Attrition – Part 1

Commissioned by The Ministry of Women’s Affairs
Authors: Sue Triggs, Elaine Mossman, Jan Jordan and Venezia Kingi

 

For a PDF (1.70MB), click HERE

 

Return to Responding to sexual violence research reports

 

Contents

List of Boxes
List of Tables
List of Figures
Acknowledgements
Executive Summary

1    Introduction
1.1    Project overview
1.2    Scope and objectives of the attrition study
1.3    Rationale for the attrition study
1.4    Key issues identified by previous research
1.5    Legislation and policy in New Zealand
1.6    Structure of this report

2    Methodology
2.1    Terminology
2.2    Scope
2.3    Data source and extraction
2.4    A caution on comparisons with official statistics
2.5    Data limitations
2.6    Data analysis
2.7    Comparisons with other research

3    Profiles of victims, offenders and incidents
3.1    Victim profile
3.2    Offender profile
3.3    Crime profile
3.4    Summary of victim, offender and case characteristics

4    Attrition rates during the police investigation
4.1    Definitions
4.2    Overview of attrition rates during the investigation
4.3    Prosecution rate
4.4    Investigation outcome by offence
4.5    Victim characteristics by investigation outcome
4.6    Investigation outcome by police district
4.7    Suspect identified, but no charges laid
4.8    No suspect identified
4.9    ‘No offence’ and ‘false complaints’
4.10    Victim withdrawal
4.11    Attrition before recording
4.12    Summary of attrition during the police investigation

5    Attrition rates through the court process
5.1    Definitions
5.2    Overview of attrition rates during the court process
5.3    Conviction rates
5.4    Cases not convicted
5.5    Victim characteristics by prosecution outcome
5.6    Prosecution outcome by police district
5.7    Summary of attrition during the court process

6    Factors influencing outcomes
6.1    Overview
6.2    Approach
6.3    Factors tested
6.4    ‘False complaints’
6.5    No offence disclosed
6.6    Suspect identification rate
6.7    Suspect prosecution rate
6.8    Conviction rate
6.9    Summary of factors influencing attrition
6.10    Conclusion

Appendix A: Sexual offences included in this study
Appendix B: Literature reviewed
References

 

 

 

 

List of Boxes

Box 1: Definitions: study scope and methods
Box 2: Definitions: victim and offender characteristics

List of Tables

Table 1: Victim-offender relationship, where known, by age of victim (%)
Table 2: Ethnicity of offender, by victim’s ethnicity
Table 3: Force, threats and injury
Table 4: Alcohol and/or drug use and uncertainty of violation

List of Figures

Figure 1: Attrition rates for cases recorded as sexual violation
Figure 2: Age distribution of victims
Figure 3: Victim-offender relationship (where known)
Figure 4: Age distribution of offenders and victims
Figure 5: Time of the day the incident occurred

Acknowledgements

My thanks to the dedicated team of police investigators who spent many weeks extracting the data for this project. Thanks also to the project teams from the Ministry of Women’s Affairs, the New Zealand Police, and the Crime and Justice Research Centre at Victoria University of Wellington for project organisation, guidance, support and comments on the report. Finally, I would like to acknowledge and thank Professor Elizabeth Stanko, Metropolitan Police Service, London, for offering peer reviewer comments on the report in its final stages.

 

Executive summary

 

Aim and scope

The aim of this study was to assess attrition in relation to adult sexual violation cases – that is, what proportion and type of cases drop out at each stage of the criminal justice process. Understanding the rate of attrition and reasons for attrition at each stage of the process is a critical first step towards improving policy and practice such that attrition is minimised and outcomes for victims and society are improved.

The study was based on 1,955 police files coded as sexual violation of an adult victim and included all such offences recorded by the New Zealand Police from 1 July 2005 to 31 December 2007. All cases initially recorded as sexual violation were included, even if subsequent investigation determined that no offence had occurred. The research data set was a summarised extract of police data, so lacked the full extent and depth of information recorded in the original file. While key outcome variables were well recorded, many of the descriptive variables had significant missing data.

An adult was defined as a person aged 16 years or older at the time of the offence. The bulk of cases involved rape (68 percent) or unlawful sexual connection (22 percent), but the sample also included attempted sexual violation (7 percent) and selected offences relating to sexual violation or exploitation (3 percent).

The study estimated attrition within the criminal justice system only. Survey data indicate that around nine in ten sexual violation offences are not reported to the police. In addition, some offences reported as sexual violation may not be recorded as such. 

 

Attrition rates for cases recorded as sexual violation

The overall attrition rate is shown in Figure 1. Just over a third of cases were classified as ‘no offence’ (including 8 percent classified as ‘false complaints’), 11 percent did not proceed due to the lack of an identified suspect, 24 percent of cases had an identified suspect but no charges were laid, and in 18 percent of cases an offender was prosecuted but not convicted for sexual violation.

Therefore, the conviction rate for sexual violation was 13 percent based on all recorded cases (or 20 percent if ‘no offence’ cases were excluded from the base). However, some cases resulted in a conviction for a related offence, giving a conviction rate of 14 percent if all convictions for any sex offence were included or 16 percent including convictions for any offence. For individual sexual violation offences (as distinct from cases, which may include multiple offences) the conviction rate was 17 percent.

 

Figure 1: Attrition rates for cases recorded as sexual violation

 

 

Of cases in which an offender was prosecuted, 42 percent resulted in a conviction, 27 percent resulted in an acquittal, 14 percent were discharged by the judge and 17 percent were withdrawn by the prosecution.

The prosecution rate (percentage of cases with charges laid) was 31 percent based on all recorded cases or 46 percent if ‘no offence’ cases were excluded from the base. The prosecution rate for cases involving multiple offences was much higher than for single-offence cases and thus the prosecution rate based on recorded offences (49 percent) was higher than the prosecution rate for cases (31 percent).

Cases in which a known suspect was not charged tended to involve either victim withdrawal or insufficient and/or conflicting evidence. The most common factors in cases with no identified suspect were that the victim withdrew from the process, there was insufficient evidence to identify a suspect who was a stranger to the victim, or the victim had limited recall due to intoxication.

‘False complaints’ were defined as cases in which the complainant was charged or warned for making a false complaint. In ‘false complaint’ cases for which further information was noted in the summary data set, the two most common file notes were that the complainant had admitted the allegation was false and that the evidence did not support the complaint. The victim had an intellectual disability or a psychiatric condition or had made previous allegations in around a third of cases.

The ‘no offence’ category accounted for 34 percent of recorded cases (including the 8 percent designated ‘false complaints’), which was lower than the 45 percent ‘no offence’ rate found in a 1981 New Zealand study (Stace, 1983). However, the inappropriate use of this category still seems to be an issue, as some of the cases classified as ‘no offence’ appeared to be offences. Based on the limited information available in the current data, the actual ‘no offence’ rate may be between about one in five cases and one in four cases.

At least one in five cases did not proceed due to victim withdrawal. That is, the investigating officer recorded that the victim did not want to proceed with the investigation or was uncooperative or could not be contacted. Withdrawn cases were more likely than other cases to involve an offender who was an ex-partner or boyfriend. The police files noted a variety of reasons for a victim not wanting to proceed, including that the victim wanted the offender warned or trespassed but not prosecuted; someone else reported the incident or the victim was pressured to report; the victim had limited recall of the incident; the victim wanted to report the incident or seek advice but take no further action; or the victim did not feel able to proceed, was not ready to proceed or felt threatened.

Comparisons with overseas research were complicated by differences between studies in scope, definitions and justice processes. Attrition rates not only differ by case type, but may vary between areas within a single study. However, the current New Zealand attrition rates were generally around the average of a range of other overseas studies and the major predictors of attrition were generally similar.

The overall conviction and prosecution rates appear to have decreased relative to an earlier New Zealand study (Stace, 1983). The type of cases has also changed markedly, with proportionally fewer cases involving young victims and stranger rapes than in the early 1980s.

Factors influencing attrition

A few key factors were the major predictors of outcomes, with the predictors varying between stages of the investigation and prosecution.

Offence type and number

Rape cases had higher attrition rates at almost every stage of the justice process, compared with other offences. The 18 percent of cases that involved more than one offence were much more likely to proceed through all stages and result in a conviction than cases involving a single offence.

Victim-offender relationship

The majority of offenders were previously known to the victim, with stranger assaults accounting for just 16 percent of cases and offenders just met (within the last 24 hours) accounting for a further 15 percent of cases. A third of cases involved victims and offenders with intimate relationships (family, current partners or

ex-partners) and 37 percent of cases involved other known offenders (such as friends, acquaintances, people known through work and caregivers).

The victim-offender relationship was a significant factor predicting attrition at almost all stages of the process, although the analysis was complicated by missing relationship data for many cases.

• Attacks by a stranger were more often associated with ‘false complaints’ and ‘no offence’ cases and had a high attrition rate due to their association with unidentified suspects. However, if prosecuted, stranger attacks were much more likely to result in a conviction, giving these cases a relatively high overall conviction rate.
• Other non-intimate offenders had a relatively high attrition rate at most stages, so a low overall conviction rate.
• Current partners and boyfriends had a high prosecution rate but a very low conviction rate for sexual violation (although half of those prosecuted were convicted of other violent offences).
• Offenders who were family members had high prosecution and conviction rates relative to other offenders.

Victim age

A third of victims were aged 16 to 19 years and over half were aged under 25 years. By comparison, just 13 percent of offenders were aged under 20 years and 28 percent were aged under 25 years. Victims aged 16 to 19 years had a higher rate of ‘false complaints’ (for stranger attacks), but their cases also had a higher rate of conviction if the offender was prosecuted.

Other victim and offender factors

The 37 percent of offenders with previous sex or violence convictions were much more likely to be prosecuted and convicted than other offenders.

A substantial minority of victims had some type of disability  psychiatric (7 percent), intellectual (6 percent) or physical (one percent). Victims with a psychiatric condition or an intellectual disability had a higher rate of ‘false complaints’ than other victims. However, prosecutions were more likely to result in a conviction in cases where the victim had an intellectual disability.

The 9 percent of victims who had made previous allegations of sexual victimisation had a higher rate of ‘false complaints’ and their cases had a low rate of prosecution.

Cases were more likely to be classified as ‘no offence’ if the victim was uncertain whether violation had occurred. The victim was uncertain in about one in seven cases, and this factor was strongly linked to alcohol or other drug use.

Suspects were less likely to be prosecuted if the victim had one or more of the following attributes: refused a medical, had a psychiatric condition, had made previous sexual allegations, was intoxicated, or did not report promptly.

A combination of several aggravating or negative factors in the case had more impact than single factors alone.

Incident factors

Cases involving force, threats or injury (at least 41 percent of cases) were less likely to be classified as ‘no offence’ and the offender was more likely to be prosecuted and convicted, especially if the injuries were serious.

Suspect identification was more likely when there was witness or forensic evidence, but less likely if the victim refused a medical examination
 

Factors not influencing attrition

No significant differences in outcomes were found between police districts once other factors had been taken into account in the modelling, with one exception – the Auckland City Police District had a lower rate of ‘no offence’ cases.

The following factors were not predictors of any outcome:

• victim gender, ethnicity and origin (New Zealand born, overseas born), victim criminal history, and victim a sex-worker
• offender gender, age, ethnicity and origin (New Zealand born, overseas born)
• incident timing (day, time of day and year).

1     Introduction

 

1.1    Project overview

As part of its work to improve women’s well-being, the Ministry of Women’s Affairs is leading a research project on effective interventions for adult victims/survivors of sexual violence. The project has four interrelated work streams, comprising:

• a study of pathways from crisis to recovery, targeting individuals who have experienced sexual violence as adults and focusing on their experiences with a variety of support sources (Kingi and Jordan, 2009)
• an environmental scan of agencies and key informants that respond to victims/survivors, focusing on systemic, organisational and other contextual factors that influence systems and agency responses (Mossman et al. 2009b)
• this retrospective analysis of attrition of sexual violation incidents recorded by the New Zealand Police (the attrition study)
• a literature review of good practice in service delivery (Mossman et al., 2009a).

 

The findings from these work streams will contribute to the Government’s considerations for policy and practice responses for victim/survivors of adult sexual violence. The Ministry of Women’s Affairs is leading the research in partnership with the Ministry of Justice and New Zealand Police.

1.2    Scope and objectives of the attrition study

This report presents the results of the retrospective study of attrition of cases recorded as sexual violation. The main aim was to assess what proportion and type of cases dropped out at each stage of the criminal justice process (the attrition rate), from recording of the offence by police to final disposition or outcome.

The focus of this project is on sexual violation against adult victims. For the purposes of the project, an adult was defined as a person aged 16 years or older at the time of the offence. This research examined sexual violation (rape and unlawful sexual connection), attempted sexual violation and selected offences relating to sexual violation or exploitation (see Appendix A). The most serious offences (rape and unlawful sexual connection) made up 90 percent of the sample.

The study was based on all sexual violation offences against adults recorded by the New Zealand Police from 1 July 2005 to 31 December 2007, a sample of 1,955 cases (2,888 offences). It provides the first detailed analysis of a large sample that tracks what happens to cases between recording and conviction in New Zealand.

The specific objectives of this project were to identify:

• conviction rates, both nationally and for the 12 New Zealand Police districts
• rates of attrition at successive stages of the criminal justice process
• factors that are associated with or predict outcomes
• a profile of the characteristics of victims/survivors, offenders, offences and cases.

 

1.3    Rationale for the attrition study

Sexual violation is one of the most serious of all offences, with devastating effects for victims and a significant impact on the quality of their lives generally. Not only is the offence traumatic, but the process of seeking justice can be a harrowing experience for many victims, both in terms of the police investigation and the court process. Many examples of this from a New Zealand perspective are provided by Jordan (1998, 2001a, 2001b, 2004a, 2004b) and earlier studies (Lee, 1983; Stace, 1983; Young, 1983).

Sexual violation affects many New Zealand women, some men, and their families. Over a thousand offences a year were recorded by the police over the study period among an adult female population of 1.7 million.

Moreover, this number reflects only a small proportion of actual levels of sexual violence against adults, as an estimated nine in ten sexual offences are not reported to police, according to the New Zealand Crime and Safety Survey 2006

(Mayhew and Reilly, 2007). The reporting rate for sexual offences was the lowest of the 12 offences cited in Mayhew and Reilly (2007). As a comparison, an estimated 36 percent of assaults were reported to the police.

International research and previous research in New Zealand have highlighted high attrition rates for sexual violation offences. Attrition occurs at all stages of the criminal justice process, but particularly during the investigation phase before prosecution. A meta-analysis of 75 studies (Daly and Bouhours, 2008) found that a third of recorded cases result in charges being laid and around one in nine cases results in a conviction.

Attrition can occur for a wide variety of reasons; for example, it is not always possible to identify a suspect, there may be insufficient evidence to proceed or poor prospects of conviction, the victim may withdraw from the process, or the investigation may show that the incident was not an offence.

Understanding the rate of attrition and reasons for attrition at each stage of the process is a critical first step towards improving policy and practice so that attrition is minimised and outcomes for victims and society are improved.

1.4    Key issues identified by previous research

Various aspects of attrition in sexual violation cases have been analysed in a large number of studies. The literature review undertaken as part of the present report focused on both New Zealand studies and the most recent studies undertaken in the two jurisdictions most similar to New Zealand – England and Australia. In addition, wider international comparisons were available from a meta-analysis of 75 studies from five English-speaking countries (Daly and Bouhours, 2008). An overview of the samples and scope of these studies is in Appendix B. Full citations for these and other studies referred to are at the end of the report.

Specific comparisons between the present project and other research are made within the results section of this report. This section summarises some of the key issues noted by many researchers.

A major focus of attrition studies is the low conviction rate for sexual violation cases, in combination with high attrition at all stages of the process. The relatively high rate of ‘no-crimed’ cases has been a particular focus of some studies, as this rate, combined with confusion between the ‘no-crime’ and ‘false complaint’ categories, has contributed to the perception held by some people that many or most rapes are false allegations.

The high attrition rate for rape cases reflects the difficulty in many cases of achieving proof beyond reasonable doubt. As rape generally occurs in the absence of witnesses and most often involves an offender known to the victim rather than a stranger, proving non-consent or even determining that an offence has occurred can be a major obstacle. As Lievore (2004: 4) notes, non-consent is critical: ‘the prosecution must prove beyond reasonable doubt that the accused knew that the victim was not consenting or was reckless as to consent’.

While clearly it is essential to have sufficient evidence of offending in order to prosecute, these issues inevitably lead to a focus on the victim that research suggests can be excessive (Kelly et al., 2005). Thus, it is argued that too much attention may be given to the characteristics, behaviour and perceived credibility of the victim, rather than the actions of the offender, resulting in revictimisation, victim withdrawal and inappropriate attrition of cases through no-criming.

Research suggests that this scrutiny of the victim can be influenced by the unrealistic stereotypes and myths held about rape and rape victims. While many agree that the influence of such stereotypes has moderated following extensive reforms of law and practice over the last 30 years, it is also clear from many studies that stereotypes still persist to an unacceptable degree, as noted in Kelly et al. (2005: 2):

‘real rapes’ continue to be understood as those committed by strangers, involving weapons and documented injury. The failure of criminal justice systems to address stereotypes means that the processes involved in responding to reported rapes – from early investigation through to courtroom advocacy – can serve to reinforce, rather than challenge, narrow understandings of the crime of rape, who it happens to and who perpetrates it. The attrition process itself reflects, and reproduces, these patterns.

The effect of this stereotyping can influence all aspects of attrition and all involved parties, from the offender’s belief as to what constitutes rape, to the families and friends who discourage the victim from proceeding, to the jury and even the victim. The main research focus in this area has been on the effect of the policies, attitudes, beliefs and behaviour of the police and, to a lesser extent, prosecutors and the Courts. All detailed studies provided examples of poor practices, leading to early withdrawal of victims or insufficient follow-up of evidence (for example, Kelly and Regan, 2001; Kelly et al., 2005; HMIC and HMCPSI, 2002, 2007).

These studies also found significant variation in attrition rates and outcomes between areas or police forces, which could only be partly explained by differences between areas in case and population profiles. The high rate of no-criming in some areas was of particular concern. Variation in outcomes between areas may reflect the diverse approaches used by different forces, and thus provides useful insights into how attrition rates can be reduced by improving investigation procedures, evidence collection and victim services.

Many researchers were also concerned about the downward trends in conviction and detection rates in some countries, especially in England and Wales. Others noted that positive changes affecting the reporting and official recording of offences explain a significant proportion of this negative trend (Feist et al., 2007; HMIC and HMCPSI, 2002, 2007).

1.5    Legislation and policy in New Zealand

1.5.1    Legislation in New Zealand

In the early 1980s, concern about an increase in the volume of reported rapes in

New Zealand, changes in attitudes to women, and law reform overseas all contributed to the commissioning of a major review of sexual violation in New Zealand (Young, 1983). Following this review, in 1985, significant amendments regarding sexual offences were made to the Crimes Act 1961, Evidence Act 2006 and Summary Proceedings Act 1957:

• sexual violation was made gender neutral and redefined to include unlawful sexual connection and rape within marriage
• the conditions under which consent is assumed not to occur were extended (these conditions are now: if there is force or threats or fear of these; if the person is asleep or unconscious or affected by drugs or alcohol; if the person is affected by an intellectual, mental, or physical condition or impairment of such a nature and degree that he or she cannot consent or refuse to consent to the activity; if he or she allows the act because he or she is mistaken about its nature and quality; or if the person is mistaken as to the identity of the offender)
•  changes to court and trial procedures to make giving evidence less traumatic for victims and limiting the publication of incident details
• removal of the requirement for the judge to warn of the dangers of convicting based on the victim’s uncorroborated evidence and clarification of the relevance of other evidence (such as the reasons a victim might not report an offence immediately).

 

The Victims of Offences Act 1987 was enacted in July 1987 to make better provision for the treatment of victims of criminal offences. This Act was later replaced by the Victims’ Rights Act 2002.

The Evidence Amendment Act 1977 (and later the Evidence Act 2006) provides a partial ‘rape shield’, as evidence of sexual experience between the complainant and any person other than the accused is not allowed without the prior agreement of the judge. However, evidence of sexual history between the complainant and accused may be raised in Court in New Zealand, unlike in many other countries (Ministry of Justice, 2008).

Legal and process reform in relation to sexual assault is again under the spotlight, with the announcement in 2007 of a taskforce to examine the effectiveness of criminal justice system responses to sexual offending against adults (Taskforce for Action on Sexual Violence), including a discussion document inviting submissions on possible changes to legislation (Ministry of Justice, 2008).

1.5.2    Police policy in New Zealand
New Zealand Police policy and practice in relation to investigations of sexual offending have changed over the last quarter of a century. The report of the Commission of Inquiry into Police Conduct (Bazley, 2007: 78) noted:

I was generally impressed at the way in which the police had steadily improved policies relating to the investigation of adult sexual assault during the period of interest to my inquiry. At the beginning of the period, policies in force reflected some very distorted views of the credibility of victims of alleged sexual assault, and the general approach to interviewing victims was rightly perceived as likely to be in itself a re-victimisation. The shift towards practices that recognise the impact of recent trauma, encourage a good working relationship with professional support agencies, and restore to the victim a sense of empowerment are all to be commended.

The police policy in place during the period relevant to the current data was the Adult Sexual Assault Investigation (ASAI) Policy. This policy was phased in from 1998, alongside a revision of the Manual of Best Practice

An overview of the ASAI Policy and past policy is provided in the Commission of Inquiry into Police Conduct report (Bazley, 2007: 78–82).

The ASAI Policy sets out the principles, commitments, procedures and training required. It requires all front-line and watch-house staff to be trained in taking initial sexual assault complaint action and dealing with victims. It further states that victims are to be offered support and medical services at an early stage. The complaint should be transferred to a trained investigator as soon as possible (within two days) for a formal interview and statement-taking.

Specialist training should incorporate not only training in sexual assault investigation but also an awareness of the needs of adult sexual assault victims, an understanding of the roles and responsibilities of specialist agencies working with sexual assault victims, and knowledge of the relevant laws and practices relating to adult sexual assault investigations. (Bazley, 2007: 74)

However, the implementation of the 1998 ASAI Policy has taken time. For example, the new training course was not implemented until 2003, five years after the ASAI Policy was introduced. Not all officers who deal with sexual assault have since received this specialist training and the executive summary of the Commission of Inquiry into Police Conduct report (Bazley, 2007: 6) notes significant implementation issues:

• There are some discrepancies between the Manual of Best Practice as it relates to sexual offending and the Adult Sexual Assault Investigation Policy. These result in an unnecessarily unwieldy and fragmented approach
• The Adult Sexual Assault Investigation Policy does not appear to have been adequately supported by training and the provision of appropriate facilities. Police described the policy as ‘aspirational’. This has both inhibited and prevented the mandatory aspects of the policy and its requirements concerning the competence of investigators from being adequately implemented.
• There are few nationally mandated training packages; the extent and content of most staff training is decided at the police district level. Consistency in the delivery of police services requires a more coordinated and strategic view of training requirements and priorities.

Recent research in New Zealand suggests that there are still significant problems with entrenched attitudes within the police force (Jordan, 1998, 2001a, 2001b, 2004a, 2004b). The research indicated that rape complainants ‘must still battle to gain credibility in the eyes of some police investigative officers’, despite changes to police policy on the investigation of sexual assault (Jordan, 2004a: 29). False complaints were still considered to be common by many officers, and myths about rape and about the victim’s expected reaction to rape still persisted.

 
New Zealand Police has commissioned initiatives and projects to implement and build on all of the recommendations made by the Commission of Inquiry in 2007. New Zealand Police recognises that the benefits of the suite of initiatives under way are unlikely to be captured by this research, as they are still in the process of being developed and implemented across police districts.

Other important work under way by New Zealand Police includes:

• revising the ASAI Policy in conjunction with sector partners Te Ohaakii a Hine – National Network Ending Sexual Violence Together and Doctors for Sexual Abuse Care
• building the capacity of the comprehensive adult sexual assault training programme to cater for more police staff attendees
• implementing the Investigative Interviewing Project, which has the potential to improve rape complainants’ experiences of police interviewing.

 

1.6    Structure of this report

The structure of this report is as follows:

• chapter 2 describes the sample, data extraction and analysis methods
• chapter 3 presents a profile of the characteristics of victims, offenders, offences and cases
• chapter 4 examines attrition rates during the police investigation phase, including prosecution rates, ‘no offence’ rates, and the reasons why cases do not proceed
• chapter 5 looks at attrition during the court process, including rates of conviction, acquittal, discharge and withdrawal
• chapter 6 assesses the factors associated with attrition at each stage of the process using multivariate analysis.

 

2    Methodology

 

Box 1: Definitions: study scope and methods

Case	The unit of measure in all analyses, unless otherwise stated, is the case, which may be a single offence or a group of related offences (at least one of which has been coded as sexual violation) against a victim documented in a single police file.
Offender	Alleged offender, suspect or defendant.
Sexual connection	The 1985 amendment to the Crimes Act 1961 defined sexual connection as (a) connection effected by the introduction into the genitalia or anus of one person, otherwise than for genuine medical purposes, of (i) a part of the body of another person; or (ii) an object held or manipulated by another person; or (b) connection between the mouth or tongue of one person and a part of another person’s genitalia or anus.
Sexual violation	The study used a broad definition of sexual violation, including sexual violation as it is defined under section 128 of the Crimes Act 1961 after the 1985 amendment (rape and unlawful sexual connection), attempted sexual violation (section 129) and other offences involving sexual violation (incest, inducing sexual connection, and sexual exploitation of a person with significant impairment), as listed in Appendix A.
Victim	Person recorded as the victim of sexual violation, including those in cases subsequently deemed ‘no offence’ or ‘false allegation’ (although the term ‘complainant’ was also used for these cases). All victims were aged 16 or over at the time of the offence.

2.1    Terminology

Throughout this report, the terms ‘victim’ and ‘offender’ are used, irrespective of the outcome of the case. These terms are not intended to imply that all recorded sexual violations have been substantiated, as this is not so. Alternative terms may be used where more appropriate to the circumstances or stage of the process, including ‘complainant’ instead of ‘victim’ for ‘false complaints’ and ‘suspect’ or ‘defendant’ for the ‘alleged offender’.

The term ‘victim’ was used as this is the common terminology for analysis of criminal justice data. The use of this terminology is not intended to negate terms, such as ‘survivor’, which are used in other sectors.

2.2    Scope

This study was based on all offences coded as sexual violation against an adult victim (aged 16 or over) recorded by New Zealand Police in the National Intelligence Application (NIA) database from 1 July 2005 to 31 December 2007. The sample includes offences committed earlier than this period, but not reported at the time of the incident. This period was selected to avoid the potential effects of the changeover to the NIA system just before July 2005, while still leaving a sufficient period for a robust sample size and an adequate follow-up.

The study used a broad definition of sexual violation, including sexual violation as it is defined under section 128 of the Crimes Act 1961 (rape and unlawful sexual connection), attempted sexual violation (section 129) and other offences involving sexual violation (incest, inducing sexual connection, and sexual exploitation of a person with significant impairment), as listed in Appendix A. The most serious offences (rape and unlawful sexual connection) made up 90 percent of the cases, while attempted violation and other offences accounted for 7 percent and 3 percent respectively.

The final sample, after removal of out-of-scope cases, comprised 2,888 sexual violation offences, involving 2,029 offenders from 1,955 cases (police files).

The sample includes all offences recorded as sexual violation, even if subsequent investigation determined that no offence had occurred.

2.3    Data source and extraction

When a crime is reported to the police, the Offence Report information is recorded electronically, using a preformatted screen in NIA. A unique file number is allocated to each report and this is retained forever. Information may be added to the file over the course of the investigation and prosecution, including a summary of facts, witness statements, forensic results, and conviction and sentencing details.

The electronic files do not always contain as much detail as the paper files on some aspects of the investigation, such as witness statements. However, the NIA database has several advantages: it provides a single point of data collection for the full range of required information that may be held across several paper files; it has more information on prosecution outcomes; and it can be used to access other relevant information, such as the criminal history of the alleged offender and information about the victim’s previous contacts with the police.

Data were extracted during July to September 2008, giving a minimum six-month period between the offence being recorded and the data being extracted.

Six percent of cases were still before the Courts when the data were extracted and a further 10 percent of case files remained open with no charges yet laid.

For privacy and security reasons, the data for this study were extracted by a team of six experienced former police investigators. Only these team members had access to the data. The team was provided with specific training for this data extraction to ensure consistency and accuracy, as well as access to counselling, due to the nature of the project. The data extraction was co-ordinated and audited by a former police investigator with considerable experience in the investigation of sexual offending.

No personally identifying information was included in the data set supplied to researchers. The original file numbers and offender identities were replaced with unique numbers to preserve the anonymity of subjects. Alleged offenders in cases that were documented as ‘false complaints’ were assigned a ‘000’ number and no further offender details were collected in such cases.

Project staff from the Ministry of Women’s Affairs and New Zealand Police worked together to define the variables to be collected and the coding schedule. Variables were selected to provide a broad range of information about the offence, offender, victim, investigation, prosecution and outcomes. Where possible, factors found to be associated with attrition rates and outcomes in international research were included in the schedule. Further details of the variables collected and their definitions are provided alongside results in chapters 3 to 6.

Information was not collected on other offences committed at the same time (e.g. indecent assault, assault, kidnapping or burglary). However, convictions for other offences that related to the same incident were routinely noted in the ‘comment’ field of the database, as all offenders were checked with respect to charge history and criminal history.

2.4    A caution on comparisons with official statistics

Caution should be observed when comparing figures produced by this study with New Zealand Police official statistics for recorded offences and apprehensions, as the statistical units and counting rules used in official statistics differ from those used in this study. Definitions and counting rules applicable to official statistics are published along with these statistics on the New Zealand Police and Statistics New Zealand internet sites. In particular, ‘no offence’ outcomes are excluded from official statistics, but are included in this study.

New Zealand Police is implementing its Statistics Strategic Plan 2006–2010, which is enhancing the quality and scope of information recorded and reported by police. For example, on 1 July 2008 the National Recording Standard was implemented, and New Zealand Police is developing a plan for reporting statistical information about victims and offenders, based on this new National Recording Standard. These initiatives are consistent with recommendations made by the Crime and Criminal Justice Statistics Domain Plan, which was not yet published by Statistics New Zealand at the time of the writing of this report.

2.5    Data limitations

The accuracy of this analysis can only be as good as the accuracy, extent, depth and objectivity of information recorded in NIA and extracted for this research. Almost all variables had at least some missing data (i.e. the information was not recorded in the file either because the information was unknown or because the data had not been transcribed from the paper file to NIA).

The key information used in this analysis was recorded for most cases, including core information about the incident (offence type, location, incident and reporting dates), the victim and case progress (arrests, court process, outcome and sentence). Offender information was recorded where this was known, excluding ‘false complaint’ cases.

Variables relating to the physical circumstances of the case (whether alcohol was a factor, the victim was injured, force was used, or there was evidence linking the offender and victim) were included as specific coded variables, but the relevant data were often missing from the file notes. This type of information was more often missing from offences coded ‘no offence’ or ‘false complaint’, but still frequently missing from other types of case. Where the data were missing, it is not possible to say whether this means the factor was absent from the incident or was present but not recorded.

Interpretative information (such as the reasons for the case not proceeding) was recorded only in text or comment fields and was subject to at least four layers of filtering:

1. the version of events in statements made by the victim, offender and witnesses
2. the accuracy and extent of recording of these statements and the perception of the veracity of participants by investigating officers
3. the summarisation of file notes by the data extraction team
4. the extraction and interpretation of information from these summarised notes by the researcher.

 

Therefore, some of the contextual analysis is derived from very summarised and incomplete information, which can provide only an indication of the circumstances of these often complex cases from a limited perspective. This issue is reiterated at various points within the presentation of results and, where possible, research from other perspectives is used to provide additional context.

2.6    Data analysis

2.6.1    Unit of measurement
The primary unit of measurement within this report, unless otherwise stated, is the case, which may be a single offence or a group of related sexual violation offences committed by one or more offenders against a single victim and documented in a single police case file. The main exception to the use of cases as the unit of measure was the analysis of offender characteristics (section 3.2). Where appropriate, some analysis is also presented for individual offences.

Thus, three data sets were developed for the analysis: the offence, offender and case data. All in-scope offences were included in the offence data. The offender data contained a single line of data for each offender, based on the offence that resulted in the most serious outcome or (if outcomes were the same) the most serious offence. The case data contained a single line of data for each case file, based again on the most serious outcome. The 1,955 cases in the sample comprised 2,888 offences and 2,029 offenders.

2.6.2    Statistical testing and modelling
The statistical significance of differences between groups was tested using

chi-squared tests, with the significance level set at 95 percent.

One objective of this report was to assess which factors were associated with or predicted the likelihood of conviction and other outcomes. A multivariate modelling technique, bivariate logistic regression, was used to identify the major predictors of the probability of each outcome (such as being convicted) from a range of possible factors within a single analysis. Only finalised cases were included in these models.

The attrition points or outcomes were all coded as binary variables (e.g. convicted or not convicted). All variables collected in the data set supplied by the police were also converted to binary or ordinal codes. For example, age was converted into several discrete groups (e.g. whether the victim was aged under 20 or not), because the relationship between age and attrition rates was not linear.

Missing information was a major problem, especially for attrition points with asymmetry in missing variables. For example, ‘false complaint’ cases rarely had data relating to the offender and circumstances of the case (such as scene, time, evidence or injury). Such variables could not be included in the models for these attrition points. Even in outcome categories with better data recording, many individual cases had some missing data for some factors. Factors with a significant proportion of missing data were checked before inclusion in the final model, to ensure it was the presence of the factor that was relevant rather than the missing cases.

The best-fit model was identified using all the relevant data and testing the full set of possible predictors. Usually, this produced a model with a few strong predictors and many weaker predictors. Given the modest sample size and the wide variety of factors tested, over-specification was a potential problem. That is, it is quite likely that some of the weaker predictors were representing just a few cases and may not be found consistently in other data sets. To test the consistency of prediction, the models were checked on random subsets of the data.

The best-fit model was selected based on a combination of explanatory power (Nagelkerke’s psedo-R2), predictive success, minimising the number of variables, and goodness of fit (Hosmer-Lemeshow statistic). Thus, predictors were not included if they added little to the explanatory power of the model or if they were not consistent predictors.

2.7    Comparisons with other research

Where possible, the results of this study were compared with previous New Zealand research and recent studies from similar overseas jurisdictions. However, studies vary greatly in scope and definitions, as discussed in Appendix B, such that comparisons must be made with caution.

 

3    Profiles of victims, offenders and incidents

 

Box 2: Definitions: victim and offender characteristics

Age	Age at the date the offence occurred. All victims were aged 16 or over at the time of the offence.
Case	The unit of measure in all analyses, unless otherwise stated, is the case, which may be a single offence or a group of related offences (at least one of which has been coded as sexual violation) against a victim documented in a single police file.
Criminal history	Previous convictions or proved charges in New Zealand, excluding convictions for traffic offences.
Disability	Psychiatric, intellectual or other recorded disability, as determined by a doctor or other relevant specialist.
Ethnicity	As recorded in the police data.
False complaint	The alleged victim was charged with or warned in relation to making a false complaint.
Offender	Alleged offender, suspect or defendant.
Victim	Person recorded as the victim of sexual violation, including those in cases subsequently deemed ‘no offence’ or ‘false allegation’ (although the term ‘complainant’ was also used for these cases). All victims were aged 16 or over at the time of the offence.

Note on the presentation of results: Some variables had a significant proportion of missing data. Therefore, the following sections indicate the extent of missing data before presenting the results. Results are shown as a percentage of all cases, excluding missing data.

3.1    Victim profile

The following profile refers to all 1,955 victims, including those cases where the alleged victim was subsequently determined to have made a ‘false complaint’.

The differences between victims at various stages of the investigation and prosecution are examined in sections 4.5 and 5.5.

All cases related to a single victim. However, in one percent of cases the file notes indicated that the offender was also charged with or had committed offences against other victims.

3.1.1    Demographic and employment characteristics of victims
Gender was recorded for 99.5 percent of victims. The vast majority of victims were female, with males making up just 5 percent of the sample.

All victims in this study were aged 16 or over. Age was recorded for 93 percent of victims. Victims of sexual violation were most often young (Figure 2), as has been found in overseas studies. Of the victims for whom age was recorded, a third were aged under 20 years. The average age was 27 years and the median age was 23 years.

Victims aged in their 20s made up a further third of the sample. The remaining victims were spread across the age range, with 19 percent in their 30s, 10 percent in their 40s and 4 percent aged 50 or over.

 

Figure 2: Age distribution of victims

 

The only other similar study from New Zealand – from a quarter of a century ago – found a higher proportion of young victims and fewer middle-aged victims than in the current study. The earlier study found 45 percent of adult victims were aged 17 to 20 years and 16 percent were aged 31 to 50 years (Stace, 1983) compared with 35 percent and 28 percent respectively in the current sample, after adjustment for age and offence differences.

This difference in age profile is likely to reflect the changing composition of recorded rape cases, with proportionally fewer cases in 1981 involving people in intimate relationships than now (see section 3.1.2). For example, the median age of victims in the current study was 21 years for rapes by a stranger or acquaintance compared with 31 years for rapes by a current or ex-partner.

Ethnicity was recorded for 88 percent of victims. Of those with ethnicity recorded, 61 percent were European, 27 percent were Māori, 6 percent were Pacific peoples, 5 percent were Indian/Asian and less than one percent were from other ethnic groups. These results were similar to the 1983 profile. The proportion of Māori victims was about twice the proportion of Māori in the adult New Zealand population, while the proportion of Asian/other victims was about half that of the population.

Employment status was recorded for less than half the victims. Of these, 45 percent were employed.

At least 48 victims (2.5 percent) were sex-workers, based on the limited information in the ‘comments’ field of the database. Most, but not all, of the offenders in these cases were clients.

3.1.2    Relationship between victim and offender
The relationship between the victim and offender was recorded for 73 percent of cases.

Sexual violation by a stranger accounted for 16 percent of cases in which the relationship was recorded, and in a further 15 percent of cases the victim had just met the offender (within the last 24 hours) (Figure 3). These figures are likely to be underestimated by several percentage points, because the victim-offender relationship information was more often missing for cases in outcome categories with a high proportion of offenders who were strangers or had just met.

Nonetheless, it is clear that the majority of offenders were previously known to the victim, either intimately (33 percent of cases with a recorded relationship) or otherwise (37 percent). Intimate relationships included current partners (10 percent), ex-partners (10 percent), boyfriends or ex-boyfriends (5 percent), and family/whānau (8 percent). Other known offenders included friends (10 percent), acquaintances (10 percent) and dates (one percent), with the remaining 17 percent in an ‘other known’ category. The ‘other known’ category covered a wide variety of relationships, such as people known through work, sports clubs, church or education, as well as relationships such as flatmates and caregivers.

 

Figure 3: Victim-offender relationship (where known)

 

 

The low proportion of rapes in which offenders were strangers has also been found in overseas studies, in contrast to the perception of the ‘traditional’ rape scenario. For example, in England, Feist et al. (2007) found that the offender was more often an acquaintance (27 percent), partner or ex-partner (22 percent), or relative or parental figure (15 percent) than a stranger (14 percent). Australian research shows similar results, with strangers accounting for 16 to 24 percent of offences, acquaintances or friends 35 percent, current or ex-partners 19 to 26 percent, and family 8 to 15 percent (OWP, 2006; Lievore, 2004).

The only other study from New Zealand (Stace, 1983) comprised 220 victims of all ages, and included rape and attempted rape only, so was not exactly comparable to the present study. Nevertheless, there was a clear difference between the overall results for the two studies. In particular, the current study had about twice the proportion of offences committed by someone in an intimate relationship and half as many stranger rapes (after adjustment for offence differences between the studies).

The inclusion of under 16-year-olds in the earlier study will have influenced these results to some extent, as teenagers are more likely to be raped by strangers and less likely to be raped by partners (although more likely to be raped by family). However, even for 16–19-year-olds in the current study, the proportion of stranger attacks was only 22 percent (Table 1), less than the 35 percent for all ages in the earlier study.

The changing composition of relationships within recorded violation cases is likely to reflect significant social changes as well as legislative changes in the mid-1980s that redefined sexual violation offences, the conditions of consent, the requirement for evidence corroboration and court procedures (section 1.5). The total number of recorded sex offences (of all types) doubled in the decade following the 1985 legislative changes, presumably due, at least in part, to an increased rate of reporting (Triggs, 1997).

Similar results have been reported in England, where the proportion of recorded rapes committed by a stranger decreased from 30 percent in 1985 to 12 percent in 1997 (Harris and Grace, 1999).

In the current study, the relationship between the victim and offender varied by age, as shown in Table 1.

As noted above, the proportion of stranger attacks may be underestimated by a few percentage points, because a substantial proportion of relationship data was missing (27 percent overall), especially within the younger age group (31 percent missing data).

 

Table 1: Victim-offender relationship, where known, by age of victim (%)

Relationship between victim and offender	Age group of victim (years)						Total
	16–19	20–24	25–29	30–39	40–59	60+
Sample size	337	304	161	278	174	15	1,421
	Offender relationship as a percentage of each victim age group (where known) (%)
Family/whānau	11	8	4	4	4	13	8
Current/ ex-partner	4	16	28	32	35	13	20
Boyfriend/ girlfriend, date	5	9	4	7	7	0	7
Friend	10	10	9	9	10	13	10
Acquaintance	17	9	8	7	9	0	10
Just met (within 24 hours)	18	21	16	9	6	0	15
Other known	14	13	17	21	18	40	17
Stranger	22	15	15	12	10	20	16
Total	100	100	100	100	100	100	100

Victim ethnicity was not strongly associated with victim-offender relationships. The significant differences were that offenders in cases with Māori victims were more likely to be current partners or family and less likely to be boyfriends or friends, compared with offenders against European victims.

For the 63 male victims where the offender relationship was recorded in the current study, 25 percent were acquaintances or friends, 17 percent were strangers, 11 percent had just met, 3 percent were family, there was one case each of partner, ex-boyfriend and date, and the remaining 38 percent were in the ‘other known’ category.

3.1.3    Mental health and disability status of victims
Disability status was recorded for 93 percent of victims. Of these, 15 percent were recorded as having some type of disability as determined by a doctor or other relevant specialist. Most of this group had either a psychiatric condition (47 percent) or an intellectual impairment (44 percent), with the rest having a physical disability (8 percent) or another unspecified disability (one percent). The median age of victims with a disability (31 years) was higher than for other victims (26 years).

The inclusion in this study of offences involving sexual exploitation of a person with a significant impairment increased the overall disability proportion, but only slightly. If the 24 victims of these offences were excluded from the total, the disability proportion of other victims was still 13 percent.

The relatively high proportion of victims with a psychiatric or an intellectual condition has been noted in overseas studies as well. For example, in England, Feist et al. (2007) found 17 percent of victims were classified as vulnerable – 57 percent with a mental disorder, 35 percent with significant impairment of intelligence, 6 percent with a physical disability and 2 percent unknown. In Victoria, Australia, a quarter of victims were identified as having a disability – of this group, 59 percent had a psychiatric disability and 22 percent had an intellectual disability (OWP, 2006).

In cases where the victims had a disability and the offender details were recorded, the offender was less likely to be a partner, an ex-partner, a boyfriend or a date (19 percent for victims with a disability compared with 28 percent for other victims), a stranger (10 percent compared with 17 percent) or someone who had just been met within the last day (7 percent compared with 15 percent). Conversely, offenders against victims with a disability were much more likely to be in the ‘other known’ category (35 percent compared with 15 percent for other victims).

Of the 64 cases in which the offender was in the ‘other known’ category and the victim had a disability, around a third had no file information that indicated the offender’s relationship to the victim and the majority of these appeared to be unsubstantiated offences (i.e. the victim agreed sex was consensual or there was no supporting evidence that an offence occurred). Where information was recorded, the two most common relationships were that the victim and offender were both patients in care (18 cases, or 10 percent of cases involving a victim with a disability) or the offender was a caregiver (12 cases, or 7 percent of cases involving a victim with a disability).

3.1.4    Previous allegations of violent victimisation by victims
Data on previous contacts with the justice system were available for 98 percent of victims. Many victims had had some previous contact with the justice system (as a victim/complainant).

Seven percent of victims had applied for a protection order and 2 percent were the respondent on a protection order, not necessarily in relation to the offender.

Previous allegations of violent victimisation had been made by 43 percent of victims (not necessarily against the same offender). The majority of these incidents related to domestic violence (71 percent), with 21 percent relating to sexual violence and 7 percent to other types of violence.

Prior allegations of domestic violence were more likely in cases where the victim and offender were ex-partners (73 percent) or current partners (62 percent) and less likely in cases where they were friends, acquaintances or dates (35 percent) or strangers (29 percent), but there was little difference between these groups in the prior rate of sexual allegations.

There was no information on whether the prior allegations had been proven. However, 77 percent of the offenders in these current or ex-partner relationships had a previous criminal history compared with 43 percent in current or ex-partner relationships where the victim had made no previous allegations.

Fifty-eight percent of victims with a psychiatric condition and 46 percent of victims with an intellectual impairment had made previous allegations of violent victimisation, and these two groups accounted for 29 percent of victims who had made previous sexual assault allegations.

Research on victimisation shows that risk is very skewed for most crime types, such that some people are victimised repeatedly (Mayhew and Reilly, 2007). International research confirms that this is also true for sexual offences, such that women with a history of being sexually assaulted are at higher risk of future sexual victimisation than women who have never been sexually assaulted (Davis et al., 2006).

3.1.5    Criminal history of victims
Criminal history information was available for 98 percent of victims. Just over a fifth of victims (22 percent) had a prior criminal history, excluding traffic convictions. Most victims with a prior history only had conviction(s) for non-violent offences (66 percent), but 33 percent had previous violent convictions and six victims (one percent, five of whom were males) had a conviction for a sex offence.

Victims who had made previous allegations of violent victimisation were more likely to have criminal convictions (36 percent) than victims who had made no prior allegations (12 percent).

3.2    Offender profile

The 1,955 cases in the study involved 2,029 alleged offenders. The vast majority of cases (94 percent) involved a single offender and most of the rest involved two offenders. Only 2 percent of cases involved three or more offenders, with a maximum of seven offenders.

In the earlier study (Stace, 1983), cases with multiple offenders made up a larger percentage of cases (22 percent), with a notably high proportion of gang rapes by multiple offenders (15 percent of cases). In the current study, gang affiliation was not specifically recorded, but the file notes included a reference to the offender being a gang member in 3 percent of cases. Most of these cases had a single offender; only seven cases had multiple offenders and a reference to gang affiliations.

The following profile refers to all alleged offenders about whom some information was recorded (n = 1,432), including those cases where the alleged offender was not charged or convicted. Not all information could be collected for all offenders: the proportion who had the relevant data recorded is indicated for each variable in the following sections.

The analysis excludes the 8 percent of alleged offenders (n = 157) who were the subject of a documented ‘false complaint’ (i.e. where the alleged victim was charged or warned for making a false complaint), because alleged offender details were not collected in these cases. Information was not available for a further 440 offenders whose identity was unknown.

3.2.1    Demographic characteristics of offenders
Gender was recorded for 99.9 percent of the 1,432 known offenders in the sample, age was recorded for 98 percent and ethnicity was recorded for 94 percent.

Male offenders accounted for 99 percent of the known offenders. Only 11 offenders were females. Male offenders accounted for over 99 percent of offenders in cases involving female victims and 97 percent of offenders in cases involving male victims.

The distribution of offenders’ ages was spread more evenly than were victims’ ages (Figure 4), with 13 percent of offenders aged under 20 years, 32 percent in their 20s, 27 percent in their 30s, 20 percent in their 40s and 9 percent aged 50 years or over. Three percent of offenders were aged under 17 years. The median age of offenders was 31 years compared with 23 years for victims.

 

Figure 4: Age distribution of offenders and victims

 


Note: All victims were aged 16 years or over.

 

Of those offenders with ethnicity recorded, 49 percent were European, 30 percent were Māori, 12 percent were Pacific peoples, 6 percent were Indian/Asian and 3 percent were from other ethnic groups. Thus, Māori and Pacific peoples were over-represented among offenders relative to their proportions in the total population.

The ethnic group of the offender matched that of the victim in just under two-thirds of cases (Table 2).

 

Table 2: Ethnicity of offender, by victim's ethnicity

Offender ethnicity	Victim ethnicity
	Māori	Pacific	European	Other		Total
Sample size	329	75	713	83		1,200
	Offender ethnicity as a percentage of each victim ethnic group (where both recorded) (%)
Māori	57	15	22	13	30
Pacific	16	64	6	2	12
European	24	12	66	23	48
Other	3	9	6	61	9
Total	100	100	100	100	100

Matching ethnicity was particularly likely if the victim and offender were family (79 percent matched) or current or ex-partners (72 percent), but also occurred for two-thirds of boyfriends or dates, acquaintances and friends. Offenders who were strangers were the least likely to have a matching ethnicity to the victim (33 percent), while half of the ‘just met’ and ‘other known’ categories had matched ethnicity.

Employment status was recorded for 69 percent of offenders. Of these, 68 percent were employed.

3.2.2    Disability status of offenders
Disability status was recorded for 98 percent of offenders. Five percent of offenders were recorded as having some type of disability, of whom 62 percent had a psychiatric condition, 20 percent had a physical disability and 18 percent an intellectual impairment. Fourteen of the 16 cases in which the offender had an intellectual disability involved a victim with an intellectual disability. Almost half the cases in which the offender had a psychiatric disability involved a victim with a psychiatric or intellectual disability.

3.2.3    Criminal history of offenders
Criminal history status (excluding traffic convictions) was recorded for all but four offenders. Forty percent of offenders had no prior convictions, 23 percent had only non-violent prior convictions, 26 percent had prior convictions for violence but no prior sex offences, and 11 percent had a previous conviction for a sex offence.

Of the 155 offenders with a previous conviction for a sex offence, 25 percent had previous convictions for sexual violence only, 15 percent had previous convictions for other sex offences only, 43 percent had a history of sexual offences and other types of violence, and 17 percent had a history of sex offences and non-violent offences.

Sixteen percent of offenders had been the respondent on a protection order, not necessarily in relation to the victim, while seven offenders (half a percent) had been a protection order applicant.

3.3    Crime profile

This section examines the number and type of offences, the location and timing of the offending, and the circumstances of the offending for cases recorded as sexual violation. The following profile refers to all recorded cases, including those cases where the incident was subsequently recorded as a ‘false complaint’ or other ‘no offence’ outcome.

3.3.1    Number and type of offences
The 1,955 cases included a total of 2,888 sexual violation offences. Most cases involved one offence (82 percent), two offences (9 percent) or three offences (4 percent). Eight cases involved more than ten offences; one case involved 117 offences. Most of the multiple-offence cases involved a single offender, but 17 percent involved more than one offender.

Rape accounted for 56 percent of all offences or 68 percent of cases (based on the most serious offence in the case). Unlawful sexual connection accounted for 33 percent of offences and 22 percent of cases, attempted violation accounted for 8 percent of offences and 7 percent of cases, and other offences accounted for 3 percent of offences and cases. ‘Other offences’ included 24 cases of sexual exploitation of a significantly impaired person, 17 cases of inducing sexual connection, 9 cases of incest and 13 other offences.

3.3.2    Reporting
Ninety percent of cases had information on the offence date and reporting date. However, both these dates were recorded in months and years, rather than as specific dates. Therefore, it was not possible to identify how many hours or days had passed before the offence was reported to the police. The data did not identify which offences were reported by the victim and which were reported by someone else.

Based on the available data, 85 percent of incidents were recorded in the same month as the offence took place. Five percent were recorded in the following month (which could be a delay of anything from one day to 60 days). A further 5 percent of cases were recorded within six months and 5 percent after more than six months.

In the earlier New Zealand study (Stace, 1983), half the incidents were recorded by the police within one hour and 83 percent within 24 hours. The offence was reported by the victim in 52 percent of cases, by a relative or friend in 27 percent of cases and by a neighbour, stranger or other person in 21 percent of cases.

The time taken to report the incident is critical for forensic evidence. In London, around two-thirds of 677 incidents were reported within the critical time-frame (MPS, 2005). In Victoria, Australia, almost half of the incidents were reported within 24 hours (OWP, 2006). Several studies noted that incidents with a substantial lapse of time before reporting more often involved victims and offenders in intimate or close relationships (Feist et al., 2007; MPS, 2005; Lea et al., 2003).

3.3.3    Scene, location, day and time
Of the 89 percent of cases with scene recorded, two-thirds of incidents occurred in a dwelling (any dwelling, including the offender’s or victim’s home). A further 22 percent occurred in a public place and one percent each in an educational institution, a hospital, a prison or licensed premises.

Just over half the incidents in the earlier New Zealand study occurred in the victim’s or offender’s home (Stace, 1983).

Recent overseas studies also found a high proportion of offences occurring in a dwelling, as a consequence of the high proportion of offences committed by an offender known to the victim. For example, a study of several areas in England found that two-thirds of rapes occurred in either the offender’s or victim’s home or joint home (Feist et al., 2007).

The distribution by police district was:

• Auckland City 11 percent
• Bay of Plenty 7 percent
• Canterbury 15 percent
• Central 7 percent
• Counties Manukau 13 percent
• Eastern 6 percent   
•  Northland 3 percent
• Southern 8 percent
• Tasman 5 percent
• Waikato 8 percent
• Waitemata 6 percent
• Wellington 12 percent.

 

The most common days of occurrence were Sunday (24 percent) and Saturday (21 percent), followed by Friday (16 percent), Thursday (11 percent), Monday (10 percent), Wednesday (9 percent) and Tuesday (9 percent). Incidents most often occurred in the evening or early hours of the morning (Figure 5). Day was recorded for 80 percent of cases and offence time for 68 percent of cases.

 

Figure 5: Time of the day the incident occurred

 

3.3.4    Force, threats and injury
The interpretation of these results was complicated by the lack of relevant information in many file notes. For example, if no information was recorded about injuries, it was not possible to know whether it meant there was no injury or that the relevant information was missing from the file.

Table 3 shows a breakdown of the results, including cases with missing data for each factor, but excluding cases where the factor was recorded as not applicable (mainly ‘false complaints’).

Given that the factor was likely to be present in at least some cases where the data were missing, but assuming the recording of such information is more common in cases where the circumstance was present, the true rate of occurrence for each factor probably lies somewhere between the two ‘% yes’ estimates in Table 3.

For example, the offender made threats of violence against the victim in 9 percent of cases overall or in 16 percent of cases in which the relevant information was recorded. Two-thirds of threats were made to force submission at the time of the offence and a third were threats to deter disclosure or reporting, either at the time of the offence or after the offence. In a small proportion of cases, threats of violence were made against family or friends of the victim at the time of the offence.

 

Table 3: Force, threats and injury

Circumstance present	Yes	No	No data	Total cases	% yes, of known	% yes, of total	% no data
Victim threatened	155	822	810	1,787	16	9	45
Family/friends threatened	24	913	851	1,788	3	1	48
Force used	508	466	812	1,786	52	28	45
Weapon	92	1,313	360	1,765	7	5	20
Victim injured	478	1,100	212	1,790	30	27	12

Notes: ‘No data’ means there was no information in the file notes about the relevant factor. Total cases exclude non-applicable. The column label ‘% yes, of known’ is the percentage of cases where the factor was present, of cases in which relevant data were recorded.

Force was used against the victim in 28 to 52 percent of cases.

Cross-tabulation of the ‘force’ variable with the ‘injury’ variable suggested that a large proportion of the missing codes might relate to cases where force was used.

If cases involving injury were assumed to involve a degree of force by the offender (and this may not always be so), then the estimate of cases involving force increased to 41 percent to 64 percent of cases, which is more in line with international findings.

Weapons were used in about one in 20 cases. A knife or cutting tool was the most common weapon used (45 percent of specified weapons), while firearms were mentioned in only four cases.

The victim sustained physical injury during the incident in 27 percent to 30 percent of cases. Where the scale of the injury was recorded, the injury was relatively minor in 61 percent of cases (e.g. superficial cuts, grazes and bruises), moderate in 21 percent of cases (e.g. extensive bruising and injuries requiring treatment by a doctor) and serious in 18 percent of cases (e.g. vaginal or anal bleeding, life-threatening injuries, broken bones and injuries requiring hospitalisation). Thirty-three cases were recorded as involving a hospital visit.

Recent studies in England and Australia found that while some degree of force was used in the majority of cases, a minority of victims were physically injured (e.g. a third of cases (Feist et al., 2007), 28 percent (OWP, 2006) and 52 percent (Lievore, 2004)). Lack of physical resistance has many causes: victims may be frozen with fear, afraid to risk injury by resisting or physically unable to resist (Lievore, 2004). Weapons were used in only a minority of rape incidents (e.g. 4 percent of cases (Feist et al., 2007) and 20 percent (Lievore, 2004)).

The previous New Zealand study (Stace, 1983) reported that two-thirds of incidents involved physical force and a fifth involved threats, but serious physical injury was infrequent.

3.3.5    Alcohol and other drugs
The use of alcohol or other drugs was not recorded for the majority of offenders and for almost half the victims (Table 4). The offender was recorded as having had alcohol or drugs (or being in an environment that suggested this, such as a bar) in 533 cases, which equates to 79 percent of incidents where this information was recorded or 30 percent of all cases. The alcohol and/or drug percentages were similar for the victim. The use of alcohol and/or drugs by both offender and victim was particularly likely in cases for which the offender was a friend or acquaintance or had ‘just met’.

 

Table 4: Alcohol and/or drug use and uncertainty of violation

Circumstance present	Yes	No	No data	Total cases	% yes, of known	% yes, of total	% no data
Offender had alcohol/drugs	533	140	1,107	1,780	79	30	62
Victim had alcohol/drugs	762	250	938	1,950	75	39	48
Victim uncertain if violated	245	1,325	249	1,819	16	13	14

Notes: ‘No data’ means there was no information in the file notes about the relevant factor. Total cases exclude non-applicable. The column label ‘% yes, of known’ means the percentage of cases where the factor was present, of cases in which relevant data were recorded.

The victim was uncertain if a sexual violation had occurred in 245 cases (13 percent of all cases). In most of these cases, the victim was recorded as having had alcohol or drugs.

Alcohol use and drug use were not recorded separately, nor was the degree of intoxication, if any, or drug type recorded. Thus, the extent of legal or illegal drug use was unknown.

The ‘comment’ field, based on file notes summarised by the data extraction team, indicated that the victim was intoxicated in at least 196 cases or 10 percent of all cases, while illegal drug use by the victim (mainly cannabis use) was noted in at least 54 cases (3 percent of cases). In 62 cases (3 percent), the file notes indicated that the victim thought she had been drugged or her drink had been spiked. Only four of the latter cases included an indication of supporting evidence in the file notes (such as toxicology results or offender admission) and over a third of these 62 cases were cleared as ‘no offence’.

A strong association between alcohol and rape has been noted by many researchers in overseas studies, especially in relation to rapes by recent acquaintances. The victim or the suspect, or both, were under the influence of alcohol in around a third to half of cases in the English studies (Feist et al., 2007; HMIC and HMCPSI, 2007; Kelly et al., 2005). In Australia, the majority of victims had used alcohol and/or drugs around the time of the offence (OWP, 2006; Lievore, 2004). Kelly et al. (2005: 81) found that the presence of alcohol increased the level of injuries.

Alcohol consumption may negatively affect the chances of a successful prosecution due to the victim’s diminished recall of events and negative stereotyping of the victim, although not all studies do find a relationship between alcohol use and conviction rates (Lievore, 2004). An intoxicated person cannot legally consent to sex.

Alcohol was much more often a factor in rapes than drugs were, but precise estimates are difficult to obtain because blood and urine samples are not always taken. Drug use, where it was documented, was most often voluntary use by the victim, rather than drugs administered without knowledge by the offender, and the use of ‘date-rape’ drugs was rare (Feist et al., 2007; HMIC and HMCPSI, 2007).

3.3.6    Evidence
All sexual violation offences reported to the police should be investigated and evidence collected as appropriate to the circumstances. This may include, for example, medical examination of the victim, witness accounts, closed-circuit television footage, scene identification and photographs, and collection of physical evidence such as DNA. The policy is for all victims to undergo a medical examination, except where there has been a delay in reporting or there was no penetration or in other circumstances where no physical evidence is expected. Medical examination is undertaken by accredited doctors from Doctors for Sexual Abuse Care or police medical officers.

Information on whether the victim had had a medical examination was recorded for only 35 percent of cases in the current study. For the 684 cases where this information was recorded, 56 percent had had a medical examination and 44 percent had not.

For the 306 cases where there was no examination, the reasons were split between the incident having occurred too long ago (27 percent), victim refusal (27 percent) and other unspecified reasons (46 percent).

Evidence was recorded in the summary database only if it related to a link between the victim and offender.

• DNA evidence linking the offender to the victim was noted in the file of 59 cases and lack of DNA evidence was noted in the file of 202 cases. The vast majority of cases (85 percent) had no information on whether such evidence was available.
• Other forensic evidence linking the offender to the victim was recorded in 114 cases.
• Information on whether there was any evidence from witnesses was recorded for 76 percent of cases. For these cases, 60 percent had witnesses and 40 percent did not.
•  Only 13 cases were recorded as having surveillance evidence linking the offender to the victim, such as security camera footage.

 

There are many reasons why DNA evidence would not be collected or analysed, including a lack of DNA in cases where penetration and ejaculation did not occur or the victim bathed or showered, or the offence was not reported promptly. In other cases, intercourse was not in dispute (i.e. the main issue was consent).

3.4    Summary of victim, offender and case characteristics

The victims/complainants in cases recorded as sexual violation had the following characteristics:

• 95 percent were women
• they were mainly young (a third were aged 16 to 19 years and over half were aged under 25 years), but not as young on average as victims in the 1981 New Zealand study (Stace, 1983)
• they over-represented Māori and under-represented Asian and other ethnic groups relative to their proportions in the total population
• a substantial minority had some type of disability – psychiatric (7 percent), intellectual (6 percent) or physical (one percent)
• 43 percent had made previous allegations of violent victimisation (not necessarily against the same offender) relating to domestic violence (71 percent), sexual violence (21 percent) or other types of violence (7 percent)
• 22 percent had a prior criminal history (excluding traffic convictions), although the majority of these had only non-violent convictions.

 
The majority of offenders were previously known to the victim, with stranger assaults accounting for just 16 percent of cases and offenders just met (within the last 24 hours) accounting for a further 15 percent of cases. A third of cases involved victims and offenders with intimate relationships (family, current partners or ex-partners) and 37 percent of cases involved other known offenders (such as friends, boyfriends, acquaintances, people known through work and caregivers). The 1981 New Zealand study involved a higher proportion of stranger attacks (Stace, 1983).

The alleged offenders in cases recorded as sexual violation had the following characteristics (based on cases for which offender information was known):

• 99 percent were men
• they were more evenly spread by age than were victims, with a median offender age of 31 years compared with 23 years for victims
• they over-represented Māori and Pacific ethnic groups relative to their proportions in the total population
• a small proportion had some type of disability – psychiatric (3 percent), intellectual (one percent) or physical (one percent)
• 60 percent had a prior criminal history (excluding traffic convictions), including 23 percent with non-violent prior convictions, 26 percent with prior convictions for violence but no prior sex offences, and 11 percent with a previous conviction for a sex offence.

 

The circumstances of the offending were as follows:

• two-thirds of incidents occurred in a dwelling, 22 percent in a public place, and one percent each in an educational institution, a hospital, a prison or licensed premises
• almost half the offending occurred during a weekend, and most incidents occurred in the evening or early hours of the morning
• around half the incidents involved some force or threat
• about 30 percent of victims were physically injured, with around 12 percent suffering moderate or severe injuries
• the use of alcohol or drugs by the victim or offender was not consistently recorded, but was a feature of a substantial proportion of incidents
• the victim was uncertain whether a sexual violation had occurred in 13 percent of all cases, and in most of these cases the victim was recorded as having had alcohol or drugs
• just over half of victims had a medical examination, in cases for which the medical examination status was known
• the availability of forensic and other evidence was not consistently recorded, but evidence linking the offender to the victim (other than witness statements) was recorded for only a minority of cases.
  

 

---

Disclaimer

This report was commissioned by the Ministry of Women’s Affairs. The views, opinions and conclusions expressed in the report are intended to inform and stimulate wider debate. They do not represent government policy.
 

Published in September 2009
by the Ministry of Women’s Affairs | Minitatanga Mō Ngā Wāhine
PO Box 10 049, Wellington, 6143, New Zealand
Phone: 0064 4 915 7112
Fax: 0064 4 916 1604
Email: mwa@mwa.govt.nz
Website: www.mwa.govt.nz

ISBN 978-0-478252-48-4 (Print)
ISBN 978-0-478252-49-1 (Digital)