Identifying first-time family violence perpetrators: The usefulness and utility of categorisations based on police offence records

Foreword | The benefits associated with focusing strategies and interventions on family violence offenders early in their offending careers are well established. Yet accurately identifying first-time family violence perpetrators is difficult because of the high under-reporting of family violence. This study involved the analysis of police narratives completed for first-time family violence perpetrators, as identified through police offence records to determine their accuracy in identifying first-time family violence offenders. The findings demonstrate that a reliance on the use of police offence record information in isolation may lead to many ‘false negatives’ and a failure to identify recidivist offenders who may reduce the effectiveness of tailored intervention programs. The use of multiple sources of data and better use of police narratives within offence records are recommended as ways to more accurately distinguish between first-time offenders and recidivist offenders.

Adam Tomison
Director

Since emerging as an issue of social and political importance in the 1970s, it is now well established that family violence is one of the most prevalent and costly (personally, socially and economically) forms of violent crime in Australia (ABS 2013; Mouzos & Makkai 2004; McPhedran & Baker 2012; People 2005) and internationally (Jewell & Wormith 2010; Sartin, Hansen & Huss 2006; Tjaden & Thoennes 2000). Correspondingly, a large and expanding body of research has attempted to understand and explain the occurrence of family violence, with a view to preventing and minimising its incidence in the future.

Within the literature, there has been consistent interest in the group of offenders hereafter referred to as ‘first-time family violence perpetrators’; that is, family violence perpetrators (FVPs) who do not have a history of being violent or abusive towards their current or past intimate partners prior to the ‘index’ (first) offence. Interest in this group of FVPs is founded on consistent evidence that suggests that first-time FVPs are less likely than those with a more extensive offending history to reoffend (Hamilton & Worthen 2011; Trujillo & Ross 2005; Waaland & Keeley 1985) and are more likely to complete and benefit from treatment (Babcock & Steiner 1999; Daly & Pelowski 2000; Gover et al. 2011; Jewell & Wormith 2010). This is primarily because the behaviours and attitudes of this group of FVPs are not as ‘entrenched’ as those of other FVPs and so may be less stable and more susceptible to modification and affected by external influences, such as contact with the criminal justice system and treatment programs (National Crime Prevention 1999). Consequently, there is understandable interest among professionals who are responsible for responding to family violence matters to accurately identify first-time FVPs.

Identification provides criminal justice and treatment agencies with an opportunity to intervene early in what could be—if ignored or not prioritised—a significant and long-term violent criminal career (National Crime Prevention 1999). By intervening early and attempting to address the underlying causes of the violent behaviour, individual offenders may be diverted away from a criminal career that would have significant and negative consequences for their intimate partners, families and the community more broadly (COAG 2010).

While there are a number of ways family violence (also referred to as spousal assault, domestic violence, intimate partner violence and wife battering) has been defined in Australia and overseas, for the purpose of this paper, the definition provided under
s 7 of the Family Violence Act (2004) (Tas) has been used. This definition was used primarily because the data used for this study was extracted from Tasmanian data collection systems.

Under the Family Violence Act (2004) (Tas), any of the following types of conduct committed by a person, directly or indirectly, against that person’s spouse or partner is a form of family violence:

• assault (including sexual assault);
• threats, coercion, intimidation or verbal abuse;
• abduction;
• stalking;
• economic abuse (eg withholding money);
• emotional abuse or intimidation; or
• contravening an external family violence order (made by a court of another jurisdiction), an interim Family Violence Order (FVO; made by the Tasmanian court system), an FVO or a Police Family Violence Order.

First-time offenders and criminal justice decision-making processes

When responding to a family violence matter, the various criminal justice parties involved are, at various stages, required to consider the reoffending ‘risk’ posed by the FVP (Bachman & Croker 1995). A consistent finding that emerges from the family violence recidivism literature is that first-time FVPs, when compared with those who have a history of family violence, are less likely to reoffend (see Gendreau, Little & Goggin 1996 for a review). Correspondingly, one of the factors that different criminal justice representatives may take into account when deciding how best to handle and resolve family violence matters is the FVP’s history of family violence offending.

With regards to initial contact with the criminal justice system, Trujillo and Ross (2005) analysed over 500 Family Violence Risk Assessment and Management Reports completed by Victoria Police officers and found that after controlling for a range of factors (and when compared with repeat offenders), first-time FVPs were statistically less likely to have charges laid against them. Hamilton and Worthen (2011) analysed all of the Domestic Violence Police Reporting Forms completed by Rhode Island police for matters involving heterosexual couples for the 2002–04 period. It was found that after controlling for a range of legal and extra-legal factors (eg presence of visible injury), offenders who had assaulted their partner on a previous occasion were almost twice (OR=1.97) as likely to be arrested than those who were first-time offenders (Hamilton & Worthen 2011).

Other research undertaken by Waaland and Keeley (1985) involved asking 26 male and 10 female US police officers from Oregon to consider a number of family violence vignettes and indicate whether they would arrest the described FVP. Analysis of participant responses found that the prior abusive behaviours of the offender had a small but significant impact on police decision-making processes (Waaland & Keeley 1985).

Further, over the last 20 years, a number of Australian and international police departments have required or encouraged frontline officers to complete formal risk-assessment tools when responding to family violence matters. Many of these tools (eg the Risk Assessment Screening Tool that is currently used by Tasmania Police), include items related to the family violence offending history of the FVP (Dutton & Kropp 2000; Felson, Ackerman & Gallagher 2005; Mason & Julian 2009; McPhedran & Baker 2012; Storey & Strand 2012). Therefore, when completing these risk assessment procedures, responding officers are required to consider the family violence offending history of FVPs and to take this into account when considering how to best handle and resolve the matter.

Once the decision has been made to charge an individual with family violence offences, the FVP’s family violence history also may influence prosecutorial and sentencing outcomes (see Henning & Feder 2005 for a review). For example, research undertaken by the NSW Bureau of Crime Statistics and Research found that FVPs who had been charged with family violence offences in New South Wales but had not been convicted of a violent offence within the preceding 10 year period were less likely to receive a sentence of imprisonment when compared with defendants who did have a conviction history (Ringland & Fitzgerald 2010).

Hartman and Belknap (2003) presented 62 municipal court judges, public prosecutors and defenders working in a large urban area in the mid-west United States with a series of legal and non-legal factors and asked them to identify the extent to which each should be taken into consideration when determining whether to prosecute or convict an FVP. Analysis of the provided responses found that the FVP’s family violence offending history was one of the highest-ranked factors for both prosecution and conviction decisions, although ratings did vary between professional groups. Finally, a study undertaken by Henning and Feder (2005) that involved the analysis of court records (n=5,461) collected for all criminal defendants arrested for a family violence offence in Shelby County, Alabama (United States) over a 12 month period found that offenders who had previously been arrested for family violence offences were statistically less likely to be released on bail.

First-time offenders and treatment attrition and effectiveness

First-time FVPs have also received a lot of attention within the family violence treatment and rehabilitation sector because of evidence that, as a group, they respond more positively to treatment (both court-mandated and voluntary) and are less likely to ‘drop out’ than FVPs with a history of family violence offending (Babcock & Steiner 1999; Daly & Pelowski 2000; Gover et al. 2011; Jewell & Wormith 2010). Babcock and Steiner (1999) analysed the treatment completion rates of over 300 men referred to one of 11 accredited family violence treatment programs operating in Seattle (United States) and found that first-time FVPs were statistically more likely to complete treatment than not (89.5% cf 77.2%). A more recent study undertaken by Gover and colleagues (2011) involved the analysis of self-report data collected from over 4,000 FVPs (81% male) who were residing in Colorado and had been referred to one of a range of treatment services. After controlling for a range of variables (eg age), multivariate analysis indicated that treatment completers were statistically more likely to be first-time FVPs when compared with drop-outs (OR=1.46), although this effect was only identified for male clients (Gover et al. 2011). Finally, a meta-analysis of 30 family violence program attrition studies undertaken between 1985 and 2010 found that prior family violence offending had a small but negative effect on treatment attrition (Jewell & Wormith 2010).

Aim of the study

Given that a FVP’s history of offending may influence how criminal justice and treatment agencies respond to, and attempt to resolve, the offending behaviour, accurately identifying first-time FVPs is important.

Individuals and agencies that are responsible for responding to family violence matters and offenders may use a range of information to determine whether a specific FVP has a history of family violence offending. This includes interviews with the victim, offender and other family members, analysis of case files and notes, speaking with colleagues and the seeking access to the FVP’s police offence record (if possible). However, in determining whether a specific FVP has a history of family violence offending, some agencies, in particular law enforcement and court personnel, may rely on police offence records more than other sources of information. This is for a range of reasons, including accessibility and privacy concerns (some agencies may not be able to access case files maintained by other agencies), and the unwillingness or inability of the victim and/or offender to provide information. For example, when police first arrive at the scene of a family violence matter, the offender may be hostile, belligerent, uncooperative or intoxicated. The victim may also be unwilling or incapable of providing information to first responders. Alternatively, the victim and/or offender may be absent when the police arrive. In these situations, law enforcement officers may not be able to interview the victim and/or offender to potentially determine whether the FVP has a history of family violence offending or not.

Some agencies may even prefer to use police offence records rather than other information because its reliability is not dependent on the truthfulness of different parties. Victims and offenders may want to hide the family violence offending history of the FVP, particularly if they believe it will influence whether the offender is arrested or convicted, the length of their sentence and eligibility for treatment. Further, victims may be intimidated into silence by their partner who may threaten them or their families with further violence if they disclose. It is also worth noting that information provided by victims in relation to the FVP’s history of family violence offending may be limited if they are not aware of the offender’s use of violence against previous partners.

However, while there appear to be valid reasons for using police offence records to differentiate between first-time and repeat FVPs, the accuracy of information extracted from these reports regarding the nature and breadth of an FVP’s family violence offending history may also be limited. This is primarily because many crimes are never reported to, or detected by, the police and this is particularly the case for violence perpetrated by intimates (ABS 2004; Bachman & Coker 1995; Barnett 2001; Felson et al. 2002; Felson, Ackerman & Gallagher 2005; Gracia 2004; McPhedran & Baker 2012; Sartin, Hansen & Huss 2006; Tjaden & Thoennes 2000). For example, of female respondents who participated in the most recent Personal Safety Survey administered by the Australian Bureau of Statistics who said their current partner had been violent towards them, only 20 percent claimed that they had reported the incident to the police (ABS 2013). This gap between reported and unreported victimisation is often referred to as the ‘dark figure’ of crime and is a limitation of all police offence record data (Coleman & Moynihan 1996).

Considering that police offence records only include information on family violence matters that are reported to or detected by police, the utility of this data to identify first-time FVPs may be limited. However, in order to test this, the aim of the current study was to determine the usefulness and utility of using police offence data for the purpose of identifying first-time FVPs.

This study represents one of the first attempts to quantify the gap between actual and police-reported family violence offending in an Australian context using police narrative data. Previous studies of this nature have been conducted predominantly in the United States, which considering the differences in legislative definitions of family violence, arrest policies and police practices, have very little relevance to the Australian context.

Method

This study involved the analysis of administrative data extracted from the Family Violence Management System (FVMS) that is currently maintained by Tasmania Police. Consequently, this study focused exclusively on a sample of FVPs reported in Tasmania. Introduced as part of the Safe at Home reforms that were implemented in Tasmania in 2005, the FVMS is a purpose-built database that stores a variety of information on all Family Violence Incidents (FVIs) that are reported to Tasmania Police and result in a callout. It is important to note that FVIs that do not result in an arrest or charge are included in the FVMS, as are ‘argument only’ incidents—callouts that do not meet the threshold for family violence as defined under the Family Violence Act (2004) (Tas).

Prior to the implementation of the FVMS in 2005, information on family violence offences occurring in Tasmania was collected through different means. Specifically, police officers responding to a family violence matter were required to complete a Domestic Violence Incident Report (DVIR; Knowles 1996).

The research team received an extract of all FVIs recorded in the FVMS for the period 1 July 2010—30 June 2011. This resulted in the identification of 1,975 unique FVPs and 2,672 FVIs. This initial sample of FVIs was then divided between those involving a first-time FVP or a repeat offender. FVIs were coded as involving a first-time FVP if the identified offender had no other FVIs recorded against them (as an offender) in the FVMS for the period between 1 June 2005 (when the FVMS was implemented) and 30 June 2010. From this preliminary cohort of first-time FVPs (n=1,206), random sampling methods were used to select 100 FVIs for closer analysis. The sample size was, in part, determined after considering the anticipated level of analysis involved and the timeframe allocated to the project.

The research team requested and received all free-text information recorded in the FVMS for this cohort of FVIs, including the FVI narrative, which was completed by the responding police officer (hereafter referred to as police narratives). At a minimum, the police narratives included information about:

• the relationship history between the victim and offender;
• the offender’s history of family violence offending (as recorded in the FVMS and DVIR database and unreported historical offences);
• the FVP’s Risk Assessment Screening Tool ‘score’;
• the sequence of events that led up to the violent behaviour; and
• the police response (eg arrest, application for a FVO or PFVO etc).

Responding officers appeared to use two sources of information to complete the narratives—the victim and FVP’s police offence records; and where possible, interviews with the victim, offender and any relevant witnesses.

Preliminary analysis of the police narratives resulted in three matters being removed from the sample. While there were no other records in the FVMS for the FVPs identified in these matters, they were the subject of a DVIR and so were identified as having a history of offending that had been reported to or detected by Tasmania Police. This left a final sample of 97 FVIs.

Of these 97 FVIs:

• 55 percent involved victims and offenders who were currently in a relationship with one another (45% were separated or divorced);
• approximately three-quarters (76%) involved a male offender; and
• 73 percent involved a female victim.

The police narratives for the final sample of FVIs was analysed by a member of the research team (the primary coder) with previous experience analysing written texts. The first step in the analysis required the primary coder to substantiate whether any of the FVPs in the sample were not actually first-time offenders. This involved the primary coder reading each of the narratives in their entirety to identify whether police had been provided with information about unreported historical family violence offences and/or historical family violence offences that were reported to police in another jurisdiction and so would not be included in the FVMS or DVIR. Reading the narratives in their entirety, rather than looking for key words and phrases (eg ‘prior violence’), was deemed necessary because of variations in police reporting styles and language.

The first scan of the narratives resulted in each of the FVIs being allocated to one of three groups—history of family violence offending, no history of family violence offending or unclear. FVIs were labelled as unclear if the researcher was unsure whether the information included in the narrative met the threshold for disconfirming the first-time FVP label.

The threshold for including FVIs in the history of family violence offending group was deliberately set fairly low, being limited to any information suggesting that the FVP had a prior history of violence, regardless of whether the violence had been witnessed by others. This was decided upon in light of research that has consistently demonstrated that rates of non-reporting for family violence are high (Barnett 2001; Felson, Ackerman & Gallagher 2005; Felson et al. 2002; Gracia 2004; Mouzos & Makkai 2004; Tjaden & Thoennes 2000); a trend that will be discussed in more depth later in this paper.

For an FVI to be included in the history of family violence offending group, the disclosure had to come from a ‘reasonable’ source—a person (or persons) who, it would be reasonable to assume, were in a position to identify whether the FVP had a history of prior violence. Such persons included the victim, the FVP themselves, a family member or a person who was or had lived with the FVP or victim previously (eg children or close friends). If it was unclear who had made the disclosure to police, the FVI was not included in the history of family violence offending group.

The second stage of the analysis involved the primary coder re-reading the narratives to confirm correct interpretation of the data and to allocate matters to more meaningful groups. As demonstrated in Figure 1, FVIs allocated to the history of family violence offending group were further subcategorised into either history of unreported family violence offending or history of reported family violence offending occurring in another jurisdiction. FVIs allocated to the no history of violence group were similarly subcategorised into either history of family violence offending (history of argument-only incidents) or no history of family violence offending (or argument-only incidents).

The narratives for FVIs labelled as unclear were also read again by the primary coder. If the second reading did not clarify whether an FVP had a history of violence or not, another research team member (secondary coder) was asked to read the narrative and the two coders discussed their interpretations and reached a consensus on the most appropriate allocation. This was required for three FVIs. However, generally, the information included in the narratives was sufficiently detailed to allow for the consistent and accurate allocation of FVIs to the categories identified in Figure 1.

Findings

As demonstrated in Figure 1, a third (33%) of the first-time FVPs who were identified through the FVMS and DVIR database had a history of family violence offending. Thirty of the FVIs (31%) who had been involved an FVP had been violent towards their partner in the past but it had not been reported to or detected by the police. Past offending behaviours included physical violence (eg shoving and hitting), threatening and intimidating behaviour (eg making threats to harm the victim, their children or their families), stalking and economic abuse. Meanwhile, two of the FVIs involved an FVP had been reported for family violence offences that occurred in another jurisdiction. It appeared that in these two matters, the police became aware of the FVP’s history of violence because a party informed them that they had moved to Tasmania from interstate, ostensibly prompting the officer to lodge an inquiry with that jurisdiction’s law enforcement agency or through the National Police Referencing System that is currently managed by CrimTrac.

Of the 65 FVIs (67%) that involved an FVP who did not appear to have a history of family violence offending, a quarter (n=16) were the subject of an argument-only FVI. The remaining 49 FVIs involved an FVP who:

• did not appear to have a history of unreported family violence offending;
• had not been reported to the police for family violence offences occurring in other jurisdictions; and
• was not the subject of an argument-only incident.

Discussion

For every three offenders identified through police offence records as first-time FVPs, only one actually had a history of family violence offending. In other words, of the 97 FVIs included in the analysis, only 65 (67%) were ‘accurately’ identified through the FVMS and DVIR database as involving a first-time FVP. Of the remaining 32 FVIs, one in three (31%) involved an FVP who had been violent towards their partner for a period of time before it was reported to or detected by the police. Although the details provided in the police narratives do not specify exactly how long victims believed that they had been subjected to the violent and abusive behaviour of the FVP, the collective accounts of victims suggested that offenders had been violent for anywhere from a matter of weeks through to the entire period of the relationship.

Figure 1 Classifications of police narratives [see PDF for figure]

Source: AIC FVMS 2010–11

This finding was not unexpected considering that many crimes are never reported to or detected by the police and this is particularly the case for violent offences perpetrated by intimates (Barnett 2001; Felson et al. 2002; Gracia 2004; Sartin, Hansen & Huss 2006). Reasons for not reporting family violence offences to the police include:

• perceptions that the offence is too trivial or minor to bother the police with;
• privacy concerns (particularly problematic within rural and Indigenous communities);
• feelings of embarrassment, humiliation and shame;
• fear of not being believed and/or blamed for the violence;
• fear of retribution from the offender and/or their family;
• fear that by making a report child protective services will also become involved;
• desire to protect the offender and the family unit; and
• the victim’s economic dependence on the offender (Barnett 2001; Felson, Ackerman & Gallagher 2005; Felson et al. 2002; Gracia 2004; Mouzos & Makkai 2004; Tjaden & Thoennes 2000).

There is also evidence that while legislative definitions of family violence have expanded over the last 40 years to include different forms of non-physical violence (eg stalking, emotional abuse, threatening/harming pets and economic abuse), these behaviours are less recognisable by victims, offenders and the community more broadly as abusive, violent and criminal (Barnett 2001; Mouzos & Makkai 2004). For example, of those women who participated in the Australian component of the International Violence Against Women Survey and reported being abused by their current husband or partner, only 11 percent thought that the behaviour was criminal and another 52 percent said it was ‘something that just happened’ (Mouzos & Makkai 2004). The inability of victims to identify the behaviours of their intimate partners as violent and abusive, and/or the minimisation of these behaviours, has obvious implications for their reporting and help-seeking behaviours.

The finding that two FVIs involved an FVP had been reported for family violence offences in another jurisdiction was not unexpected. While most Australian police services have implemented intra-jurisdictional data sharing protocols and have access to the National Police Referencing System, there is still very little integration between the offence history records maintained by different Australian jurisdictions. This is for a range of reasons, including differing definitions of family violence, reporting thresholds and privacy concerns (McPhedran & Baker 2012). Consequently, while information concerning offences perpetrated in other jurisdictions may be recorded in police narratives, it is not recorded in any consistent manner in the police offence record data.

Finally, it is noteworthy that 25 percent of ‘true’ first-time FVPs had been reported to or detected by the police on prior occasions for argument-only incidents. While argument-only incidents did not involve actual physical violence, they do indicate that this group of offenders have displayed behaviours (eg shouting and screaming at their partner) that were of sufficient concern that the police were required to respond. According to escalation models of family violence offending, argument-only incidents may be a precursor for family violence occurring in the future (McPhedran & Baker 2012). Alternatively, it may indicate that family violence offending is already occurring, but has not been detected by or reported to the police. However, testing these hypotheses was beyond the scope of the current study.

Implications for policy and practice

The findings from the literature review indicated that one of the factors that criminal justice and treatment agency representatives may take into consideration when determining how best to respond to individual family violence offenders is whether the FVP has a history of family violence offending. While many professionals may take a range of information into account when differentiating between first-time and repeat FVPs, for a number of reasons they may base their classification of the offender mainly on the offender’s police offence record (where they have delegated permission to access this information). The findings from this study suggest that caution should be exercised when using information extracted from police offence records in isolation and when considering the results of any subsequently undertaken formalised decision-making process (eg risk assessment tool) that includes the consideration of this information.

Information included in police narratives may provide a more accurate picture of the offending histories of FVPs than police offence records alone. Narratives included information concerning not only the FVP’s reported history of violent offending, but historical offences that had not been reported to the police and offences that had been committed in other jurisdictions. While this information may be uncovered through other means, such as directly speaking to the victim or offender, there may be benefit in using information included in police narratives to ‘flag’ issues that are of relevance to different agencies and parties (eg police prosecutors, Magistrates and treatment practitioners) in police offence history records. Specific flags may include ‘historical family violence offences—unreported’ and ‘offending in another jurisdiction’. The inclusion of such ‘flags’ in police offence records would ensure that key information collected by responding officers through conversations with the victim and other parties are communicated to other relevant involved parties.

Limitations

While the findings from this study indicated that one in three first-time FVPs identified through police offence records may have a history of family violence, there is reason to suspect that this may actually be an underestimation of prior offending within this cohort. In some cases, the ability of responding officers to interview the victim, offender and any other relevant witnesses was limited by the intoxication of one or all parties or their hostility and aggression. Consequently, in these cases, the quality and amount of narrative information provided by the responding officers was quite limited. Further, officers responding to FVIs included in the sample were only able to speak to the current partner, meaning that the offender’s use of violence against previous partners was not included in this analysis. Finally, while officers were given basic guidelines in terms of what they were asked to include in the narrative (see Methods section above), there was some variation in how this information was recorded and the level of detail provided. Consequently, while it appears that most, if not all, of the narratives addressed the question of whether there was any unreported family violence occurring, it may not have been recorded or included for a range of reasons.

It is also worth noting that the current study used data extracted from the FVMS, which is maintained by Tasmania Police. The data collection protocols associated with the FVMS may be inconsistent with the data collected by other Australian and international jurisdictions. Therefore, the findings should not be taken as representative of the experiences of other jurisdictions, although perhaps it may be indicative.

Conclusion

As a group, first-time FVPs have been a source of interest within the family violence sector for the last 40 years. This is primarily because of evidence that suggests that as a group they are less likely to reoffend and more likely to complete and benefit from treatment. This may be because the behaviours and attitudes that facilitate and encourage the use of family violence are less stable within this cohort, are more able to be manipulated and are more responsive to external influences, such as contact with the criminal justice system and treatment. Therefore, intervening early with first-time FVPs may reduce the likelihood of future offending.

However, determinations made by agencies and individuals responsible for responding to and managing family violence offences as to whether an individual is a first-time FVP has typically involved a reliance on police offence records. Considering that most crimes, particularly those perpetrated by intimates, are never reported to or detected by police, there are some obvious issues associated with using police offence records to identify first-time FVPs, although this has been a relatively untested proposition.

In the current study, one in every three first-time FVPs identified through police offence records actually had a history of offending (unreported or otherwise). In addition, 25 percent of the 65 ‘true’ first-time FVPs were involved in argument-only incidents, which may be a precursor or indicator of family violence offending.

The findings from this study have implications for both criminal justice agencies and practitioners who come into contact with FVPs at different stages. If the information they use to determine the ‘risk’ posed by an FVP is based solely or predominantly on police offence records (where available and accessible), the effectiveness of program responses to prevent reoffending may be affected by the inclusion of unsuitable offenders in FVP programs. Giving consideration to expanding the information police officers should attempt to collect may go some way towards reducing the false identification of FVPs. Other practitioners may develop a more accurate identification of offending histories by actively seeking information from a range of sources in addition to police records.

References

All URLs correct at August 2014

• Australian Bureau of Statistics (ABS) 2004. Information paper: Measuring crime victimisation, Australia: The impact of different data collection methodologies, 2002. Canberra: ABS. http://www.abs.gov.au/ausstats/abs@.nsf/mf/4522.0.55.001#CRIME%20VICTIMISATION%20DATA%20SOURCES
• Australian Bureau of Statistics (ABS) 2013. Personal safety, Australia, 2012. http://www.abs.gov.au/ausstats/abs@.nsf/mf/4906.0
• Bachman R & Coker A 1995. Police involvement in domestic violence: The interactive effects of victim injury: Offender’s history of violence, and race. Violence and Victims 10(2): 91–106
• Babcock J & Steiner R 1999. Relationships between treatment, incarceration, and recidivism of battering: A program evaluation of Seattle’s coordinated community response to domestic violence. Journal of Family Psychology 13(1): 46–59
• Barnett O 2001. Why battered women do not leave Part 2: External inhibiting factors—social support and internal inhibiting factors. Trauma, Violence & Abuse 2(1): 3–35
• Coleman C & Moynihan J 1996. Understanding crime data: Haunted by the dark figure. Buckingham: Open University Press
• Council of Australian Governments (COAG) 2010. National plan to reduce violence against women and the children. COAG: Canberra. http://www.dss.gov.au/our-responsibilities/women/programs-services/reducing-violence/the-national-plan-to-reduce-violence-against-women-and-their-children/national-plan-to-reduce-violence-against-women-and-their-children
• Daly J & Pelowski S 2000. Predictors of dropout among men who batter: A review of studies with implications for research and practice. Violence & Victims 15(2): 137–160
• Dutton D & Kropp PR 2000. A review of domestic violence risk instruments. Trauma, Violence & Abuse 1(2): 171–181
• Felson R et al. 2002. Reasons for reporting and not reporting domestic violence to the police. Criminology 40(3): 617–648
• Felson R, Ackerman J & Gallagher C 2005. Police intervention and the repeat of domestic assault. Criminology 43(3): 563–588
• Gendreau P, Little T & Goggin C 1996. A meta-analysis of the predictors of adult offender recidivism: What works! Criminology 34(4): 575–608
• Gracia E 2004. Unreported cases of domestic violence against women: Towards an epidemiology of social silence, tolerance, and inhibition. Journal of Epidemiology and Community Health 58(7): 536–537
• Gover A et al. 2011. Factors related to the completion of domestic violence offender treatment: The Colorado experience. Victims & Offenders: An International Journal of Evidence -based Research, Policy and Practice 6(2): 137–156
• Hamilton M & Worthen M 2011. Sex disparities in arrest outcomes for domestic violence. Journal of Interpersonal Violence 26(8): 1559–1578
• Hartman J & Belknap J 2003. Beyond the gatekeepers: Court professionals’ self-reported attitudes about and experiences with misdemeanor domestic violence cases. Criminal Justice and Behavior 30(3): 349–373
• Henning K & Feder L 2005. Criminal prosecution of domestic violence offenses: An investigation of factors predictive of court outcomes. Criminal Justice and Behavior 32(6): 612–642
• Jewell L & Wormith S 2010. Variables associated with attrition from domestic violence treatment programs targeting male batterers: A meta-analysis. Criminal Justice and Behavior 37(10): 1086–1113
• Knowles J 1996. Police culture and the handling of domestic violence: An urban/rural comparison. Canberra: Criminology Research Council. http://www.criminologyresearchcouncil.gov.au/reports/41-93-4.pdf
• Mason R & Julian R 2009. Analysis of the Tasmania Police risk assessment screening tool (RAST): Final report. Sandy Bay: Tasmanian Institute of Law Enforcement Studies, University of Tasmania. http://www.safeathome.tas.gov.au/__data/assets/pdf_file/0011/142310/RAST_Report_Analysis_of_Risk_Assessment_Screening_Tool.pdf
• McPhedran S & Baker J 2012. Lethal and non-lethal violence against women in Australia: Measurement challenges, conceptual frameworks, and limitations in knowledge. Violence against Women 18(8): 958–972
• Mouzos J & Makkai T 2004. Women’s experiences of male violence: Findings from the Australian component of the International Violence Against Women Survey (IVAWS). Research and Public Policy series no. 56. Canberra: Australian Institute of Criminology. http://www.aic.gov.au/publications/current%20series/rpp/41-60/rpp56.html
• National Crime Prevention 1999. Pathways to prevention: Developmental and early intervention approaches to crime in Australia. Canberra: National Crime Prevention, Attorney-General’s Department
• People J 2005. Trends and patterns in domestic violence assaults. Contemporary Issues in Crime and Justice no. 89. Sydney: NSW Bureau of Crime Statistics and Research. http://www.bocsar.nsw.gov.au/lawlink/bocsar/ll_bocsar.nsf/vwFiles/CJB89.pdf/$file/CJB89.pdf
• Ringland C & Fitzgerald J 2010. Factors which influence the sentencing of domestic violence offenders. Crime and Justice Statistics Brief no. 48. Sydney: NSW Bureau of Crime Statistics and Research
• Sartin R, Hansen D & Huss M 2006. Domestic violence treatment response and recidivism: A review and implications for the study of family violence. Aggression and Violent Behaviour 11(5): 425–440
• Storey J & Strand S 2012. The characteristics and violence risk management of women arrested by the police for intimate partner violence. European Journal of Criminology 9(6): 636–651
• Tjaden P & Thoennes N 2000. Extent, nature and consequences of intimate partner violence series: Research report: Findings from the National Violence Against Women Survey. Washington: US Department of Justice, Office of Justice Programs. https://www.ncjrs.gov/pdffiles1/nij/181867.pdf
• Trujillo M & Ross S 2008. Police response to domestic violence: Making decisions about risk and risk management. Journal of Interpersonal Violence 23(4): 454–473
• Waaland P & Keeley S 1985. Police decision making in wife abuse: The impact of legal and extralegal factors. Law and Human Behavior 9(4): 355–366

About the authors

Hayley Boxall is an A/g Snr Research Analyst in the Crime Prevention and Evaluation Research team at the AIC.

Dr Lisa Rosevear is a former Research Analyst with the AIC.

Dr Jason Payne is a former Research Manager with the AIC and is currently a Snr Lecturer in Criminology at the Australian National University.

Acknowledgements

The authors would like to thank Dr Lorana Bartels, Josh Sweeney, Willow Bryant and the Tasmanian Safe at Home team for their assistance and support at various stages of this project. The authors would also like to acknowledge the assistance provided by Ben Young, Senior Research Officer Tasmania Police, who provided the AIC with the extract of data necessary to undertake this project.