of these studies were to examine the processing, investigation and prosecution of sexual offences cases as well as to analyse the possible reasons for high attrition rates in sexual offences cases. While the number of cases that simply ‘drop out’ of the system is alarming, it is the factors that contribute to attrition that require in-depth attention. Much like current discussion on reporting/conviction rates, it is not the numbers that are relevant, but the actual reasons for attrition that are worth examining. ARTZ AND SMYTHE SA CRIME QUARTERLY No 22 • DECEMBER 2007 15 convictions are a good indicator of effective justice for rape complainants. Bearing in mind the original emphasis of the Sexual Offences Bill – to effect fair, sensitive and appropriate justice and to protect rape survivors within the courtroom – the conviction rate question becomes somewhat negligible. In addition, the goals of case disposition also vary between criminal justice agents. For instance, for the police, low reporting of cases is considered a good indicator of policing (effective crime prevention), and a ‘high’ rate of referrals to prosecution is a good indicator of effective case disposition. For the prosecution service, high numbers of case referrals mean crime prevention is failing (because of increased levels of crime). Similarly, successful prosecution is based on conviction rates, to the exclusion of other indicators that may signal successful prosecuting practice. In rape cases, this may include key performance indicators that may not be easily quantifiable, but reflect the more qualitative aspects of investigations and prosecutions as well as the experiences of victims throughout the investigation and trial processes. How attrition happens At each stage of the criminal justice process there are multiple opportunities for discretion to be exercised. Rape cases within the criminal justice system are disposed of or finalised in a number of different ways. These include: • Undetected • Undetected – complainant not traced • Withdrawn – no consequence • Undetected – warrant issued • Nolle prosequi • Withdrawn in District Court • Acquittal at Regional Court See the box below for more detail on these processes. Ways in which rape cases can be disposed of ARTZ AND SMYTHE • In order for a case to be categorised as ‘undetected’, the police standing orders on closing of dockets specifies that the investigation should have failed to disclose the identity of the offender, although the police are convinced on the basis of prima facie evidence that an offence has been committed. In other words, undetected cases are those where a rape is believed to have occurred, but the police have been unable to positively identify the offender. In police terms it constitutes a failed investigation. • The police standing orders for closing of dockets allows for a docket to be closed as ‘undetected – complainant not traced’ when a complainant cannot be found after reporting the matter. This category accounts for around one in ten rape cases reported. • In terms of the standing orders an investigating officer may only withdraw a case of ‘no consequence’ upon an affidavit from the complainant requesting withdrawal. • If the identity of the perpetrator is known, but his whereabouts are not, the police standing orders provide that a case may be filed as ‘undetected – warrant issued’. Should the perpetrator resurface at a later stage he may be arrested on this warrant. • A prosecutor may decline to prosecute an alleged offence when he/she does not believe that there is a reasonable prospect of instituting a successful prosecution. In other words, there is no prima facie case on the basis of which to pursue prosecution at that time. A case may be dismissed nolle prosequi at any stage before the accused pleads to the charges. • Bail applications are heard in the District Magistrates Court and, for the most part, cases are withdrawn at this level for further investigation. Note that indications that a case has been ‘referred to court’ (that is, successfully investigated) must be seen against the fact that at some stations as many as one third have in fact been withdrawn (at the District Court), for one reason or another, without having been captured on the system. Example of attrition based on the findings of a 1998 CIETafrica study For every 394 women raped in the Southern Johannesburg Metropole 272 (69%) reported the attack to the police Of those who reported, only 17 (6%) became ‘rape cases’1 This means that a rapist in the Southern Johannesburg Metropole had a one in approximately 394 chance of being convicted. 16 ARTZ AND SMYTHE SA CRIME QUARTERLY No 22 • DECEMBER 2007 In South Africa, the practice of ‘filtering’ rape cases through the criminal justice system was only recently identified as a serious concern. The findings of a 1998 CIETafrica study in the box below demonstrate how attrition works. At each point at which cases have been shed from the system there has been attrition. Interestingly, each of these attrition points coincides with the stages of the criminal justice process where criminal justice personnel exercise the most discretion. From a reformist perspective, it seems obvious that it is exactly these sites of discretion that need to be more strictly regulated if attrition rates are to decrease and we are to see increased prosecutions and convictions (assuming that increased convictions are a good indicator of effective justice). But attrition does not work only in the reasonably linear broad strokes painted by CIETafrica. At each stage there are multiple opportunities for discretion to be exercised and incentives for exercising them in a particular way. It could also be argued that in some instances victims also exercise choice and agency in their own dealings with the system, contributing to attrition. The CIETafrica study illustrates that the number of reported rapes is relatively high, but only a small proportion of these cases actually make it to trial or result in conviction. Key factors contributing to attrition in rape cases include the victim’s decision to report the rape, the likelihood of arresting the accused, the scope of the investigation, the dismissal of the case by the prosecutor, and acquittal at trial. Other studies (Kelly 2002) have shown that factors increasing the likelihood of arrest include: • The use of a weapon and/or the use of force • The level of resistance used by the victim • The existence or availability of other witnesses If the victim appears ambivalent, ‘difficult’, intoxicated, or confused about the facts of the case, police are less likely to vigorously pursue the case. Our analysis of rape cases has shown that there is considerable variance from station to station and court to court, even within the same magisterial jurisdiction, in the disposition of cases. It is acknowledged, of course, that the nature and extent of attrition is dependent on the individual circumstances of each reported case. Our findings, however, show that attrition goes far beyond these individual factors and implicates more serious systemic disparities in the management of rape • At the Regional Magistrates Court cases may result in convictions or acquittals. Many are however withdrawn by the prosecution under section 6(a) or 6(b) of the Criminal Procedure Act (51 of 1977). Section 6(a) refers to the authority of the prosecutor to withdraw a charge before the accused pleads to that charge, in which event the accused is not entitled to a verdict of acquittal in respect of that charge. Section 6(b) refers to the authority of the prosecutor to, at any time after an accused has pleaded (but before conviction) stop the prosecution in respect of that charge, in which event the court trying the accused must acquit the accused in respect of that charge 1 of the 17 was ‘lost’ in a manner considered fraudulent 5 were referred to court for prosecution 1 resulted in a conviction Table 1: How rape cases were ‘finalised’ in two neighbouring urban police stations How cases % cases filed % cases filed were filed in police in police station A station B Undetected 33 4 Case withdrawn due to complainant 16 7 Withdrawn in court 21 8 Nolle prosequi 16 61 Finalised in regional court 12 17 SA CRIME QUARTERLY No 22 • DECEMBER 2007 17ARTZ AND SMYTHE cases. This is illustrated by the example provided in Table 1 below. social interactions and what they perceive as constituting a genuine incident of rape. These myths, as Kelly (2002) has argued, are ‘non-factual presumptions that serve (intentionally or unintentionally) to deny, minimise or misrepresent what we know from both research and accounts of victims and perpetrators about rape. Other stages of the criminal justice process – including reporting, forensic medical examinations, statement taking, investigations/evidence gathering and arrest of accused persons – are also key attrition points in our criminal justice system. The ability to find the accused, and in some cases the complainant, has a great impact on the ability of the criminal justice system to assist rape complainants. Without the accused the case cannot proceed. Other aspects of the investigation, including the ability of investigating officers to collect appropriate and relevant evidence for the prosecution of rape cases, are also questionable in a significant proportion of cases. The quality of medico-legal examinations by medical practitioners is also similarly critical to the effectiveness and integrity of investigations and prosecutions. The medico-legal examination often forms a crucial aspect of rape cases and therefore requires detailed attention to injuries and complaints made by the survivor at the time of the examination. When these examinations are not properly documented or are incomplete, the tenacity of the evidence and the strength of prosecution can be severely compromised. The impact of discretion in rape cases International studies on police investigation and prosecution of rape cases have found that police officers and prosecutors become particularly sceptical of rape victims when their stories do not coincide with what Estrich (1987) has called the ‘real rape’ template. Kelly (2002) makes a similar finding. As a result the credibility of rape victims is questioned. Temkin (1999) similarly found that the police and prosecution service still held a culture that anticipates high levels of false reporting and that the majority of cases are ‘lost’ due to their designation as false allegations by the police or because of victims withdrawing their statements. Seen in isolation from prosecutorial decision- making, police performance at station B above seems very impressive (Table 1). This is certainly true when one compares the 33 per cent of cases filed as undetected at station A against the mere 4 per cent at station B. However, on closer examination it appears that we may not in fact be looking at better performance and certainly, as far as victims are concerned, are not achieving the goal of better service provision. Comparative studies illustrate that the highest proportion of cases fall out at the early stages, with half or more cases dropping out even before referral to prosecutors. Similar to international findings on attrition, the results of our own research have also highlighted key attrition points in the criminal justice system. The first, and most difficult to investigate, is the decision of rape victims not to make an official report – a phenomenon that is simply incalculable in terms of criminal justice statistics. The next most noteworthy attrition point is police discretion. Police use their wide discretionary powers to establish whether an incident is ‘criminal’ or warrants investigation in ways that replicate traditional interpretations – often based on stereotypical assumptions – of what constitutes ‘real rape’ and what is considered ‘criminal’ activity. This is sometimes based on what the police perceive to be acts that occur ‘naturally’ within intimate or Our research in the South African context has repeatedly demonstrated a pervasive belief by criminal justice personnel that there are a large proportion of cases that are withdrawn because of ‘false allegations’. Our analysis of dockets, however, shows little evidence of this, with less than 1 per cent cases being withdrawn by complainants on the basis of false allegations. These instances were cases of alleged statutory rape. Comparative studies on attrition have shown that the manner in which police discretion is exercised is crucial to the effective management of sexual assault cases. An investigation into attrition rates should therefore consider the ‘modes of discretion’ used by charge officers, investigating officers and prosecutors in rape cases to declare the cases ‘unfounded’ (when cases are dropped because of lack of merit), or to continue (investigate and bring to trial) rape cases. Our research found that the following elements of case disposition are particularly relevant in examining attrition: • The factors and elements used by police and prosecutors to determine whether the case is ‘unfounded’ or worthy of investigation and prosecution. For example, what they believe they are expected to do by law in terms of substantive definitions and evidentiary procedures. • The factors important to criminal justice agents in deciding whether to arrest, investigate or prosecute in a rape case. This includes factors considered to be important in producing successful judicial outcomes. • Investigation and prosecutorial methods, strategies or policies applied and considered useful in processing rape cases. • Factors that limit or hamper effective investigation and prosecution of rape cases, including infrastructural/material, procedural, circumstantial and personal obstacles. The quality of investigations by the police is universally cited, and locally confirmed, as a major factor in attrition of rape cases. The accessibility of investigating officers, high case loads and the extent to which investigating officers are ‘qualified’ to investigate rape cases are contributing factors to the quality of rape investigations. Information regarding the status of a case, including of an arrest, is difficult SA CRIME QUARTERLY No 22 • DECEMBER 200718 ARTZ AND SMYTHE to establish. Statement taking by the police is also problematic, with dockets containing often-vague victim statements, which not only contain irrelevant information and details, but do not even set out the basic elements of the offence. Clearly, a high proportion of these cases are getting lost at the early stages of the criminal investigation. This may be due to the police designating cases as ‘false reports’ or as ‘withdrawals by the victim’. Internationally – and increasingly in South Africa – research is also beginning to reveal that the two most important factors influencing the outcome of a rape trial are the evidence of physical injury and admission to offence by the perpetrator. Other significant studies, spanning a period of 30 years, support our contention that inappropriate or poor police discretion directly affects attrition rates (Temkin 1997; Kelly 2002). International studies also alert us to the importance of addressing poor administration – such as delays and postponements, lack of pre- and post-court support and courtroom intimidation – in curbing attrition, and that policing and prosecutorial agencies still rely on stereotypes about rape victim credibility (Bargen & Fishwick 1995; Polk 1985; Frohmann 1991; Kersetter 1990; La Free 1989; Martin & Powell 1994). Similar to our experiences in South Africa, Kelly’s research on attrition in the UK found that the reporting and investigation stage is the point at which the largest number of rape cases leak from the system. She argues that: the initial responses of police officers, their skill and expertise as investigators and evidence gatherers, as well as their treatment of complainants are vital elements in criminal justice system responses [to rape cases] (Kelly 2002). She also found that a staggering 62 per cent of cases reported to the police fall out during the investigation process, either because the perpetrator cannot be identified or found, or because insufficient evidence is collected. Of those cases that are referred to the prosecution, many are dismissed by the prosecution (nolle prosequi) without ever going to trial. SA CRIME QUARTERLY No 22 • DECEMBER 2007 19 Adler’s (1987) analysis of rape cases found that the success of the rape complaint was consistently based on six predicators: • The victim’s sexual inexperience • Her respectability • Absence of consensual contact with the perpetrator prior to the rape • Resistance and injury • Early complaint • A lack of acquaintance with the accused The way in which the responses of police, prosecutors and judges shape the construction of rape within the criminal justice system has been the subject of scathing critique, most notably in the form of Estrich’s landmark book, ‘Real Rape’ (1987). ‘Real rapes’, according to Estrich, are still those involving a weapon and injury, committed by strangers, outdoors. These are the cases that criminal justice personnel take seriously. Kelly (2002) also speaks of the ‘real rape template’ adopted by criminal justice agents, arguing that conformity to this template, (which informs the victim’s self-conception of the assault as a rape and her belief that the police will also see it that way) is one of the strongest predictors of whether a rape complaint will make it all the way through the system. Frohmann’s 1997 study into prosecutorial discretion within two US jurisdictions provides a useful insight into this aspect of criminal justice practice. She illustrates in this study the exercise of prosecutorial discretion through ‘official typifications of rape- relevant behavior’ (Frohmann 1997:217), in respect of ‘rape scenarios’, ‘post-incident interaction’ (Frohmann 1997:218), ‘rape reporting’ (Frohmann 1997:219), and ‘victim’s demeanor’, used by prosecutors to inform their decisions as to whether the complainant is credible. In respect of each of these aspects she shows how prosecutors draw on a store of subjective ‘knowledge’, through which they have constructed a ‘typical’ rape scenario against which complaints are measured (Frohmann 1997:217-219). She emphasises the important role that ‘convictability’ (Frohmann 1997:399) plays in shaping prosecutorial decisions and argues that this narrow approach is self-reinforcing: prosecutorial assessments of the way in which decision-makers ‘downstream’ (ultimately the jury or judge) will evaluate the complaint inform their decision whether to send the matter to trial. However, this means that only a narrowly defined group of potentially convictable cases gets seen in court, which reinforces stereotypical perceptions of what amounts to ‘real rape’. In South Africa, the decision whether to follow through with a rape case appears, at least in part, also to be based on what the criminal justice agent anticipates will happen at the next stage of the criminal justice process: • For the reporting officer the question is whether the investigating officer will have enough information to proceed with the investigation • For the investigating officer the question is whether the prosecutor will prosecute the case on the basis of investigation • For the prosecutor the question is whether the court will find the offender guilty of the offence Both investigating officers and prosecutors inevitably approach a rape complaint from a cost- benefit perspective that is ultimately focused on the ‘convictability’ of the case and an evaluation of whether the case has evidential difficulties. That is, given the resources to hand, will the time, energy and money spent on investigation and preparation for trial, result in a realistic possibility of conviction? Conclusion In 1993 Henderson (1993:41) wrote that: Two decades of feminist law reform efforts to hold men responsible for raping women have yielded disappointing results. Rape myths, woman-blaming, and resistance to taking rape seriously flourish, and successful prosecution of cases not meeting the stereotype of real rape, while no longer impossible, remains improbable. It is our contention that attrition rates will remain inordinately high despite new law reform efforts. Contributing to perspectives on the ‘successful’ ARTZ AND SMYTHE SA CRIME QUARTERLY No 22 • DECEMBER 200720 investigation and prosecution of rape cases, and the resultant impact on official attrition rates, scholars point to evidence that laws are not being applied (Adler 1987; NSW Department of Women 1996) and, when they are applied, that they are narrowly interpreted (Adler 1987; NSW Department of Women 1996) and have thus been ‘rapidly undermined’ (Kelly 2002:33). While the significance of the new Sexual Offences Bill lies in its expanded definition of rape and other sexual offences (see Lisa Vetten’s article in this issue), there is little evidence that it will make a recognisable impact on the management of sexual offences cases. In order to achieve its stated objectives of providing protection, reducing secondary victimisation and trauma within the criminal justice process, and offering timeous, effective and non-discriminatory investigation and prosecution, the proposed Bill offers little in this regard. Without creating enforceable, regulatory provisions for the effective administration and implementation of the law, the high levels of attrition will continue unabated. Key to reducing attrition is ensuring that mechanisms are created that reduce excessive discretion, increase accountability to victims and other criminal justice agents, and ensure that systems are created to fully trust, support and encourage victim participation in the criminal justice process. Acknowledgements An extended version of this article was originally published as Artz, L & Smythe, D 2007. Case attrition in rape cases: a comparative analysis. South African Journal of Criminal Justice, 20(2):158- 181. Endnotes 1 It should be noted that some of the ‘excluded’ cases may result from re-labelling of the offence to, for example, indecent assault. A substantial portion reflects, however, police decisions to close the matter or victims that decide (or are persuaded) to drop the case. See CIETafrica (1988:44). References Adler, Z 1987. Rape on Trial. London: Routledge and Keegan. Artz, L and Smythe, D 2007. Case attrition in rape cases: A comparative analysis. South African Journal of Criminal Justice, 20(2). Bargen, J and Fishwick, E 1995. Sexual Assault Law Reform: A National Perspective. Canberra: Office of the Status of Women. Andersson N, Mhatre S, Mqotsi N and Penderis M 1998. Prevention of sexual violence: a social audit of the role of the police in the jurisdiction of Johannesburg’s Southern Metropolitan Local Council. Johannesburg: CIETafrica in collaboration with the Southern Metropolitan Local Council. Criminal Law (Sexual Offences) Amendment Bill 50 [2003]. Estrich, S 1987. Real Rape: How the Legal System Victimizes Women Who Say No. Cambridge: Harvard University Press. Frohmann, L 1990. Discrediting Victims’ Allegations of Sexual Assault: Prosecutorial Accounts of Case Rejections. Social Problems, 38(2): 213-226. Henderson, L 1993. Getting to Know: Honoring Women in Law and in Fact. Texas Journal of Women & Law, 41(2): 42. Kelly, L 2002. Routes to Injustice: A Research Review on the Reporting, Investigation and Prosecution of Rape Cases. London: HMCPSI. Kersetter, W 1990. Gateway to justice: Police and prosecutorial response to sexual assaults against women. Journal of Criminal Law and Criminology, 81(2): 267-313. La Free, G 1989. Rape and Criminal Justice: The Social Construction of Sexual Assault. CA: Wadsworth. Martin, P and Powell, R 1994. Accounting for the "second assault": Legal organizations’ framing of rape victims. Law and Social Inquiry, 19(4): 853-890. New South Wales Department of Women 1996. Heroines of Fortitude: The Experiences of Women in Court as Victims of Sexual Assault. NSW: Department of Women. Polk, K 1985. Rape reform and criminal justice processing. Crime and Delinquency, 31(2): 191-205. South African Law Commission Discussion Paper 1999. The Substantive Law of Rape. Temkin J 1997. Plus ça change: Reporting Rape in the 1990s. British Journal of Criminology, 37(4): 507-528. ARTZ AND SMYTHE