Police Classification of Sexual Assault Cases as Unfounded An Exploratory Study Prepared by Linda Light and Gisela Ruebsaat Research and Statistics Division Department of Justice 2006 The views expressed in this publication are those of the authors and do not necessarily represent the views of the Department of Justice Canada or the Government of Canada. Information contained in this publication or product may be reproduced, in part or in whole, and by any means, for personal or public non-commercial purposes, without charge or further permission, unless otherwise specified. • You are asked to: exercise due diligence in ensuring the accuracy of the materials reproduced; indicate both the complete title of the materials reproduced, as well as the author organization; and indicate that the reproduction is a copy of an official work that is published by the Government of Canada and that the reproduction has not been produced in affiliation with, or with the endorsement of the Government of Canada. • Commercial reproduction and distribution is prohibited except with written permission from the Department of Justice Canada. For more information, please contact the Department of Justice Canada at: www.justice.gc.ca ©Her Majesty the Queen in Right of Canada, represented by the Minister of Justice and Attorney General of Canada, 2017 2 Acknowledgements The Justice Institute of British Columbia wishes to acknowledge and thank the following for their support and assistance in the research process:  The Policy Centre for Victim Issues and the Research and Statistics Division at the Department of Justice Canada  The Victim Services and Crime Prevention Division at the BC Ministry of Public Safety and Solicitor General  The BC Ministry of Community Services  The BC Association of Specialized Victim Assistance and Counselling Programs  The Vancouver Police Department  The Richmond, Chilliwack, and Langley RCMP Detachments  Vancouver Police Department  Operations Policy Branch, “E” Division RCMP  Operations Support Officer, Langley RCMP Detachment  The key informants who gave so generously of their time and expertise. 3 Contents Acknowledgements....................................................................................................................................... 3 List of Tables ................................................................................................................................................. 6 Executive Summary....................................................................................................................................... 7 1.0 Introduction .................................................................................................................................. 7 2.0 Targeted Literature Review .......................................................................................................... 7 3.0 Research Approach ....................................................................................................................... 8 4.0 Investigating and Classifying Sexual Assault Allegations .............................................................. 9 5.0 Conclusions and Implications of the Research Findings ............................................................. 10 1.0 Introduction .................................................................................................................................... 11 2.0 Targeted Literature Review ............................................................................................................ 12 2.1 Trends in unfounded rates for sexual assault ............................................................................. 13 2.2 Factors affecting police decision-making ................................................................................... 14 3.0 Research Approach ......................................................................................................................... 15 3.1 Site selection ............................................................................................................................... 15 3.2 File review ................................................................................................................................... 15 3.3 Data analysis ............................................................................................................................... 16 3.4 Interviews and discussions with key informants ........................................................................ 17 3.5 Ethics review ............................................................................................................................... 17 Study value and limitations......................................................................................................................... 18 4.0 Investigating and Classifying Sexual Assault Allegations ................................................................ 18 4.1 Police policy on sexual assault .................................................................................................... 18 4.2 Case classification categories ...................................................................................................... 19 4.3 Completing Violent Crime Linkage System (VICLAS) Booklets .................................................... 20 4.4 Investigating sexual assault allegations and recommending charges ........................................ 20 4.5 Classifying and scoring cases ...................................................................................................... 21 4.6 Quality control ............................................................................................................................ 21 5.0 Research Findings and Discussion................................................................................................... 22 5.1 Investigative issues ..................................................................................................................... 22 5.1.1 Whether the victim and suspect were strangers ................................................................ 23 5.1.2 Whether there were mental health issues/disabilities....................................................... 24 5.1.3 Whether force was used .................................................................................................... 27 5.1.4 Whether the victim physically resisted ............................................................................... 28 4 5.1.5 Whether the victim said “no” ............................................................................................ 29 5.1.6 Whether the victim appeared upset .................................................................................. 30 5.1.7 Formality of the police investigation .................................................................................. 32 5.2 Scoring files and other data reporting issues ............................................................................. 36 5.2.1 Confusion between unfounded and founded but not cleared categories ......................... 36 5.2.2 Confusion between unfounded and RCMP unsubstantiated categories............................ 39 5.2.3 Inconsistent use of VICLAS booklets ................................................................................... 39 5.2.4 Inconsistent use of codes for spousal violence................................................................... 40 5.2.5 Deficiencies in quality control ............................................................................................. 40 5.2.6 Lack of national data on unfounded determinations ......................................................... 40 6.0 Conclusions and Implications of the Research Findings ................................................................. 41 6.1 Implications for sexual assault investigations ............................................................................ 41 6.1.1 Training ............................................................................................................................... 41 6.1.2 Policy and operational guidelines ....................................................................................... 42 6.1.3 Specialization ...................................................................................................................... 42 6.2 Implications for statistical data collection ................................................................................. 42 6.3 Implications for further research ................................................................................................ 43 6.4 Implementation of the research findings: the way forward....................................................... 44 Appendix I ................................................................................................................................................... 45 References .................................................................................................................................................. 48 Case Law...................................................................................................................................................... 50 5 List of Tables Table 1: Whether victim and suspect were strangers: victims of all ages .................................................. 23 Table 2: Whether victim and suspect were strangers: victims 14 years and over ..................................... 24 Table 3: Victim’s/parent’s mental health/mental disability issues: victims of all ages in Chilliwack and Langley (jurisdictions with relatively high unfounded rates)...................................................................... 25 Table 4: Victim’s/parent’s mental health/mental disability issues: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) ....................................................................................... 26 Table 5: Whether force was used: victims of all ages in Chilliwack and Langley (jurisdictions with relatively high unfounded rates)................................................................................................................. 27 Table 6: Whether force was used: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions)................................................................................................................................................ 27 Table 7: Whether the victim physically resisted: victims 14 years and over .............................................. 28 Table 8: Whether the victim said “no”: victims 14 years and over in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) ............................................................................................................................. 29 Table 9: Whether the victim appeared upset or not: victims of all ages.................................................... 31 Table 10: Whether the victim appeared upset or not: victims 14 years and over ..................................... 31 Table 11: Victim-generated evidence in the file: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) ............................................................................................................................. 33 Table 12: Witness-generated evidence in the file: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) ............................................................................................................................. 33 Table 13: Suspect-generated evidence in the file: victims of all ages ........................................................ 34 6 Executive Summary 1.0 Introduction The General Social Survey (GSS) on Victimization indicates that the non-reporting rate in sexual assault cases, already significantly higher than for other violent offences, rose from 78% in 1999 to 88% in 2004 (Statistics Canada, 1999, 2004). This research addresses one of the considerations thought to contribute to high non-reporting rates: the high proportion of sexual assault allegations that are classified as “unfounded” by police, compared to other violent offences. In fact, the rate of sexual assault cases deemed to be unfounded – that is, not to have occurred or not to have been attempted – has been shown to be twice as high as for other violent offences, and this difference has been sustained over several decades (Kong et al, 2003). Further, rates of unfounded classifications have been shown to vary widely from jurisdiction to jurisdiction. Any erroneous classification of sexual assault cases as unfounded has serious implications for reported crime statistics since, once determined to be unfounded, these cases are not included in numbers at either the local or the national level. In addition, there can be serious implications for victims’ access to justice and support services as well as their safety and the prevention of further violence, since complainants that have been determined not to be victims would be unlikely to be referred to victim support services. Given the above, high unfounded rates have long been a source of tension between police and individual victims, as well as between police and the wider women- and victim-serving community. Understandably, erroneous unfounded classifications can contribute to an erosion of women’s confidence in the justice system which, in turn, can lead to their reluctance to report sexual assault to the police. This research was undertaken for the Justice Institute of British Columbia, with funding from the Policy Centre for Victim Issues and the Research and Statistics Division at the Department of Justice Canada, the BC Ministry of Community Services, and the BC Ministry of Public Safety and Solicitor General. 2.0 Targeted Literature Review The targeted literature review undertaken for this project identifies the range of factors that researchers have examined to assess their potential impact on police decision-making with respect to the classification of sexual assault cases as founded or unfounded. These include the quality and consistency of the information obtained on the incident; victim characteristics, including moral character, behaviour and credibility; the relationship between victim and suspect; offence characteristics; suspect characteristics; the existence of corroboration for either the complainant or the accused; and expectations about the responses of Crown, defence, or judges. Not all study findings as to the impact of these factors are consistent. 7 Studies have also examined the impact of system expectations and the need for “efficiency”, as well as definitional problems (both in the Criminal Code definition of sexual assault and in the interpretation of the definition of unfounded) on unfounded rates. These also appear to be influential in terms of police decision-making. In spite of the promising research conducted after the 1983 legislative reforms which introduced three levels of sexual assault to replace those of rape and indecent assault, little follow-up research has been done to explore local or regional differences in sexual assault founding rates, what factors might contribute to those differences, and what steps might be taken to improve police processing of sexual assault cases. 3.0 Research Approach This exploratory research was undertaken using a collaborative design, working with the Vancouver Police Department, the “E” Division of the RCMP, the BC Association of Specialized Victim Assistance and Counselling Programs (BCASVACP)1, and the Justice Institute of BC (JIBC), under whose auspices the research was conducted. The research took a primarily quantitative methodological approach, supplemented by interviews and discussions with key informants. The research sites included in the study were all within driving distance of the Lower Mainland in BC. They were selected with a view to including both independent police and RCMP jurisdictions, and balancing jurisdictions where the unfounded rate for sexual assault was higher than the provincial average and jurisdictions where the rate was lower. Ultimately, four sites were included in the study: the Chilliwack2, Langley, and Richmond RCMP detachments, and the Vancouver Police Department (VPD). A total of 148 police sexual assault files were selected from the four sites. An extensive File Review Guide was used to collect information concerning factors such as the nature of the alleged incident; the initial disclosure or report to police; the police or Crown’s response or action; victim characteristics and reactions; and suspect characteristics and behaviours. Following a preliminary statistical analysis of the file data, interviews and/or small group discussions were conducted with a total of 18 key informants within the BC policing community. The purpose of the interviews and discussions was to solicit the views of police members and records staff on investigative and statistical scoring issues that had arisen during the file review and data analysis; to fill in gaps in the researchers’ knowledge about investigative and statistical scoring practices concerning sexual assault allegations; and to examine the implications of the preliminary research findings. The reader is reminded that this research is exploratory in nature. Resource limitations, as is frequently the case, have limited the number of research sites, the number of files reviewed, the number of key informants who provided input, and the extent to which the findings can be generalized. In spite of these limitations, some compelling results have emerged from this study, which provide important insights into police founding decisions in sexual assault cases. In addition, given that the investigative and offence classification issues identified in this report are consistent with those that are considered in 1 BCASVACP is a provincial organization representing community-based front-line victim service providers. 2 Chilliwack is now part of the Upper Fraser Valley Detachment, but is referred to in this report as the Chilliwack Detachment. 8 the literature (Light and Ruebsaat, 2005), there is no reason to believe that the sites included in the study are unique in terms of police beliefs, attitudes, and practices concerning sexual assault. 4.0 Investigating and Classifying Sexual Assault Allegations Important as the legislative response may be, no amount of legislation can outweigh the importance of an appropriate and effective police response to sexual assault allegations. Police are the point of entry into the criminal justice system, and their interventions can have significant consequences for subsequent stages of the criminal justice process. Further, police attitudes and behaviours can impact on whether or not particular victims feel supported, and whether or not they decide to engage in criminal proceedings. It is the role of the police to thoroughly investigate all sexual assault allegations and, ultimately, to conclude the investigation with a determination as to whether or not there is evidence that an offence occurred or was attempted. The conclusion of the investigation involves the classification of the case as founded or unfounded, and, if founded, as cleared by charge, cleared otherwise, or not cleared. Once a case is classified as unfounded, all investigation will cease and no more evidence will be collected. An additional case classification category is used in RCMP detachments, that of unsubstantiated. This is used as an “in-house term to track caseload for resource allocations”. A complaint is deemed to be unsubstantiated when there is no evidence to confirm either that it occurred or that it did not occur. It is important to note that complaints that are classified as unfounded or unsubstantiated are excluded from national crime statistics, thus impacting on reported sexual assault levels and clearance rates, both of which are based on founded occurrences. Over the course of the interviews and small group discussions, police members repeatedly emphasized the importance of a thorough investigation in determining whether or not a sexual assault allegation is founded or unfounded. All agreed that in order to classify a case as unfounded, there must be clear evidence that an offence did not occur or was not attempted. If no such evidence is obtained and documented, the case must be classified as founded but “not cleared”3. Key informants also agreed that the lack of evidence to proceed, or the victim’s shortcomings as a witness (as in the cases of a very young child or a person with severe mental health issues) is not sufficient for a case to be classified as unfounded. Again, it was felt that such a case must be classified as founded but “not cleared”. However, such understanding did not always appear to guide practice. Specifically, the file review uncovered a number of cases that concretely illustrate the confusion between, and inconsistent use of the unfounded and founded but not cleared categories. Such cases included, but were not limited to, those where the victim was a young child or had mental health issues or communication difficulties. Similarly, the RCMP classification of cases as unsubstantiated was difficult to differentiate from unfounded cases, even for some of those RCMP members who were questioned about this distinction. 3 Not cleared applies where there is no evidence that a crime did not take place but where there is not enough evidence to proceed. In such a case, the incident is reported to Statistics Canada as founded but not cleared. 9 5.0 Conclusions and Implications of the Research Findings The results of this exploratory research provide important insights into the founding process and the factors that may contribute to higher unfounded rates for sexual assault compared to other violent offences. Many of the files reviewed over the course of this research indicated that police had conducted thorough investigations, with a demonstrated sensitivity to sexual assault survivors. However, the files also indicate that there is room for improvement, both in police members’ understanding of the nature and dynamics of sexual assault, and their investigation of sexual assault allegations. Increased police focus on sexual assault training, policy, guidelines, quality control and specialization may contribute to the creation of a culture that reinforces an awareness and understanding of sexual assault, and of the importance of rigorous investigation of these cases despite difficulties in obtaining evidence stemming from factors such as the young age or mental health issues of complainants. The research findings also indicate a need for greater clarity and consistency in the classification and scoring of cases as founded or unfounded, particularly in relation to the distinction between the unfounded and founded but not cleared categories, and particularly when the issue is one of insufficient evidence. Overshadowing much of what is put forward in this report is the discovery that Statistics Canada no longer requests data on offences that are classified as unfounded from police services, and no longer analyses the data that it still receives due to data quality issues. This situation impacts on Statistics Canada’s ability to analyze and report on trends in the criminal justice response to sexual assault cases in BC and across the country. Without national data on unfounded rates, there will be no way of monitoring founded-unfounded case classification trends and no way of assessing the impact of any corrective measures that may be taken to ensure accuracy in decision-making regarding founded and unfounded determinations. In the absence of reliable data on a national basis, the issues raised in this report can be explored only on a site-by-site basis. Ideally, future studies would be based on larger and/or more diverse samples in order to allow for more extensive statistical analysis and a higher degree of statistical reliability. The issues addressed in this report also require more extensive qualitative research, in order to elicit information and opinions from a larger number of key informants, and informants from a wider range of disciplines, including community-based victim service providers and advocates. The open, constructive responses to these findings from police members who supported and participated in this study, from those working in the areas of police data-gathering and statistics, and from the women and victim-serving community have been encouraging to date. While there is a need for further exploration of the issues raised in this report, there are also clear indications of areas that need immediate attention. It is hoped that discussion and collaborative efforts to explore and address these issues will continue., utilizing these findings as a springboard from which to launch further initiatives at the local, provincial, and national levels. 10 1.0 Introduction The General Social Survey (GSS) on Victimization indicates that the non-reporting rate in sexual assault cases, already significantly higher than for other violent offences, rose from 78% in 1999 to 88% in 2004 (Statistics Canada, 1999, 2004). Thus, in spite of legislative reforms and significant police efforts to improve policies and practices in this area, sexual assault reporting rates are declining rather than improving. This research addresses one of the considerations thought to contribute to high non-reporting rates: the high proportion of sexual assault allegations that are classified as “unfounded” by police, compared to other violent offences. In fact, the rate of sexual assault cases deemed to be unfounded – that is, not to have occurred or not to have been attempted – has been shown to be twice as high as for other violent offences, and this difference has been sustained over several decades (Kong et al, 2003). Further, rates of unfounded classifications have been shown to vary widely from jurisdiction to jurisdiction. For instance, in 1994, unfounded rates for sexual assault level I ranged from 9% in Quebec to 21% in the Yukon (Roberts, 1994). Unfounded rates in BC also varied widely across jurisdictions. For example, the unfounded rates in the police sites included in this study ranged from 7% to 28% in 2002 and 2003. It is a well-known fact that most sexual assault survivors are female (Kong et al, 2003). In 2004, women were sexually assaulted at a rate five times higher than that for men, a gender difference that has not significantly changed since the 1999 GSS on Victimization (Gannon and Mihorean, 2005). For many of those working in the violence against women area, the high rate of unfounded determinations in sexual assault cases is part of a generally problematic justice system response to the offence. This concern is supported by research with sexual assault survivors, who report that their experiences during police investigations and court proceedings tend to make them feel like they are being dehumanized, blamed, and disbelieved (Hattem, 2000; Tomlinson, 1996). Police, on the other hand, contend that higher rates of false or otherwise unfounded sexual assault reports are to be expected, because of societal values related to sexual activity, the nature and dynamics of this offence, and the complexity of consent issues in these cases. However, widely differing unfounded rates across police jurisdictions as well as over time undermine this argument. Police have also contended that the problem may be, at least in part, a result of imprecise definitions of police scoring categories or victims’ misunderstanding of what constitutes sexual assault. There is some evidence in the literature to support these contentions (Muzychka, 1991; Roberts, 1990; Gunn and Minch, 1988). Any erroneous classification of sexual assault cases as unfounded has serious implications for reported crime statistics since, once determined to be unfounded, these cases are not included in numbers at either the local or the national level. In addition, there can be serious implications for victims’ access to justice and support services as well as their safety and the prevention of further violence, since complainants that have been determined not to be victims would be unlikely to be referred to victim support services. 11 Given the above, high unfounded rates have long been a source of tension between police and individual victims, as well as between police and the wider women- and victim-serving community. Understandably, erroneous unfounded classifications can contribute to an erosion of women’s confidence in the justice system which, in turn, can lead to their reluctance to report sexual assault to the police. The idea for this research on police classification of sexual assault cases as unfounded emerged at the 2003 National Victims of Crime Conference in Ottawa, where participants reiterated the need for more focused attention on sexual assault, in terms of both research and action to improve the criminal justice response. The research was undertaken for the Justice Institute of British Columbia, with funding from the Policy Centre for Victim Issues and the Research and Statistics Division at the Department of Justice Canada, the BC Ministry of Community Services, and the BC Ministry of Public Safety and Solicitor General. It is hoped that the research findings will contribute to a better understanding of police founding decisions, and to any changes in investigative and statistical practices that may be required. 2.0 Targeted Literature Review The targeted nature of this research project did not allow for a full review of the relevant literature. Therefore, with one or two exceptions, the literature review focused on research and reports produced in Canada since 1985. Thus, we were able to include the evaluation research and analytical work that followed the 1983 sexual assault legislative reform in Canada4. The review included literature that directly addressed the high unfounded rates for sexual assaults reported to police. Most of this literature focused on the effects of the 1983 reform legislation. Some of the studies cited used UCR (Uniform Crime Reporting)5 data to analyze national founding and charging trends before and after the legislative changes. Fewer still examined police files to identify the factors influencing police classification of sexual assault complaints as founded or unfounded. Since the 1990s, little research has been undertaken specifically addressing low founding rates in sexual assault. A number of more general reports in the early 1990s addressed the issue of gender bias in the justice system and its possible impact on justice system processing of sexual assault cases. These include: Gender Equality in the Canadian Justice System Federal / Provincial / Territorial Working Group of 4 In 1983, Bill C-127 introduced three levels of sexual assault that paralleled the three levels of physical assault. The new offences replaced those of rape and indecent assault, and were intended to emphasize their physical rather than their sexual nature. Unlike rape, sexual assault is a gender neutral offence which can be committed against women or men. (Bill C-127 also introduced restrictions on the admissibility of evidence of complainants’ prior sexual history, and eliminated spousal immunity.) Although by definition a gender neutral offence, the presence of unique male/female power dynamics in sexual assault cases continues to be recognized in case law: “It cannot be forgotten that a sexual assault is very different from other assaults. It is true that it, like all the other forms of assault, is an act of violence. Yet it is something more than a simple act of violence. Sexual assault is in the vast majority of cases gender based. It is an assault upon human dignity and constitutes a denial of any concept of equality for women.” (R. v. Osolin, 1994 at p. 521) 5 Many policing jurisdictions in Canada report criminal activity using the Uniform Crime Reporting (UCR) Aggregate Survey, the data collection system developed by Statistics Canada to capture police-reported crime. 12 Attorneys General Officials on Gender Equality in the Canadian Justice System, 1992); Gender Equality in the Justice System, (Law Society of British Columbia, 1992); and the British Columbia Task Force on Family Violence report, Is Anyone Listening? (Province of British Columbia, 1992). Only one study (Hattem, 2000) was based on interviews with sexual assault survivors, but this research explored the considerations that may influence a woman’s decision to report to police, rather than factors affecting police decision-making. 2.1 Trends in unfounded rates for sexual assault Unfounded rates for sexual assault have declined dramatically since Clark and Lewis’ ground-breaking, 1977 study which found that 63.7% of 116 reported cases were classified as unfounded. However, evaluations have produced mixed results on the extent to which the 1983 Criminal Code reforms have resulted in lower unfounded rates. For example, a Winnipeg evaluation (Gunn and Linden, 1991) found a significant decrease in the unfounded rate post-reform, while a Vancouver evaluation (Ekos Research, 1988) found no such difference. While no one suggested that police discretion could or should be eliminated, Johnson (1996) reflected the tone of other researchers when she commented that the 1983 reforms have not restricted the degree of police discretion in classifying specific sexual assault incidents. Julian Roberts (1990), who has studied this subject more than any other researcher in Canada, noted wide regional variations in sexual assault unfounded rates, ranging from 6.6% in Quebec to 29% in the Yukon in 1988. In addition, he suggested that the decline in these rates for Level III sexual assault in 1986, 1987 and 1988 was more likely a result of changes in how police were classifying the levels of sexual assault, rather than a real decline in the level of unfounded determinations. Overall, Roberts (1994) and Roberts and Gebotys (1993) concluded that there has been little change in the proportion of sexual assault cases classified as unfounded – not more than two percentage points since the year prior to the legislative changes – and that the unfounded rates in 1992 and in 1994 were still almost double that for assault level I. These authors also argued that any changes in clearance rates for sexual offences can be attributed to factors other than legislative changes, because they correspond to similar trends for other, non-sexual offences. Further, while Roberts and Mohr (1994) acknowledged that more sexual assault cases are filtered out of the criminal justice system than other violent offences, they argued that it is founding, not clearance rates, that are the culprit. They also highlighted the urgent need for research that compares attrition rates in sexual assault to those for other offences involving physical violence, given the far-reaching effects of significantly lower founding rates for sexual assault. These include: • weakening women’s confidence in the justice system; • discouraging the reporting of sexual assault; and • lowering the rate of successful prosecutions. 13 2.2 Factors affecting police decision-making 6 The targeted literature review undertaken for this project identifies the range of factors that researchers have examined to assess their potential impact on police decision-making with respect to the classification of sexual assault cases as founded or unfounded. These include the quality and consistency of the information obtained on the incident; victim characteristics, including moral character, behaviour and credibility; the relationship between victim and suspect; offence characteristics; suspect characteristics; the existence of corroboration for either the complainant or the accused; and expectations about the responses of Crown, defence, or judges. Factors affecting police decision-making are both subjective (such as police judgments about the complainant’s affect or credibility) and practical (such as police judgments about the likelihood of charges being laid or conviction). Not all study findings are consistent. Most of the researchers cited pointed to preconceptions about complainants’ credibility as a significant factor in founding decisions. Some also pointed to social biases inherent in police decision-making (Tomlinson, 1996; Muzychka, 1991; Gunn and Minch, 1988), or found police bias to be the motivating factor in only a minority of cases did (Stanley, 1985; Clark and Lewis, 1977). System expectations and the need for “efficiency”, as well as definitional problems (both in the Criminal Code definition of sexual assault and in the interpretation of the definition of unfounded) also appear to be influential in terms of police decision-making. Several authors pointed out that police still look for “hard” or corroborating evidence, such as injuries, torn clothing or other witnesses, even though the Criminal Code no longer requires it in sexual assault cases (Du Mont and Myhr, 2000; Gregory and Lees, 1996; Muzychka, 1991). Ekos Research (1988), on the other hand, found no relationship between corroborating evidence or witnesses and the founding decision. There is also inconsistency between studies regarding whether certain victim characteristics such as age, nationality, and marital status impact on police founding decisions. Some of the factors emerging from the literature review as ones associated with the founding decision, in particular those cited in Appendix I under Expectations of Crown Counsel and the Courts, are clearly inconsistent with an appropriate determination of unfounded. As will be shown, these are illustrative of the confusion identified in the current study between the classifications of unfounded and founded but not cleared. The association of these factors with the founding decision indicates that cases are being inappropriately classified as unfounded, not on the basis of whether an offence occurred or not, but on the basis of whether the case is likely to be successfully prosecuted. In spite of the promising research conducted after the 1983 legislative reforms, little follow-up research has been done to explore local or regional differences in sexual assault founding rates; what factors might contribute to those differences; and what steps might be taken to improve police processing of sexual assault cases. 6 A list of all factors that were considered in the literature reviewed for this current study is included in Appendix I. Any factor that emerged in the current study as statistically associated with the founding decision is noted in the list, following the citations. 14 3.0 Research Approach This exploratory research was undertaken using a collaborative approach, working with the Vancouver Police Department, the “E” Division of the RCMP, the BC Association of Specialized Victim Assistance and Counselling Programs (BCASVACP)7, and the Justice Institute of BC (JIBC), under whose auspices the research was conducted. A collaborative working relationship with the policing community is essential in order to gain access to police files, to ensure the ongoing participation of police members and records staff, and to encourage the use of research findings to effect change where required. The BCASVAP’s involvement, as well as the researchers’ background and experience in the area, were important in ensuring that the research benefited from the perspectives of the women-serving and victim-serving community. The research took a primarily quantitative approach, supplemented by discussions with key informants. 3.1 Site selection The research sites included in the study were all within driving distance of the Lower Mainland in BC. They were selected with a view to including both independent police and RCMP jurisdictions, and balancing jurisdictions where the unfounded rate for sexual assault was higher than the provincial average and jurisdictions where the rate was lower. Hence, site selection was based in large part on police statistics concerning founding decisions for 2002 and 2003, the most recent years for which statistics were available at the time the research was initiated. Ultimately, four sites were included in the study: the Chilliwack8, Langley, and Richmond RCMP detachments, and the Vancouver Police Department (VPD). The VPD and the Richmond RCMP detachment had low unfounded rates, ranging from 7% to 12% for 2002 and 2003. The Chilliwack and Langley RCMP detachments had high unfounded rates, ranging from 19% to 28% during the same period9. In addition, the VPD has a specialized sexual offences unit10. 3.2 File review A total of 148 police sexual assault files were selected from the four sites: 34 from the Chilliwack detachment; 40 from the Langley detachment; 32 from the Richmond detachment; and 42 from the VPD. The large majority of files (122) dealt with level I sexual assault. Seventy-six of the files were 7 BCASVACP is a provincial organization representing community-based front-line victim service providers. Chilliwack is now part of the Upper Fraser Valley Detachment, but is referred to in this report as the Chilliwack Detachment. 9 More recent statistics indicate that sexual assault unfounded rates in Chilliwack and Langley declined in 2004. Chilliwack, at 21%, remained above the Lower Mainland average of 15%, while Langley, at 16%, was close to that average. Richmond’s unfounded rate, on the other hand, rose to 27%, almost doubling Vancouver’s unfounded rate of 8% remained lower than the average. 10 While the VPD unit is one of a very small number of specialized police sex offences units in BC, other jurisdictions have designated a particular investigator who specializes in sex offence cases. 8 15 classified as unfounded; two others were considered unsubstantiated and therefore not submitted to Statistics Canada; and seventy-two files were classified as founded11. An extensive File Review Guide was used to collect information concerning factors such as the following: the nature of the alleged incident; the initial disclosure or report to police; the police or Crown’s response or action; victim characteristics and reactions; and suspect characteristics and behaviours. In addition, each file was summarized in a brief narrative on the front page of the Guide, in order to facilitate quick recognition of the case during the data analysis process. Given the exploratory nature of the research, the File Review Guide included approximately 200 factors that had been identified in the literature as possibly impacting on police classification of sexual assault allegations12, whether or not these findings were replicated in other studies. Given that the literature is widely acknowledged to be limited, additional factors were included based on the experience of the researchers and the research partners in the area of sexual assault. These included:    3.3 Alternative forms of resistance, such as feigning illness in order to appear repulsive or claiming an STD/HIV infection (in order to explore police understanding of the full range of victims’ resistance) Details about the collection and preservation of medical and forensic evidence (in order to explore the possible impact of a lack of such evidence, or its improper storage and preservation) Information about whether the case was handled by general duty members or sexual assault specialists (in order to explore the impact of specialized units and/or specialized training on the founding decision). Data analysis Frequencies were calculated for all of the factors included in the File Review Guide. Given the large number of factors addressed and the comparatively small number of files included in the study, there were no responses to a number of questions. For many others, the number of responses was too small to allow for further statistical analysis. However, as will be demonstrated below, the lack of information was sometimes instructive in and of itself. Using SPSS13, cross-tabulations were performed for factors where it was speculated that a statistically significant association with the founding decision may emerge, and where the frequencies were high enough to support meaningful cross-tabulations. In order to focus specifically on the central question of this research, all cross-tabulations were run using the founded/unfounded split as the dependent variable. 11 Two unsubstantiated files were inadvertently included in the sample. However, because the facts of these files were not dissimilar to cases that were determined to be unfounded, they were added to the cases classified as such. 12 See Appendix I for a listing of these factors. 13 SPSS – Statistical Package for the Social Sciences – is a statistical software program. 16 This first run of cross-tabulations was conducted on the full sample of 148 cases. Those revealing no statistically significant associations were rejected. In addition, responses to certain questions were combined to create response categories with sufficient numbers to allow for further cross-tabulations. In order to isolate certain factors applicable only to victims of at least 14 years of age (such as saying “no” or physically resisting sexual advances), another sub-sample was created consisting of those cases where the victim was 14 years of age or older. The age of 14 was used as the cut-off because, for most purposes, it is considered to be the age of consent. Further, in order to facilitate comparison between jurisdictions with relatively low and high unfounded rates, the database was divided into two smaller data bases, with the Vancouver and Richmond sites in one, and the Chilliwack and Langley sites in the other. Two more sub-samples were created to facilitate comparison between the VPD, an independent municipal force, and the Chilliwack, Langley, and Richmond RCMP detachments. 3.4 Interviews and discussions with key informants Following a preliminary analysis of the file data, interviews and/or small group discussions were conducted with a total of 18 key informants within the BC policing community. The purpose of the interviews and discussions was to solicit the views of police members and records staff on investigative and statistical scoring issues that had arisen during the file review and data analysis; to fill in gaps in the researchers’ knowledge about investigative and statistical scoring practices concerning sexual assault allegations; and to examine the implications of the preliminary research findings. Key informants included police members and records staff from the RCMP and VPD, as well as staff from the Police Services Division of the Ministry of Public Safety and Solicitor General. Discussions were guided by two sets of questions, one for police members with expertise in the investigation of sexual assault, and another for civilian staff with expertise in the records and statistics area. The questions were sent out prior to the interview or discussion group. Although questions for police focused primarily on investigational issues, they were asked some questions about statistical scoring issues as well. Records staff and Police Services Division staff were asked about statistical scoring issues only. 3.5 Ethics review The research proposal was submitted to and approved by the Ethics Review Committee of the Justice Institute of BC. This body must approve all research involving human subjects that is conducted by or for the Institute. The ethics review process is consistent with that followed by BC universities. As previously noted, a considerable amount of personal data on both victims and suspects in sexual assault cases was collected from police files. In order to maintain confidentiality, File Review Guides were identified with a research number. A list that matched the research file numbers with police file numbers was maintained for validation purposes but kept separate from the completed Guides. 17 Study value and limitations The reader is reminded that this research is exploratory in nature. Resource limitations, as is frequently the case, have limited the number of research sites, the number of files reviewed, the number of key informants who provided input, and the extent to which the findings can be generalized. In spite of these limitations, some compelling results have emerged from this study, which provide important insights into police founding decisions in sexual assault cases. In addition, given that the investigative and offence classification issues identified in this report are consistent with those that are considered in the literature (Light and Ruebsaat, 2005), there is no reason to believe that the sites included in the study are unique in terms of police beliefs, attitudes, and practices concerning sexual assault. 4.0 Investigating and Classifying Sexual Assault Allegations Important as the legislative response may be, no amount of legislation can outweigh the importance of an appropriate and effective police response to sexual assault allegations. Police are the point of entry into the criminal justice system, and their interventions can have significant consequences for subsequent stages of the criminal justice process. Further, police attitudes and behaviours can impact on whether or not particular victims feel supported, and whether or not they decide to engage in criminal proceedings. As Roberts and Gebotys (1993) point out, improving police policies and practices in relation to sexual assault requires a better understanding of their role in investigating and classifying sexual assault allegations. 4.1 Police policy on sexual assault Police policy can play a key role in guiding police practice and the use of police discretion in investigating and classifying sexual assault cases. Most of the RCMP policy on sexual assault applies only to child sexual abuse. However, some policy applies to police relations with sexual assault centres, including third party reporting, and to police practice with respect to s. 810.2 recognisances for high risk offenders. In addition, the general policy on the use of polygraph tests states that it is generally not appropriate to administer such tests to victims of crime. Although the RCMP policy on sexual assault is slated for revision, at the time of writing, there was no indication as to when that revision might take place or what issues it might encompass. The VPD operational policy on sexual offences, which is only available to police, addresses investigational procedures, including: the preservation of evidence; securing the scene; documenting the victim’s condition; interviewing victims, witnesses and suspects; liaising with the Forensic Identification Squad and Sexual Offence Squad supervisor; liaising with medical personnel regarding medical forensic evidence; liaising with the Ministry of Children and Family Development; and 18 completing a General Occurrence (GO) Report and Violent Crime Linkage System (VICLAS) booklet. In addition, the VPD has a policy that it does not polygraph victims of sexual assault. 4.2 Case classification categories The following categories are used by police members and records staff to classify cases as founded or unfounded and to submit information to Statistics Canada. Official definitions or segments of definitions are in quotes14. Unfounded Unfounded “means that the police investigation has established that a crime did not happen or was not attempted.” Therefore, police can classify a case as unfounded only when they have evidence to confirm that an incident did not occur or was not attempted. Founded  Not cleared applies where there is no evidence that a crime did not take place but where there is not enough evidence to proceed. In such a case, the incident is reported to Statistics Canada as founded but not cleared.  Cleared by other means applies “when police decide not to complete a Report to Crown counsel (RCC) even though they have identified the offender and have sufficient evidence to support a charge. This would happen, for instance, in a case of diplomatic immunity, if the victim/complainant refuses to sign a formal complaint, if the alleged offender dies before he/she can be formally charged, if there are specific policy directives in place which, under special circumstances, preclude police from filing RCCs, for example, government directives relating to diversion and youth, or finally, if police exercise discretion and decide, for departmental reasons, not to charge.”  Cleared by charge applies “when police file a Report to Crown counsel alleging responsibility for an offence. This does not necessarily imply the swearing of an information or prosecution since Crown counsel may return the case to police due to insufficient evidence or take no further action.” 14 These official definitions and quotes from Statistics Canada or RCMP sources or correspondence were provided by the Research and Evaluation Unit of the Police Services Division at the BC Ministry of Public Safety and Solicitor General. 19 It is important to note that an additional case classification category is used in RCMP detachments, that of unsubstantiated. This is used as an “in-house term to track caseload for resource allocations”. A complaint is deemed to be unsubstantiated when there is no evidence to confirm either that it occurred or that it did not occur. These complaints are not reportable to Statistics Canada and therefore are excluded from national crime statistics. 4.3 Completing Violent Crime Linkage System (VICLAS) Booklets Police members are also required to fill out a VICLAS booklet for every eligible violent crime they investigate. VICLAS is a national system consisting of a central repository that is intended to capture, collate and compare information on violent offences, including sexual assault. Clearly, the completeness of information in this system is directly related to its potential effectiveness. An incomplete VICLAS database will have a negative impact on its potential to identify patterns of violent crime and violent offenders for investigative purposes. 4.4 Investigating sexual assault allegations and recommending charges It is the role of the police to thoroughly investigate all sexual assault allegations and, ultimately, to conclude the investigation with a determination as to whether or not there is evidence that an offence occurred or was attempted. The conclusion of the investigation involves the classification or scoring of a given case as founded or unfounded, and, if founded, as cleared by charge, cleared otherwise, or not cleared. It is important to note that once a case is classified as unfounded, all investigation will cease and no more evidence will be collected. BC is one of the few jurisdictions in Canada where police do not lay charges; they can only recommend to Crown counsel that a charge be laid. It is then up to the Crown to decide whether there is enough evidence to proceed. Unlike the police, the Crown’s decision to charge must be based on whether there is a substantial likelihood of conviction and whether charges are in the public interest. Over the course of the interviews and small group discussions, police members repeatedly emphasized the importance of a thorough investigation in determining whether or not a sexual assault allegation is founded or unfounded, and that this determination must be based on a broad range of factors. A member’s opinion that a victim is not telling the truth does not constitute sufficient grounds to classify a case as unfounded. Such a classification must be based on clear evidence that an offence did not occur or was not attempted. If no such evidence is obtained and documented, the case must be classified as founded but “not cleared”. Key informants also agreed that the fact that there is not enough evidence to proceed, or that the victim would not make a good witness (such as in the cases of a very young child or a person with severe mental health difficulties) is not sufficient for a case to be classified as unfounded. Again, it was felt that such a case must be classified as founded but “not cleared”. 20 4.5 Classifying and scoring cases Practices vary between the VPD and RCMP jurisdictions with respect to police and records staff roles in classifying and scoring15 criminal complaints. In some jurisdictions, records staff do all the classifying and scoring while in others, it is a shared responsibility between police members and records staff16. For instance, in the Chilliwack and Langley detachments, civilians do all the statistical scoring of police files. Dispatchers enter the first offence codes when they receive the call. Records staff members are responsible for selecting and recording the appropriate codes on the file jackets. A supervisor may review those codes and correct them where required, depending on the nature of the call. As the case proceeds, the codes will be amended by records staff to reflect changes in the status of the case. In Chilliwack, such changes may be requested by a member; in Langley, these are generally the responsibility of records staff. In both Chilliwack and Langley, codes are input into the system by a data entry clerk. It is these codes that are submitted to the national RCMP database and ultimately, through various conversion processes, to Statistics Canada. In Vancouver, police members initially select the offence codes to classify their report. Detectives can update these codes as the case proceeds, but general patrol members cannot. Updates of files come through the records queue, where records staff members make the necessary updates. If an experienced records person disagrees with a code, they will change it, sometimes in consultation with the member. Inexperienced records staff will often leave the file as the member submitted it17. 4.6 Quality control Quality control in all the research sites is done by civilian records staff, who review open files to ensure that the scoring is correct, complete, and adheres to policy. Therefore, if the allegation is classified as unfounded and the file is closed early in the process, it may not make its way to records and be subject to quality control at the local level. At the time of the research, the VPD did not have the staff to read all files to check offence codes, although key informants indicated that sexual assault files would be reviewed by a police supervisor. In general, researchers were told, records staff function as the first level of quality control and Statistics Canada or, for the RCMP, Informatics, function as the second level. At the national level, error reports are generated automatically by Statistics Canada, or, in the case of the RCMP, by Informatics, and sent back to the submitting jurisdiction. Errors are corrected at the local level and the information is then re- 15 The terms “classification” and “scoring” are used interchangeably in this report, depending on the context. These terms describe two parts of the same process of “classifying” a case in terms of its outcome (e.g., founded or unfounded) and “scoring” it by entering the appropriate code for statistical purposes into the paper file and/or into the electronic file that is submitted to Statistics Canada. 16 If police conclude a file without noting whether it is founded or not, civilian staff from the police records section will usually make that determination on the basis of the information in the file, with or without police consultation. In practice, therefore, there is sometimes an overlap between “concluding” a file and “classifying” a file, in that, for example, police may conclude a file by determining it to be unfounded. 17 It is interesting to note that in Vancouver, as a result of a recent procedural change, when a designation is changed from founded to unfounded, a records staff member will change “suspect” to “subject of complaint”, a designation that carries less stigma than “suspect”. Researchers were told that the RCMP do not always do this. 21 submitted to Statistics Canada or Informatics. In addition, Informatics runs a daily error report which also indicates the jurisdiction where the error originated. If a jurisdiction appears on the top ten of the national list of error reports, that jurisdiction is reportedly watched closely. 5.0 Research Findings and Discussion Due to the exploratory nature of this study, caution must be exercised in interpreting and generalizing the results. Nonetheless, the file review, combined with the key informant interviews and discussions, provide important insights into the founding process in sexual assault cases. They also provide valuable insights into the production of crime statistics generally, and sexual assault statistics in particular. These insights can be grouped under two headings: investigative issues and statistical scoring issues. 5.1 Investigative issues In spite of general agreement about the stringent requirements to be met before a case is determined to be unfounded, none of the key informants was surprised by the higher rate of sexual assault allegations that are classified as such, compared to other violent offences. It was generally felt that this is understandable, given the complexities inherent in sexual assault investigations. For instance, key respondents suggested that victims, witnesses or suspects may not understand what legally constitutes sexual assault or consent to sexual activity. What one person considers a clear lack of consent may not be considered as such by another. Further, it was noted that drug and/or alcohol use can compromise a victim’s ability to communicate a lack of consent as well as her memory of events. Key informants also emphasized that there are generally no witnesses to a sexual assault, resulting in “he said – she said” situations. While police acknowledged that suspects may be motivated to lie or to “misunderstand” a sexual situation, it was felt that some victims may as well. A number of possible reasons for false allegations were identified, including their use as a “tool” to get back at someone, to win a custody battle, to seek attention, or to extricate oneself from a difficult personal situation, such as the revelation of pre- or extra-marital sexual encounters, particularly in a cultural context where premarital sex is forbidden. According to a senior police member, the serious ramifications of wrongful convictions underscore the importance of thorough police investigations, including a thorough assessment of the victim’s and suspect’s credibility. There were not enough cases in the study sample where investigators believed allegations to be “maliciously” false, fabricated by a victim to extricate herself from a difficult situation or falsely alleged in the context of a custody dispute, to draw distinct conclusions about these cases. However, comments about false allegations included in the files as well as those made by informants indicate that, while there are undoubtedly cases where false allegations of sexual assault are made for these reasons, there are also risks of preconceptions in these areas. 22 A number of factors were found to be statistically associated with the classification of sexual assault allegations as either founded or unfounded18. These include whether the victim and suspect were strangers, whether mental health issues or disabilities, the use of force, the victim’s physical resistance, the fact that she had said “no” or the fact that she appeared to be upset were noted in the file, as well as formality of the police investigation. The most interesting results emerged where these statistically significant associations could be compared across jurisdictions with high and low unfounded rates, and across the VPD and the RCMP. 5.1.1 Whether the victim and suspect were strangers It is widely acknowledged that most sexual assault victims and offenders know each other (Kong et al, 2003). The authors’ own literature review indicated that such cases are more likely to be classified as unfounded, whereas cases where victim and suspect were strangers are more likely to be classified as founded (Du Mont and Myhr, 2000; Gregory and Lees, 1996; Tomlinson, 1996). Table 1: Whether victim and suspect were strangers: victims of all ages Number of valid cases = 132 Founding decision Founded Unfounded Total Victim and suspect knew each other Victim and suspect were strangers Number 41 24 Percentage 41.8% 70.6% Number 57 10 Percentage 58.2% 29.4% Number 98 34 Percentage 100% 100% Similarly, for the sample composed of victims of all ages, Table 1 shows that when the victim and suspect were strangers, significantly more cases were classified as founded (70.6%) than unfounded (29.4%). However, when the victim and suspect knew each other, significantly more cases were classified as unfounded (58.2%) than founded (41.8%). According to Table 2, for the sample of victims aged 14 years and over, it was three times more likely that a case would be classified as founded (75.9%) than unfounded (24.1%) when victim and suspect were strangers. Further, this pattern was generally consistent across jurisdictions with high and low unfounded rates. 18 Factors were considered to be statistically associated when levels of association between the variables reached .05, the standard level of significance. 23 Table 2: Whether victim and suspect were strangers: victims 14 years and over19 Number of valid cases = 90 Funding decision Founded Unfounded Total Victim and suspect knew each other Victim and suspect were strangers Number 26 22 Percentage 42.6% 75.9% Number 35 7 Percentage 57.4% 24.1% Number 61 29 Percentage 100% 100% Similarly, police key informants stated that, in their view, it is much less likely that a sexual assault allegation involving strangers will be classified as unfounded. 5.1.2 Whether there were mental health issues/disabilities Victims’ mental health issues or disabilities have been identified in the literature as a factor contributing to higher unfounded rates (Du Mont and Myhr, 2000; Gregory and Lees, 1996; Ekos Research 1988; Stanley, 1985). The analysis also revealed that victims’ (or, where victims were children, their parents’) mental health issues or disabilities were, in some jurisdictions, statistically associated with higher rates of unfounded determinations. 19 As noted, the age of 14 years is, for most purposes, the age of consent. 24 Table 3: Victim’s/parent’s mental health/mental disability issues: victims of all ages in Chilliwack and Langley (jurisdictions with relatively high unfounded rates) Number of valid cases = 74 Founding decision Founded Unfounded Total Complainant’s mental health issues or mental disability noted in file20 No indication that complainant had mental health issues or mental disability Number 4 30 Percentage 21.1% 54.5% Number 15 25 Percentage 78.9% 45.5% Number 19 55 Percentage 100.0% 100.0% According to Table 3, when it was noted in the file that victims or their parents had a mental health issue or disability, nearly four times as many incidents were classified as unfounded (78.9%) than founded (21.1%) in jurisdictions with high unfounded rates. For those cases where no such mental health issue or disability was noted in the file, 9% more cases were classified as founded (54.5%) than unfounded (45.5%). 20 Consistent with VICLAS, alcohol and drug abuse was considered to be a mental health issue. In many cases, substance abuse was combined with other mental health issues. 25 Table 4: Victim’s/parent’s mental health/mental disability issues: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) Number of valid cases = 106 Founding decision Founded Unfounded Total Complainant’s mental health issues or mental disability noted in file21 No indication that complainant had mental health issues or mental disability Number 6 43 Percentage 21.4% 55.1% Number 22 35 Percentage 78.6% 44.9% Number 28 78 Percentage 100.0% 100.0% The findings were similar for the three RCMP detachments as a whole. As shown by Table 4, when it was noted in the file that victims or their parents had a mental health issue or disability, cases were nearly four times more likely to be classified as unfounded (78.6%) than founded (21.4%)22. For those cases where no such mental health issue or disability was noted, approximately 10% more cases were classified as founded (55.1%) than unfounded (44.9%). While incidents involving a victim or parent with mental health issues or disabilities were much more likely to be classified as unfounded in jurisdictions with comparatively high unfounded rates and the three RCMP detachments, there was no such association in jurisdictions with comparatively low unfounded rates, including the VPD alone. Clearly, a complainant’s mental health issues or disabilities can make it more difficult for police to determine whether a sexual assault has taken place or not. However, the findings suggest that there is no intrinsic reason why such issues should affect the founding decision. It may be that they are less likely to do so where police members better understand that affected individuals are particularly vulnerable to victimization in general, and to sexual assault in particular, and where they are adequately trained to interview such complainants. 21 Consistent with VICLAS, alcohol and drug abuse was considered to be a mental health issue. In many cases, substance abuse was combined with other mental health issues. 22 Note that the percentages in this column are almost identical to those for Chilliwack and Langley (see Table 3), where unfounded rates were significantly higher than they were in Richmond. 26 5.1.3 Whether force was used 23 Research shows that sexual assault most often does not involve the use of physical force. Nonetheless, both the use of force and the seriousness of the alleged assault are identified in the literature as factors associated with the classification of sexual assault cases as founded (Tomlinson, 1996; Ekos Research, 1988). Table 5: Whether force was used: victims of all ages in Chilliwack and Langley (jurisdictions with relatively high unfounded rates) Number of valid cases = 74 Founding decision Founded Unfounded Total Use of physical force noted in file No use of force noted in file Number 11 23 Percentage 73.3% 39.0% Number 4 36 Percentage 26.7% 61.0% Number 15 59 Percentage 100.0% 100.0% Table 5 indicates that in jurisdictions with high unfounded rates, there was a greater likelihood that a sexual assault allegation would be classified as founded if the file noted that the suspect used physical force. In such cases, 73.3% were classified as founded and 26.7% were classified as unfounded. When no physical force was noted, 39% of cases were classified as founded and 61% were classified as unfounded. Table 6: Whether force was used: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) Number of valid cases = 106 Founding decision Founded Unfounded Total Use of physical force noted in file No use of force noted in file Number 14 35 Percentage 70.0% 40.7% Number 6 51 Percentage 30.0% 59.3% Number 20 86 Percentage 100.0% 100.0% 23 In the context of this report, the word “force” includes only those situations where the accused employs strength or power when touching the victim. 27 Table 6 reveals a similar pattern in the three RCMP detachments as a whole, where it was also found that sexual assault allegations were more likely to be classified as founded if the file noted that the suspect had used physical force. In fact, when the use of force was noted, 70% of the cases were classified as founded and 30% were classified as unfounded. When no use of force was noted, 40.7% of cases were classified as founded and 59.3% were classified as unfounded. Whether or not the suspect’s use of force was noted in the file did not affect the founding decision to a statistically significant degree in jurisdictions with low unfounded rates, including the VPD sample. Again, these findings suggest that there is no intrinsic reason why the suspect’s use of force should affect the founding decision. This factor might be less likely to contribute to an unfounded classification where police fully understand the nature and dynamics of sexual assault. 5.1.4 Whether the victim physically resisted Like the use of force, a victim’s physical resistance to a sexual assault may be viewed by some as one of the few responses that truly communicates no consent to sexual activity. In fact, the victim’s lack of physical resistance to a sexual assault was cited in the literature as a factor contributing to higher unfounded rates (Du Mont and Myhr, 2000; Ekos, 1988). Table 7: Whether the victim physically resisted: victims 14 years and over Number of valid cases = 102 Founding decision Founded Unfounded Total Noted in file that victim physically resisted Not noted in file that victim physically resisted Number 18 33 Percentage 85.7% 40.7% Number 3 48 Percentage 14.3% 59.3% Number 21 81 Percentage 100.0% 100.0% Table 7 includes files where the reference to physical resistance was based solely on the victim’s word and files where the victim’s account was corroborated by other evidence. Where the victim’s physical resistance was noted in the file, 85.7% of cases were classified as founded and only 14.3% were classified as unfounded. Where there was no mention of physical resistance, 40.7% of cases were classified as founded and 59.3% were classified as unfounded. 28 It is the reference to the victim’s physical resistance that shows the most marked impact on the founding decision, resulting in significantly higher rates of cases being classified as founded. This finding is consistent with what one would expect, as physical resistance to an assault is a very clear statement of lack of consent. There were too few victims who physically resisted to allow comparisons between those sites with high and low unfounded rates, and between the RCMP and Vancouver. There is no doubt that physical resistance makes it clear that a victim is not consenting to sexual activity. However, it is well recognized that there are reasons why a victim may not physically resist and that there are other non-physical means of resistance. It is interesting to note that in cases where the victim did not physically resist because she was asleep/unconscious/drugged/intoxicated, there was no significant difference in the founding rates. This finding indicates that the unconscious state of the victim was not used by police to invalidate the allegation of sexual assault. 5.1.5 Whether the victim said “no” 24 Like the use of force and a victim’s physical resistance, the importance afforded to a victim saying “no” to sexual activity may be part of a stereotypical view of the nature of consent and resistance. Obviously, saying “no” makes it clear that a victim is not consenting to the activity. However, just as there are many reasons why a victim may not physically resist, there are other ways of communicating nonconsent to sexual activity. Table 8: Whether the victim said “no”: victims 14 years and over in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) Number of valid files = 66 Founding decision Founded Unfounded Total Noted in file that victim said “no” Not noted in file that victim said “no” Number 24 7 Percentage 68.6% 22.6% Number 11 24 Percentage 31.4% 77.4% Number 31 35 Percentage 100.0% 100.0% 24 This analysis is based on cases where the victim was 14 years of age or over since consent to sexual activity would generally not be an issue for victims under 14. Further, it includes cases where the reference to saying “no” was based on the victim’s word only, as well as cases where the victim’s account was corroborated by other evidence. 29 Table 8 indicates that in the three RCMP detachments, saying “no” and having it noted in the file was more than twice as likely to result in an allegation being classified as founded (68.6%) than unfounded (31.4%). For those files where there is no documentation that a victim said “no”, the allegation was classified as founded in 22.6% of cases and as unfounded in 77.4% of cases. While the statistical relationship between a victim saying “no” and classifying a case as founded was highly significant in the RCMP detachments, there was no such association in the VPD. In this site, the founding decision was apparently unrelated to whether or not the victim said “no”. In fact, VPD cases where there was no documentation in the file that the victim said “no” were almost evenly divided between founded and unfounded classifications. This difference may indicate a tendency among some investigators in the RCMP detachments to classify sexual assault cases as unfounded based on an insufficient understanding about the nature of sexual assault, consent, and victims’ resistance. It is also possible that the VPD’s specialized sex offences unit and training in the investigation of sexual assault may have enhanced members’ awareness and knowledge of these issues. Key informants agreed that it would never be appropriate to classify a case as unfounded primarily because a victim did not say “no”, without adequate investigation of other relevant factors. Police must have other reasons to doubt a victim’s credibility or her version of events. Victims can have valid reasons for not saying “no” to the suspect, and can communicate a lack of consent in other ways. Nonetheless, police members acknowledged that the fact that a victim clearly said “no” to sexual activity makes the investigation into the issue of consent more straightforward. 5.1.6 Whether the victim appeared upset 25 The issue of whether or not a victim appeared upset also raises the issue of stereotypical notions about victims’ “appropriate” emotional reactions to sexual assault. Gunn and Minch (1988) found that police classification of sexual assault cases as unfounded was based, at least in part, on such stereotypical notions. 25 In order to bring the numbers up to a meaningful level, this section is based on cases where it was noted in the file that a victim appeared to be “upset”, “hysterical” or “out of control”. 30 Table 9: Whether the victim appeared upset or not: victims of all ages Number of valid files = 41 Founding decision Founded Unfounded Total Noted in file that victim did not appear to be upset Noted in file that victim appeared Number 3 18 Percentage 25.0% 62.1% Number 9 11 Percentage 75.0% 37.9% Number 12 29 Percentage 100.0% 100.0% to be upset Table 9 shows a strong association between documentation in the file that a victim appeared to be upset and a decision that the incident was founded. For files with such a notation, 62.1% were classified as founded and 37.1% as unfounded. Conversely, although the numbers are small, there was a strong relationship between documentation that a victim did not appear to be upset, and a decision that the incident was unfounded. For those with such a notation, 75% were classified as unfounded and 25% as unfounded. Table 10: Whether the victim appeared upset or not: victims 14 years and over Number of valid files = 32 Founding decision Founded Unfounded Total Noted in file that victim did not appear to be upset Noted in file that victim appeared Number 2 16 Percentage 25.0% 66.7% Number 6 8 Percentage 75.0% 33.3% Number 8 24 Percentage 100.0% 100.0% 31 to be upset Although the numbers were smaller still, Table 10 indicates that this pattern held true when the analysis was based only on those victims aged 14 years or over. While the numbers were still smaller, the pattern held true when the sample was divided between jurisdictions with high unfounded rates (Chilliwack and Langley) and those with low unfounded rates (Richmond and Vancouver), as well as when the sample was divided between Vancouver and the RCMP. However, because of the small numbers, the association was statistically significant only in the sample that included Chilliwack and Langley. There was also a strong pattern, although still not one that achieved statistical significance, in the three RCMP detachments. While it may be that a victim’s reaction after an assault is an indication of whether or not it occurred, especially where other factors cast doubt on her credibility, it is well known that sexual assault victims may exhibit a wide range of reactions to the traumatic experience – from openly expressive of strong emotion to deceptively calm. It would appear that this dynamic is less well understood in RCMP detachments than in the VPD. However, police key informants from all jurisdictions appeared to be very cognizant of this fact. 5.1.7 Formality of the police investigation The component of the police response that showed a significant association with the founding decision was the nature and extent of investigative materials included in the file. These ranged from police accounts of what they were told, to written statements from victims, witnesses, and suspects and audio or videotaped interviews with them. For the purposes of this analysis, police accounts only were compared to all other types of evidence, based on the assumption that the latter indicate a more formal approach to obtaining evidence than the former26. The following tables consider the relationship between the founding decision and the nature of victim, witness, and suspect-generated information contained in the files. With some exceptions, particularly with respect to Vancouver, the general finding was that the more formal the information in the police files, the more likely the case was to be classified as founded. 26 The numbers for other investigative tools such as photographs, physical evidence, DNA evidence and polygraph tests were not sufficiently high for statistical analysis. This finding may be suggestive, in itself, of a lack of formality in the investigation of some sexual assault cases. 32 Victim-generated information Table 11: Victim-generated evidence in the file: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) Number of valid files = 95 Founding decision Founded Unfounded Total Police account of victim’s narrative only Formal victim interview or statement Number 5 41 Percentage 17.2% 62.1% Number 24 25 Percentage 82.8% 37.9% Number 29 66 Percentage 100.0% 100.0% Based on Table 11, RCMP cases were much more likely to be classified as unfounded (82.8%) than founded (17.2%) when only the police account of the victim’s narrative was in the file. However, when other records of victims’ statements were included, cases were more likely to be classified as founded (62.1%) than unfounded (37.9%). Interestingly, the pattern was reversed for the VPD, but the relationship was not statistically significant. It should be noted that considerably more formal information was obtained from victims than from witnesses or suspects in all the sites included in the study. This may be indicative of an over-reliance on victim-generated evidence or a pre-disposition to focus the success or failure of the investigation on the victim rather than on the evidence as a whole. Witness-generated information Table 12: Witness-generated evidence in the file: victims of all ages in Chilliwack, Langley, and Richmond (the RCMP jurisdictions) Number of valid files = 71 Founding decision Founded Unfounded Police account of witness statements only Formal witness interviews or statements Number 7 29 Percentage 28.0% 63.0% Number 18 17 33 Total Percentage 72.0% 37.0% Number 25 46 Percentage 100.0% 100.0% The data in Table 12 indicate that RCMP files were approximately two and a half times more likely to be classified as unfounded (72.0%) than founded (28.0%) when only the police account of witness’s narratives was in the file. When other records were included, such as witness’s written statements or audio or video-taped statements, the case was much more likely to be classified as founded (63.0%) than unfounded (37.0%). While the pattern remains the same for the VPD sample, where the total number is the smallest, the association is not statistically significant. Suspect-generated information Table 13: Suspect-generated evidence in the file: victims of all ages Number of valid files = 77 Founding decision Founded Unfounded Total Police account of suspect’s narrative only Formal suspect interview or statement Number 15 23 Percentage 36.6% 63.9% Number 26 13 Percentage 63.4% 36.1% Number 41 36 Percentage 100.0% 100.0% Again, a similar pattern is evident from the data on suspect information in police files for the sample as a whole. As indicated in Table 13, files were much more likely to be classified as unfounded (63.4%) than founded (36.6%) when only the police accounts of suspects’ narratives were in the file. Conversely, when additional records of suspects’ statements were included, such as audio or video-tapes or written statements from the suspect, cases were much more likely to be classified as founded (63.9%) than unfounded (36.1%). This pattern of statistically significant association held true for the three RCMP detachments and for the Richmond-Vancouver sample, where unfounded rates were low. However, it did not hold for the VPD sample alone, where unfounded rates were also low; however, this finding cannot be considered reliable since it is based on only 21 cases. 34 Thus, for the sample as a whole, sexual assault allegations were more likely to be classified as founded when there was an indication in the file of a formal interview with or statement from the victim, witness, or suspect than when there was no such indication. However, limited numbers did not allow for further analysis of the relationship between the formality of police investigations and the founding rate. Whatever the potential impact of the formality of the investigation on founding decisions, most key informants agreed that, in general, effective sexual assault investigations should include more rather than less formal investigative strategies, and a serious attempt to interview any identified suspects. While there were substantial numbers of files in all jurisdictions where formal interviews or statements were not obtained from victims, witnesses, or suspects, it was with suspects that the lowest proportion of formal investigative strategies was undertaken. The frequent lack of contact or statement from suspects surprised some police members and not others. However, all agreed that it was a matter of concern. It was emphasized that police should always make serious attempts to contact and interview any suspect in a case, and document such attempts, along with reasons for failing to do so. Some police members went so far as to say that, unless there is other clear evidence that a sexual assault did not occur, it would not be appropriate to classify a case as unfounded without talking to the suspect. A number of reasons were suggested for why a suspect might not be interviewed by police. For instance, if a suspect refuses to come in to provide a statement, there is not much else that an investigator can do if there are insufficient grounds to arrest him. Further, if a warrant is issued against a suspect, he will not be taken to the police station to be interviewed, but directly to court. Finally, if there is strong evidence that no offence had taken place, there would be no need to interview a suspect. The file review revealed a number of cases where unfounded determinations were made without formal evidence from suspects. In some of these instances, it appeared to the researchers that a determination of unfounded was made without adequate investigation in terms of formal suspect interviews, and that a more appropriate classification might have been founded but not cleared27. In one case, the seven-year old daughter of a former sex trade worker was exhibiting sexualized behaviour in general, and with an elderly suspect in particular. While the child protection ministry was involved, the case was concluded as unfounded with no apparent attempt to interview the suspect. In another case, a number of neighbours alleged sexual abuse of a child by her father, primarily on the basis of the child’s sexualized behaviour and inappropriate sexual questions and comments. The father had unsupervised access to the child and a history of spousal abuse. The case was concluded as unfounded and was determined to be part of a custody battle. However, there was no evidence in the file that the allegation was part of a custody battle and no apparent attempt was made to interview the father. In a case involving an adult victim, a young woman woke up in the same bed as a male friend of a friend, to find him against her with an erection. The woman said that she had memories of him fondling her but had not given consent. The man argued it was consensual. Police concluded the file as unfounded 27 See 5.2.1 Confusion between unfounded and not cleared categories for a discussion of this issue. 35 because the victim and suspect had conflicting stories, although there was no formal statement from or interview with the suspect. A final case involved a deaf teenager who alleged inappropriate touching and suggestions (written in notes) by her employer. Her mother and grandmother thought she was misunderstanding his intentions. However, a child and youth worker reported to police. Police concluded the file as unfounded rather than founded but not cleared in spite of the fact that there had been two previous sexual assault allegations against the suspect from nieces, and that he refused to come in to make a formal statement. 5.2 Scoring files and other data reporting issues In addition to the investigative factors described above, the research findings raise a number of issues related to file scoring and data recording that may relate to unfounded rates in sexual assault cases. These include the confusion between the unfounded and the founded but not cleared categories; the confusion between the unfounded and the RCMP unsubstantiated categories; the inconsistent use of VICLAS for “unfounded” sexual assault cases; the inconsistent use of codes for spousal violence; the deficiencies of quality control; and the lack of national data on sexual assault cases classified as unfounded. 5.2.1 Confusion between unfounded and founded but not cleared categories The researchers observed some confusion in the files between the unfounded and founded but not cleared categories. Almost all key informants understood that cases should be classified as unfounded only where there is evidence that an offence did not occur or was not attempted. Key informants also agreed that where there is evidence that an offence took place, but not enough evidence to proceed, cases should be classified as founded but not cleared. However, such understanding did not always appear to guide practice. In terms of the determination that a case is either unfounded or founded but not cleared, key informants agreed that clarity exists at both ends of the continuum. Some cases may be clearly unfounded, such as when credible witnesses or physical evidence contradict the victim’s allegations, or when the facts as stated do not constitute a criminal offence. Other cases are clearly founded but not cleared, such as when there is strong evidence that an offence occurred but no suspect can be identified. However, respondents also acknowledged that there is a substantial “grey area” in between these two situations, particularly in cases where there is insufficient evidence to determine whether or not an offence occurred or was attempted, or where the police face significant challenges in obtaining evidence from a victim. They agreed that the cases in this “grey area” are problematic. Nonetheless, they also agreed that “you can’t jump from unsure to unfounded” without clear evidence that an offence did not occur. Thus, as informants explained, a case involving a “he said – she said” scenario should be classified as founded but not cleared in the absence of additional evidence28. 28 Confusion between a case that is legitimately unfounded and a case that is likely founded but shows little likelihood of a conviction due to insufficient evidence is also highlighted in the literature as a reason for high 36 Furthermore, in the VPD sample, the research revealed that sexual assault cases with insufficient evidence to determine whether or not an offence had occurred were sometimes deliberately classified as unfounded in order to protect suspects from the repercussions of an allegation that may (or may not) be unwarranted. The motivation to avoid the unnecessary stigmatization of suspects is understandable. However, it could be argued that it is also important to avoid the potential stigmatization of victims whose allegations are classified as unfounded on the basis of insufficient evidence, rather than being left as founded but not cleared. The file review uncovered a number of cases that concretely illustrate the confusion between, and inconsistent use of the unfounded and founded but not cleared categories. Such cases included, but were not limited to, those where the victim was a young child or had mental health problems or communication difficulties. While there appeared to be insufficient evidence to proceed with charges or a low likelihood of conviction, there did not appear to be clear evidence that no offence had occurred or been attempted. Nonetheless, many of these cases were classified as unfounded, rather than founded but not cleared. For instance, there were several allegations of child sexual abuse by the father or another adult that, in the absence of sufficient evidence to proceed with charges, were classified as unfounded either without contact with the suspect or with only informal contact. In some of these cases, there was sufficient concern to prompt a comment in the file that the child was not in danger because the child would not be alone with the suspect, or that the suspect should not continue to see the children. In one such case, a nine-year old girl was overheard saying that her father was “mean” and rubbed his hard penis against her. A year later, the girl told her mother that her dad puts his hand down her pants. The mother reported to child protection authorities who, in turn, reported to police. The child told police that her dad puts his hand down the back of her pants, on top of her underwear, “just to bug her.” Without interviewing the father, police decided that the case was unfounded, with a comment in the file that the child was not in danger because she was never alone with the father. In another case, lifeguards observed inappropriate touching of three teen and pre-teen girls in a swimming pool. The suspect, the girls’ English tutor, denied inappropriate touching. Neither the girls nor their parents thought that anything untoward had occurred. However, the police continued to have concerns because the lifeguards were adamant that the touching was inappropriate. Police advised the suspect to discontinue seeing the girls and provided the same advice to the parents. Nevertheless, the case was classified as unfounded. In yet another case, conflicting information was obtained from both a mother and her ten-year old daughter about whether the mother’s partner had asked the girl to touch him. The girl initially said that he had, then recanted. The mother agreed that it was probably a misunderstanding. But the case was classified as unfounded without any contact with the suspect. unfounded rates in sexual assault cases (Gregory and Lees 1996; Johnson 1996; Tomlinson 1996; Muzychka 1991; Gunn and Minch 1988; Minch et al. 1987; Stanley 1985; Clark and Lewis 1977). A related consideration is the assumption that a victim’s reluctance to proceed through the justice system is an indication that the allegation is likely false (Gregory and Lees 1996; Tomlinson 1996; Muzychka 1991; Minch et al. 1987; Stanley 1985). 37 There were also several cases in which allegations of a child sexually abusing another child appear to have been classified inconsistently. In one case, a mother reported that her 12-year old daughter lay on top of an eight-year old boy and gave him a hickey. Both mothers played down the incident when it was followed up by police. Counselling referrals were made but the families were not cooperative. The case was reported to Statistics Canada as founded but not cleared. In another case, a mother reported to police that a seven-year old pushed her six-year old son down on the ground while they were playing and pulled his shorts down and started kissing him on his penis and bum. The case was also reported to Statistics Canada as founded but not cleared. Yet, in another case, an 11-year old disclosed to his teacher that his 12-year old cousin had performed oral sex and anal intercourse on him five or six times in a locked bedroom and told him not to tell his mother. The victim alleged that he said “no” but his cousin didn’t listen. The file was classified as unfounded because of the young ages of the victim and suspect, contrary to the preceding example, and because the incidents were considered to be of an “exploratory nature”. In another incident, after visiting with her 13-year old male cousin, a three-year old held her vaginal area and said it hurt and that she had touched the place where her cousin pees. She disclosed the same thing a couple of months later to her father’s girlfriend. The adults were interviewed, but the suspect was not, although he apparently denied any sexual activity with the three-year old. The police felt the family was doing the right thing by participating in family counselling, yet the case was classified as unfounded. There were also cases of inconsistency in case classification where the victim had a communication disability. Investigative practices varied for three cases in which the victim was deaf. In one case, caregivers of a deaf-blind woman suspected abuse because of the victim’s behaviour when she returned from visits with her parents. In spite of disagreements among staff about the accuracy of the observations of the victim’s behaviour, an extensive interview was conducted by a qualified expert, and the case was classified as founded but not cleared. In another case, a young deaf woman accused an elderly “senile” man in her apartment complex of making sexual advances towards her without her consent. She did not want the police to speak with him about his behaviour, but only to speak with the building manager. While the case was one that apparently could not proceed with charges, it was classified as unfounded rather than founded but not cleared. Further, there were examples of inconsistency in classifying cases involving adults where there was insufficient evidence to know if an assault had occurred. Some of these were classified as unfounded and others were classified as founded but not cleared. In one case, hospital staff called police about a disturbed patient who did not fully recall what had happened but alleged that she was drugged and sexually assaulted over several days and had been unable to get away. She said she wanted to talk to police later, but did not return phone calls from the police. The case was classified as unfounded. In another case, the victim left a house after an argument with her boyfriend. When the boyfriend left half an hour later, he heard her screaming, and found her in the street, hysterical, with torn pants. 38 When the police came the victim was in a state of “psychosis” and unable to give details. At the hospital, she denied being sexually assaulted. The police noted that “in view of the circumstances, this file will be concluded at this time. The actions of the female are not consistent with her having been a victim of a crime.” Yet the case was classified as founded but not cleared. Finally, there were several cases where the victim and suspect had been drinking and/or taking drugs and the victim went to sleep or passed out, only to wake with the suspect performing a sexual act on her, or to find evidence that a sexual act had been performed. Most of these cases were classified as unfounded, on the basis of inconsistencies in the evidence given by the victim and the suspect. 5.2.2 Confusion between unfounded and RCMP unsubstantiated categories In RCMP sites, cases can be classified as unsubstantiated where there is no evidence to confirm whether or not an offence has been committed. Such cases are not reportable to Statistics Canada. Some key RCMP informants were unclear about the difference between the “unfounded” and “unsubstantiated” designations, or even unaware that the latter exists. A senior police officer suggested that a new Statistics Canada category of “insufficient evidence” would remove the need to decide whether a case is unfounded or founded but not cleared when there is not enough evidence to make that determination. Most records staff, however, felt that current Statistics Canada categories were sufficient. They suggested better written definitions with clearer criteria; training for staff who were doing the scoring; and more consistent monitoring for quality control. The file review uncovered a number of cases that concretely illustrate the confusion between, and inconsistent use of the unfounded and unsubstantiated categories. In fact, the latter was difficult to differentiate from unfounded cases, even for some of those RCMP members who were questioned about this distinction. When a mother’s boyfriend was arrested on an unrelated matter and removed from the home, the fiveyear old daughter told her mother that the suspect had touched her private parts and had her touch his private parts on a number of occasions. The suspect also reportedly danced and gyrated in front of the victim in his underwear. This case was considered unsubstantiated. In another case, the mother of a seven-year old girl told police that her daughter had dreamt about a naked male and that there were “several other little things” that may have indicated sexual abuse. However, there was no sexual assault incident, no disclosure, and no suspect. Yet this case was coded as unfounded. 5.2.3 Inconsistent use of VICLAS booklets The file review revealed some inconsistency as to whether or not a VICLAS booklet was completed when a sexual assault was classified as unfounded. In more than one case, a debate between members was recorded in the file regarding whether or not a booklet was required. Key informants also appeared unclear on this issue. Most members agreed, however, that it would be useful to have clearer guidelines regarding when such a booklet is required. 39 5.2.4 Inconsistent use of codes for spousal violence The file review and information from key informants also revealed inconsistencies in the use of supplementary spouse assault codes to capture sexual assault that occurs within the context of an intimate relationship, as well as a lack of quality control in the scoring of this factor. Such inconsistencies present a particular problem for sexual assault statistics, making it impossible to determine the incidence of sexual assault committed within an intimate relationship. Further, such inconsistencies make it difficult to assess what impact, if any, the fact that a sexual assault is committed within the context of a current or past intimate relationship has on the founding decision. 5.2.5 Deficiencies in quality control Several key informants noted that scoring errors were not uncommon and that quality control on scoring matters was deficient. However, their opinions were divided as to who was most qualified to make a decision about whether a case should be classified as unfounded – police members or records staff. Police respondents saw the founding decision as part of their investigative role, and a component of concluding a case. Records staff, on the other hand, viewed the decision as a scoring matter, based on the information contained in the file. Hence, the former tended to believe that only police, with their training and investigative experience, were qualified to determine whether a case should be classified as founded or unfounded. The latter tended to believe that they, with their training in offence scoring classifications, were better qualified to decide whether or not a case met the criteria for being unfounded. Whatever their opinions about who should make founding decisions, most key informants agreed that central to accurate classification and scoring of sexual assault allegations was good communication and collaboration between police members and records staff. Police members’ concluding remarks were seen as an effective strategy for promoting such communication. Several key informants – both members and records staff – emphasized the need for clear, concise remarks that summarize each file and document what was done and the reasons for closing the file and for a particular case classification or score. These remarks, key informants noted, would make quality control easier and more straightforward. 5.2.6 Lack of national data on unfounded determinations A crucial piece of information was provided by the last key informant to be interviewed over the course of this research. The key informant reported that Statistics Canada no longer requests unfounded data (on any offences) from police services, and no longer analyzes the data on unfounded cases that it still receives due to data quality issues. No other key informants appeared to be aware of this state of affairs, which clearly has very serious implications for future research and monitoring capabilities in the area of sexual assault. When clarification was sought from Statistics Canada, officials said that some police services across the country do not systematically collect data on the offences that they have classified as unfounded. While 40 other police services do, there are data quality problems with the data that is forwarded. Limited coverage, combined with poor quality data, would preclude a valid statistical analysis. 6.0 Conclusions and Implications of the Research Findings Once police classify a sexual assault allegation as unfounded, there is no further investigation of the complaint. If the classification is incorrect, the case will not receive the investigative attention it warrants. In addition, complaints that – rightly or wrongly – are classified as unfounded are excluded from national crime statistics, thus impacting on reported sexual assault levels and clearance rates, both of which are based on founded occurrences. As previously stated, higher unfounded rates for sexual assault compared to other violent offences, have long been a source of tension between police and individual victims, as well as between police and women’s advocates and service providers. Such tension can lead to an erosion of women’s confidence in the justice system which, in turn, can contribute to their reluctance to report sexual assault to the police. The results of this exploratory research provide important insights into the founding process and the factors that may contribute to higher unfounded rates for sexual assault. Many of the files reviewed over the course of this research indicated that police had conducted thorough investigations, with a demonstrated sensitivity to the nature and dynamics of sexual assault. However, the findings also highlight the need to enhance police understanding of these issues, and to promote more optimal approaches to investigating sexual assault allegations. Possible strategies for enhancing sexual assault investigations and increasing the accuracy and consistency of the scoring process, as well as possible directions for further research, are described below. 6.1 Implications for sexual assault investigations Increased police focus on sexual assault training as well as policy and operational guidelines may combine to foster a culture that reinforces an awareness and understanding of the nature and dynamics of sexual assault and of the importance of rigorous investigation of these cases in spite of challenges involved. The creation or enhancement of such a culture may, in turn, result in the allocation of specialized resources necessary to investigate these cases with maximum effectiveness. 6.1.1 Training Significant time, effort, and funds have been invested in providing training on both adult sexual assault and child sexual abuse to police by both the RCMP and municipal forces in BC. However, the complex nature of this offence requires ongoing and specialized training. The study findings suggest that such training should include: 41     enhanced training on the nature and dynamics of sexual assault, including the various forms it may take, its gendered nature, issues of consent and victim resistance, and the range of possible reactions to the trauma of victimization; enhanced training on optimal investigative procedures in sexual assault cases, including formal interviewing of victims, witnesses, and suspects, the dangers of over-reliance on victimgenerated information, and the requirements of an unfounded classification; enhanced training on the investigation of sexual offences involving very young children, including interviewing skills and optimal investigative strategies in these cases; and enhanced training on the greater susceptibility to victimization of those with mental health issues or disabilities, and on strategies and techniques for conducting investigations in these cases. 6.1.2 Policy and operational guidelines Study findings suggest that optimal sexual assault investigations would also be supported by:    the development of detailed sexual assault policy and operational guidelines for the RCMP to guide police practice in the investigation of adult sexual assault and child sexual abuse cases, and to address issues raised in this report, including the thoroughness of investigations, the dangers of over-reliance on victim-generated information, and the requirements of unfounded classifications; the review of the VPD’s sexual assault policy to ensure that it adequately covers the issues raised in this report, including the thoroughness of investigations, the dangers of overreliance on victim-generated information, and the requirements of unfounded classifications; improved quality control and accountability mechanisms to ensure that sexual assault investigations, including founding decisions, are conducted in the most thorough and effective manner possible. 6.1.3 Specialization Further, the study findings suggest that sexual assault investigations may also benefit from:  6.2 further sexual assault specialization in police jurisdictions that do not currently have a specialized sex offences unit, in order to reap the benefits of specialized training and experience, and to foster a culture of awareness of and sensitivity to sexual assault issues. Implications for statistical data collection With respect to the scoring of files, this research indicates a need for greater clarity and consistency in the classification of cases as founded or unfounded, particularly in relation to the distinction between the unfounded and founded but not cleared categories, and particularly when the issue is one of insufficient evidence. In addition, there is a need for more consistent use of the spousal assault supplementary code and of VICLAS booklets. 42 The findings of this research suggest that what is required is:     enhanced training and guidelines, with definitions, criteria, and instructive examples for records staff and police members, in order to promote their understanding of the requirements of each scoring category for case outcome; enhanced training and guidelines to be applied not only to the scoring of case outcomes, but to the utilization of spousal assault supplementary codes; for maximum effectiveness and enhanced communication and collaboration, training to be conducted jointly with records staff and police members; and effective quality control to ensure that the knowledge of those who are making classification decisions is consistently translated into practice. Overshadowing much of what is put forward in this report is the discovery that Statistics Canada no longer requests data on incidents that are classified as unfounded from police services, and no longer analyzes the data that it still receives due to data quality issues. This impact of this situation on Statistics Canada’s ability to analyze and report on trends in the criminal justice response to sexual assault cases in BC and across the country is far-reaching. 6.3 Implications for further research The implications of the discontinuation of systematic data-collection on “unfounded” cases, both for research on sexual assault and for monitoring of the criminal justice response to this offence are serious. Without national data on unfounded rates, there will be no way of monitoring trends in founded and unfounded rates, and no way of assessing the impact of any corrective measures that may be taken to ensure accuracy in decision-making regarding these determinations. In the absence of reliable, national data, the issues raised in this report can be explored only on a site-by-site basis. Ideally, future studies would be based on larger and/or more diverse samples in order to allow for more extensive statistical analysis and a higher degree of statistical reliability. A wider range of sites would allow greater exploration of the issues addressed in this report in a variety of jurisdictional, cultural and operational contexts. These could include the RCMP, provincial and municipal police forces, police forces with and without specialized sexual offence units, forces utilizing different training approaches, and samples consisting of policing jurisdictions that have more clearly and consistently high and low sexual assault unfounded rates over extended periods of time. The issues addressed in this report also require more extensive qualitative research, in order to elicit information and opinions from a larger number and broader range of key informants, including community-based victim service providers and advocates. Finally, given the differences between research sites highlighted in this report, it would be useful to better understand the factors that may impact on a police agency’s attitudes and responses to sexual assault, including factors that may contribute to the creation of a “culture” of responsiveness among police members29. 29 For example, Abbotsford Police Department, an independent police force in BC’s Fraser Valley, had 43 6.4 Implementation of the research findings: the way forward The goal of this research was to better understand police founding decisions in sexual assault cases, and why the unfounded rate for sexual assault is significantly higher than for other violent offences. The hope is that this increased understanding will lead to changes in investigative and offence classification or scoring practices where required. The collaborative approach adopted in this research can help to facilitate this goal of constructive change. Already, senior police members who reviewed and provided feedback on the first draft of this report have indicated that they are considering ways to further explore the research findings and to implement changes based on these findings. A presentation on the research and its preliminary findings was made to the Police Information and Statistics Committee of the Canadian Association of Chiefs of Police (POLIS)30, where the research results were received with interest. Another presentation was made to the provincial Working Group of the Community Coordination for Women’s Safety (CCWS) program of the BCASVACP31. This working group includes senior police members and other justice system personnel who may be in a position to effect change in their organizations. It also includes the Justice Institute of BC and other provincial nongovernmental organizations representing women – including Aboriginal women, immigrant women and women with disabilities – who face particular barriers accessing the justice system. The open, constructive responses to these findings from police members, from records staff, and from the women and victim-serving community have been encouraging. While there is a need to further explore the issues raised in this report, there are also clear indications of areas that need immediate attention. It is our hope that collaborative efforts to address these areas will continue, using the research as a springboard from which to launch further initiatives at the local, provincial, and national levels. sexual assault unfounded rates from 1999 to 2003 that were comparable to the VPD’s, and consistently lower than other BC Lower Mainland jurisdictions during that period. While Abbotsford does not have a specialized sex offences unit, it has a history of progressive responses to violence against women, an active violence against women coordinating committee, proactive police-based and community-based victim services with a reputation for working together collaboratively, and a police chief with a reputation for proactive response to violence against women. 30 More information on POLIS can be found at http://www.cacp.ca/. 31 More information on BCASVACP and CCWS can be found at: http://www.endingviolence.org 44 Appendix I List of Factors Examined in the Literature Complainant Factors Complainants’ moral character, reputation, behaviour, and credibility were highlighted in all the literature reviewed on this topic, which addresses the following factors:                     age (Ekos Research, 1988; Stanley, 1985) gender (Ekos Research, 1988; Stanley, 1985) marital status (Ekos Research, 1988; Stanley, 1985) nationality (Ekos Research, 1988; Stanley, 1985) occupation/socio-economic background (Muzychka, 1991; Stanley, 1985) employment status (Ekos Research, 1988) perceived promiscuity, presumed or actual prior sexual history (Federal/Provincial/Territorial Working Group of Attorneys General Officials on Gender Equality in the Canadian Justice System, 1992; Muzychka, 1991) prior criminal or juvenile record (Ekos Research, 1988) alcohol or drug use/intoxicated or drugged state at the time of the offence (Law Society of British Columbia 1992; Muzychka, 1991; Ekos Research, 1988; Minch et al. 1987; Stanley, 1985) mental or physical disabilities (Law Society of British Columbia, 1992; Ekos Research, 1988)32 perceived “mental instability” or mental health history of complainant (Gregory and Lees 1996; Muzychka, 1991; Stanley, 1985)33 whether the victim resisted the assailant (Gunn and Linden, 1991, 1993; Ekos Research, 1988; Minch et al, 1987)34 the victim’s appearance or dress at time of assault (Law Society of British Columbia, 1992) whether she is a “respectable” victim assaulted under “acceptable” conditions (Muzychka, 1991) whether the victim violated traditional sex role norms (Tomlinson, 1996) housing status (Muzychka, 1991, found that police were more likely to doubt the veracity of a report from a complainant who came from a transition house or was homeless) status as a parent (Muzychka, 1991, found that single mothers were more likely to be doubted) previous complaints (Musychka, 1991, found that police were more likely to doubt the veracity of a report from a complainant who had made previous complaints) time that elapsed between the incident and the first complaint (Gunn and Linden, 1991, 1993; Law Society of British Columbia, 1992; Stanley, 1985) to whom the complainant first reported the assault (Ekos Research, 1988 found that assaults first reported to social agencies were less likely to be founded than those reported to police, family members or friends) 32 This factor was positively associated with a lower unfounded rate in the current study. This factor was positively associated with a lower unfounded rate in the current study. 34 This factor was positively associated with a lower unfounded rate in the current study. 33 45     complainants’ emotional response to the assault (Muyzchka, 1991; Gunn and Minch, 1988; Stanley, 1985)35 complainants’ reluctance to proceed (Gregory and Lees, 1996; Musychka, 1991; Minch et al, 1987) whether the victim had had a medical examination (Gunn and Linden, 1991, 1993; Stanley, 1985) whether the victim had destroyed evidence (Tomlinson, 1996). Suspect Factors Suspect factors include:     suspect’s prior criminal record (Gunn and Linden, 1991; Ekos Research, 1988) alcohol or drug use/intoxicated or drugged state at the time of the offence (Stanley, 1985) socio-demographic details of the suspect (Gunn and Linden, 1991, 1993) character and reputation of the suspect (Law Society of British Columbia, 1992). Relationship between Complainant and Suspect Relationship between the complainant and the suspect (Ekos Research, 1988; Gregory and Lees, 1996; Gunn and Linden, 1991, 1993; Stanley, 1985; Tomlinson, 1996) includes situations where the complainant and offender are: strangers; current or former intimate partners; acquaintances with whom there was previous sexual contact; or acquaintances with whom there was no previous sexual contact36. Characteristics of the Alleged Offence Characteristics of the alleged offence include:        nature and seriousness of the offence(s) committed (Gunn and Linden, 1991; Ekos Research, 1988) number of suspects (Gunn and Linden, 1991, 1993) use of physical violence (Gunn and Linden, 1991, 1993; Ekos Research, 1988)37 use of a weapon (Tomlinson, 1996) physical injuries sustained by victim (Gunn and Linden, 1991, 1993) concurrent commission of other crimes by same offender(s) (Stanley, 1985) location of the offence (Gunn and Linden, 1991, 1993) and 35 Whether a file noted that a complainant appeared to be upset was positively associated with a higher founded rate in the current study. 36 In this study, cases were more likely to be classified as founded when the complainant and suspect were strangers. 37 In this study, cases were more likely to be classified as founded when it was noted in the file that the suspect used force. 46  time of the offence (Gunn and Linden, 1991, 1993). Quality and Consistency of Information Quality and consistency of information obtained includes:    insufficient evidence or complainants’ failure to substantiate the allegation (Gregory and Lees, 1996; Ekos Research, 1988)38 complainant-initiated withdrawals (Gregory and Lees, 1996 reported that these were more likely to occur where there was some degree of prior intimacy or acquaintance between complainant and suspect) corroborative or physical evidence for either complainant or accused, including injuries, medical-forensic evidence, torn clothing, other witnesses (Ekos Research, 1988; Gregory and Lees, 1996; Law Society of British Columbia, 1992; Muzychka, 1991; Ruebsaat, 1985; Tomlinson, 1996). Expectations of Crown Counsel and the Courts Expectations of responses of Crown, defence and judges (Gregory and Lees, 1996; Johnson, 1996; Muzychka, 1991; Gunn and Minch, 1988; Minch et al, 1987) include:       expectations of what Crown would consider to be sufficient and credible evidence for a likelihood of conviction and therefore what they would be prepared to charge expectations of what judges would consider to be sufficient and credible evidence for conviction expectations of what defence lawyer tactics might be and how the witness (and the case) would stand up to cross-examination perceptions of what constitutes the responsibility of justice system players to consider the efficiency of the overall system in making their decisions about whether to proceed with a case or not presence of circumstances which might be viewed with disfavour by the court, including, for example, “inappropriate female behaviour” (Tomlinson, 1996) complainants’ reluctance to proceed (complainants considered to be “uncooperative” or “apprehensive”) was a reason for classifying cases as unfounded (Minch et al, 1987). 38 This issue relates to the current study’s finding that there is confusion between unsubstantiated and unfounded determinations and between unfounded and founded but not cleared classifications. 47 References Ad Hoc Federal-Provincial-Territorial Working Group Reviewing Spousal Abuse Policies and Legislation. 2003. Final Report of the Ad Hoc Federal-Provincial-Territorial Working Group Reviewing Spousal Abuse Policies and Legislation. Ottawa, ON: Department of Justice Canada. Alberta Justice. 2005. Domestic Violence Handbook for Police and Crown Prosecutors in Alberta. Edmonton, AB: Author. Clark, L. and Lewis, D. 1977. Rape: The Price of Coercive Sexuality. Toronto, ON: The Women’s Press. Clark, S. and Hepworth, D. 1994. Effects of Reform Legislation on the Processing of Sexual Assault Cases. In Julian Roberts and Renate Mohr (Eds.) Confronting Sexual Assault: A Decade of Legal and Social Change (pp.113-135). Toronto, ON: University of Toronto Press. Dionne, M. 1997. Voices of Women Not Heard. The Bernardo Investigation Review: Report of Mr. Justice Archie Campbell. Canadian Journal of Women and the Law, 9, 394-417. Du Mont, J. and Myhr, T. L. 2000. So Few Convictions. The role of Client-Related Characteristics in the Legal Processing of Sexual Assaults. Violence Against Women, 6, 1109 – 1136. Ekos Research Associates Inc. 1988. Report on the Treatment of Sexual Assault Cases in Vancouver, British Columbia. Ottawa, ON: Research and Statistics Division, Department of Justice Canada. Fazlollah, M. 2001a. National Rape Statistics Highly Suspect. Philadelphia: Womensenews. Fazlollah, M. 2001b. Women Activists Monitor Philadelphia Rape Squad. Philadelphia: Womensenews. Fazlollah, M. 2000. Police May Ignore Rape Complaints to Boost Image. Philadelphia: Womensenews. Federal/Provincial/Territorial Working Group of Attorneys General Officials on Gender Equality in the Canadian Justice System. 1992. Gender Equality in the Canadian Justice System, Ottawa, ON: Department of Justice Canada. Gannon, M. and Mihorean, K. 2005. Criminal Victimization in Canada, 2004. Juristat, Canadian Centre for Justice Statistics, 25:7, Ottawa, ON: Statistics Canada. Gregory, G. and Lees, S. 1996. Attrition in Rape and Sexual Assault Cases. The British Journal of Criminology, 36, 1-17. Gunn, R. and Linden, R. 1991. Factors Affecting the Disposition of Sexual Assault Cases Before and After a Change in Sexual Assault Laws. The Canadian Journal of Program Evaluation, 6, 71-82. Gunn, R. and Linden, R. 1993. Evaluating the Impact of Legal Reform: Canada’s New Sexual Assault Laws. In Julian V. Roberts and Joe Hudson (Eds.) Evaluating Justice (pp. 135–154). Toronto, ON: Thompson Education Publishing, Inc. Gunn, R. and Minch, C. 1988. Sexual Assault: The Dilemma of Disclosure, The Question of Conviction. Winnipeg, MAN: University of Manitoba Press. Hattem, T. 2000. Survey of Sexual Assault Survivors. Ottawa, ON: Research and Statistics Division, Department of Justice Canada. Johnson, H. 1996. Dangerous Domains. Toronto, ON: Nelson Canada. 48 Kong, R., Johnson, H., Beattie, S. and Cardillo, A. 2003. Sexual Offences in Canada. Juristat, Canadian Centre for Justice Statistics, 23:6, Ottawa, ON: Canada. Lakeman, L. 2004. Canada’s Promises to Keep: The Charter and Violence Against Women. Vancouver, BC: Canadian Association of Sexual Assault Centres. Law Society of British Columbia. 1992. Gender Equality in the Justice System. Vancouver, BC: Law Society of British Columbia. Light, L. and Ruebsaat, G. 2005. Police Classification of Sexual Assault Cases as Unfounded. Targeted Literature Review and Annotated Bibliography. Vancouver, BC: unpublished. Manitoba Police Commission. 1988. Protocol for the Police Handling of Sexual Assault Complainants. Manitoba: unpublished. Minch, C., Linden, R., and Johnson, S. 1987. Attrition in the Processing of Rape Cases. Canadian Journal of Criminology, 29, 389-404. Ministry of Public Safety and Solicitor General. 2005. Police and Crime Summary Statistics, 1995 – 2004. Police Services Branch. Vancouver, BC: Author. Muzychka, M. 1991. Beyond Reasonable Doubt: The Influence of Victim Stereotypes and Social Biases on Police Response to Women’s Complaints of Sexual Assault. St. John’s, NFLD: Provincial Advisory Council on the Status of Women, Newfoundland and Labrador. Province of British Columbia. 1992. Is Anyone Listening? Report of the British Columbia Task Force on Family Violence. Ministry of Women’s Equality. Victoria, BC: Author. Province of British Columbia. 1997. Survey of Sexual Assaults Reported to Police in British Columbia 1993-1994. Police Services Division, Ministry of Attorney General. Victoria, BC: Author. Roberts, J. 1990. Sexual Assault Legislation in Canada: An Evaluation. An Analysis of National Statistics. Report No. 4. Programs and Research Section, Ottawa, ON: Department of Justice Canada. Roberts, J. 1994. Criminal Justice Processing of Sexual Assault Cases. Juristat Service Bulletin, Canadian Centre for Justice Statistics, 4:7 Ottawa, ON: Statistics Canada. Roberts, J. 1996. Sexual Assault in Canada: Recent Statistical Trends. Queen’s Law Journal, 21, 395-421. Roberts, J. and Mohr, R. M. 1994. Sexual Assault: Future Research Priorities. In Julian Roberts and Renate Mohr (Eds.) Confronting Sexual Assault: A Decade of Legal and Social Change (pp. 327338). Toronto, ON: University of Toronto Press. Roberts, J. and Gebotys, R.J. 1993. Evaluating Criminal Justice Legislation: Effects of Rape Reform in Canada. In Julian V. Roberts and Joe Hudson, (Eds.) Evaluating Justice (pp. 155-169). Toronto, ON: Thompson Educational Publishing, Inc. RCMP “E” Division. 2001. K. Sexual Offences. “E” Division Policy. Available from RCMP “E” Division. Ruebsaat, G. 1985. Sexual Assault Legislation in Canada: An Evaluation. The New Sexual Assault Offences: Emerging Legal Issues. Report No.2. Programs and Research Section, Ottawa, ON: Department of Justice, Canada. Stanley, M. G. 1985. Sexual Assault Legislation in Canada: An Evaluation. The Experience of the Rape Victim with the Criminal Justice System Prior to Bill C-127. Report No.1. Programs and Research Section, Ottawa, ON: Department of Justice, Canada. 49 Statistics Canada. 1993. The Violence Against Women Survey. The Daily. November 18. Ottawa, ON: Author. Statistics Canada. 1999. General Social Survey on Victimization. Ottawa, ON: Author. Tomlinson, D. 1996. An Examination of Sexual Assault Victims’ Experiences with the Criminal Justice System and Related Support Services in Alberta. Victims’ Programs’ Assistance Fund. Alberta: Unpublished report prepared for Alberta Ministry of Justice. Woods, J. 1998. Recommendations for Amendments to “E” Division R.C.M.P. Operational Policies Pertaining to Relationship Violence and the Processing of Firearms Applications. Royal Canadian Mounted Police. Case Law R. v. Darrach, (2000) 148 C.C.C. (3d) 97 (S.C.C.) R. v. Mills, (1999) 28 C.R. (5th) 207 (S.C.C.) R. v. O’Connor, (1995) 22 C.R. 273 (S.C.C.) R. v. Osolin, (1994) 86 C.C.C. (3d) 481 (S.C.C.) R. v. Seaboyer, (1991) 66 C.C.C. (3d) 321, (S.C.C.) 50