Police Information

Arresting Evidence: Domestic Violence

Wed, September 19, 2001

Source: Jalna Hanmer, Sue Griffiths and Dave Jerwood

Arresting Evidence: Domestic Violence And Repeat Victimisation Policing & Reducing Crime Briefing Note

Summary

This project is part of a continuing programme of research and development on policing responses to repeat victimisation. The aim was to reduce repeated domestic violence by adopting a graded response to incidents in the Killingbeck Division of West Yorkshire Police. An extended briefing note (1/98) was published in April 1998 describing the project and setting out the findings from the first six months. This paper presents findings from the first full year of the project, and discusses the operational issues involved in implementing and monitoring such initiatives. 

The model

The project builds on the force domestic violence policy - to arrest whenever possible, to gather evidence as with any other crime, and to ensure the safety of the victim. It involves grading responses to domestic violence incidents according to the number of times the police have attended the offender in the past year. At its simplest, a first police attendance triggers a Level 1 response, a second a Level 2 and a third a Level 3. However, in assessing the initial level of response, the full police record is reviewed along with information from the woman on previous unreported assaults, and entry at higher levels may be required. The main premise is that an increase in incidents prompts an increase in activity by the police and other agencies. This activity is directed at both victims and offenders, and takes account of whether the offence is common law or criminal and whether the woman is living with or separated from the perpetrator.  

Main findings

The project reduced repeat attendances

To assess the model’s effectiveness, the level of domestic violence attendances in 1997 was compared with a three-month period in 1996 before the project began. The aim was to reduce the number of attendances at Levels 2 and above. Table 1 shows that the numbers of men entering the project at Levels 2 and 3 are significantly reduced since the project began. During April to June 1996 (pre-project period) Level 1 entries were less frequent and Levels 2 and 3 more frequent than during the project. The proportion of attendances that were one-off, for example, increased from 66% to 85%. The change at each level is progressive over each quarter of 1997, suggesting that implementation of the model improved over the year.

Over the year more women received police assistance

Table 1 also shows that the overall number of incidents remained constant but with more Level 1 entries, suggesting that more women are reporting incidents to the police. Over time, if repeat attendance continues to be effectively prevented, we would expect to see overall numbers of incidents reduce. Even with stable incidence numbers, however, there is a saving in police resources as Level 1 interventions require less time to be spent on responding to domestic violence calls than Levels 2 and 3.

Early intervention reduces repeat attendances

The main report shows that early intervention achieves the greatest reduction in repeat attendances. Level 1 entrants have the lowest proportion of re-attendances, with three-quarters (75%) not requiring a second attendance in 1997, compared to 54% of Level 2 entrants and 36% of Level 3. The percentage of Level 1 entrants requiring two or three or more attendances was consistently less than those at the other levels. 

This indicates that introducing the first intervention at Level 2 or 3 is not as effective as first intervention at Level 1. The most likely reasons for this are either that the offender has become desensitised or has concluded that the police do not object to his behaviour or both. There is also evidence that recording individual repeat patterns in this way allows the identification of the small number of chronic repeat offenders who the project has not been able to demotivate. Police and other agencies can then begin to develop tailored responses, targeting specialist services where they are most needed.try level of men by quarterly time periods

The time interval between calls for assistance increased

The project demonstrates greater time intervals between repeat attendance than that shown by earlier research, where 35% of households required a second attendance within five weeks. In Killingbeck, within five weeks, 9% who entered at Level 1 were attended again; of those requiring Level 2 and 3 interventions, 15% and 26% were attended again within five weeks.  

The recording of domestic violence incidents improved 

West Yorkshire Police has an independent database for recording domestic violence incidents – the Domestic Violence Index (DVI). When the project began, a daily manual check of all calls received at Killingbeck was undertaken and all attendances that appeared to be ‘domestic’, irrespective of their final codes, were cross- checked with the DVI. This found that only 50% of attendances were allocated a domestic FIN-code when attendances were finalised. With intensive follow-up and improved awareness amongst attending officers and control room staff, by the end of the year the recording rate had increased to 80%. 

Points for action

The Killingbeck experience demonstrates the benefits to be gained from adopting a strategic approach to domestic violence at the local level. The new Crime and Disorder partnerships offer an infrastructure for delivering an effective response to victims and offenders. In developing their local strategies, consideration should be given to the following issues: 

Recording practices - systems need to be put in place to allow accurate identification of repeats, both to enable rapid appropriate interventions and for monitoring purposes. For the police, this may involve developing common definitions of ‘domestic’ and consistent use of final codes. Asking the victim about previous unreported assaults can inform the development of an effective response. To gain a picture of the extent of domestic violence in a local area, police records can be supplemented by data from other sources, for example women’s refuges, GPs, Accident and Emergency departments and local surveys of women. 

Graded response - a graded response is simply a way of organising existing resources in a systematic, planned way, matching the level of response to the level of risk. The precise measures will vary according to what is available in the local area. While the levels of intervention are set, there should be room for discretion. If a first attack is particularly serious, for example, you may want to implement a measure from a higher level such as a panic alarm.  

Training - training should be a key part of the strategy. In Killingbeck, to prepare for the project, all officers received face-to-face training that aimed to increase their understanding of domestic violence and its potential consequences and to convey the project rationale, its interventions and their role within it. Initial training should be supplemented with ongoing updates and feedback sessions. Cross-agency training can help to improve awareness of each agency’s roles and responsibilities. 

Inter-agency working - a wide range of agencies and organisations need to be involved in an effective response to domestic violence, and early consultation before any initiative can help to improve ‘ownership’. Attention should be given to improved communication flows and information sharing within and between agencies. 

Performance indicators - it is widely recognized that domestic violence is under-reported. Encouraging increased reporting should be a key aim of any initiative and forces should expect the overall number of recorded incidents to increase in the short term. This should be reflected in the development of local performance indicators to ensure that they do not act as a disincentive. Monitoring the patterns of attendances over time is a useful indicator of effectiveness. In the short term, the aim is to reduce repeat attendance and increase one-off attendances, resulting in stable or increasing calls overall. At some point, saturation may be reached, as there is not a limitless supply of men who require police attendance (and /or women who require police assistance). When every woman who requires attendance receives it and if repeat attendance continues to be effectively prevented, we would expect to see overall numbers of incidents decrease.

To view the full paper click the URL below