Data and statistics

Recommendations for improving data practices

In the following you may find recommendations in relation to data harmonisation and consolidation for improved data practices based on the results of the IMPRODOVA project.

1) Efforts to harmonise data should be underpinned by a clear understanding of the aims, meaning and feasibility of ‘data harmonisation’ and ‘consolidation’ in relation to different data sources.

Key to this understanding is clarity about the purpose of data harmonisation and how data will be gathered and used. Data harmonisation may occur within countries (across agencies such as police, prosecutors, health, and housing), or across countries for the purposes of (a) comparability or (b) conceptual alignment (based on research evidence and knowledge about domestic violence and forming the basis of minimum standards/indicators). Surveys are best placed to elicit insights directly from victims and facilitate comparability across countries, while administrative data gathering benefits from conceptual alignment across agencies and countries on key indicators such as age, sex and relationship between victim and perpetrator.

It should be noted that a harmonised EU definition of domestic violence is likely to be reduced to High Impact Domestic Violence (HIDV) which privileges physical violence. This could effectively obscure all other forms of domestic violence and, in turn, have adverse implications for national interventions

2) Measuring the extent of domestic violence reported to the police in terms of the numbers of victims, perpetrators and offences, as recommended by EIGE (2019), should be a minimum standard for police data gathering.

This data works to raise awareness of the scale of the problem, monitor change over time, and inform the allocation of adequate resources to tackle the problem. As it currently stands, police data on the number of offences is more readily available that number of victims and perpetrators.

3) Data should be recorded on police action taken in response to acts (incidents) reported to them as domestic violence, including those incidents not later recorded as a crime of offence.

This measure provides important information about incidents coming to the attention of the police, and how the police respond to incidents reported to them.

4) Data on types of abuse (e.g. physical, sexual, psychological and economic) should be priority categories for survey data collections.

The collation of this data within surveys should be prioritised and recognised as complimentary to administrative data due to the limitations of administrative data in relation to these variables. Indicators relating to types of abuse (and their seriousness) are populated using crime codes as a proxy, yet there are notable limitations of this approach since some types of abuse (e.g. economic and psychological) and not well recognised or defined in criminal codes.

5) As identified by the Istanbul Convention, measuring the sex of the victim and perpetrator and the relationship between them, should be a minimum standard for police and survey data gathering. In addition to collecting data on sex, measuring the gender identity of victims and perpetrators would be a further step towards inclusivity.

This data is crucial to understanding the gendered dynamics of domestic violence and, in particular, intimate partner violence. Data should be able to be disaggregated in order to be of optimum use for FLRs.

6) Where domestic violence data is gathered on violence/abuse perpetrated in a range of family relationships or a domestic unit (as per the definition of domestic violence adopted by the Istanbul Convention), there should be a clear delineation of these relationship categories and this must include categories for violence/abuse perpetrated by intimate partners and/or ex-partners.

Family and other close relationships form a context of violence in which power relations and other factors relevant to the dynamics of violence contribute to the eminently damaging nature of it in these relationships while simultaneously making it particularly difficult for FLRs to detect and intervene. Such can be, for instance, in the case of parents’ violence against their children, violence perpetrated by adult children against their elderly parents, or violence perpetrated by affinal kin. However, violence perpetrated by partners or ex- partners has a distinctive dynamic and should be clearly delineated in the gathering and reporting of data as recommended by Walby (2005).

7) Data should be gathered regarding repeat offending and victimisation, and the impact of domestic violence and abuse on victims a minimum standard for survey data gathering.

This data is central to understanding the gendered, ongoing and coercive nature of domestic violence. These dimensions are important to operational police responses though they are difficult to record consistently within police administrative data, hence the importance of capturing this information directly from victims within surveys.

8) Consideration should be given to how cases reported to the police can be tracked through the criminal justice system (e.g. through the use of unique identifier for individual cases).

The capacity to track cases throughout the criminal justice system will provide the basis of an in-depth understanding of individual cases as they progress through the system. While this recommendation extends beyond police data gathering, this process begins with the police. With victim privacy and data sharing concerns in mind, the use of a unique identifier should be strictly in relation to case tracking and it should not be shared with agencies out with the criminal justice system (e.g. health and housing). Unique identifiers pose a threat to privacy and the rights of the accused and so, whilst they have undoubted advantages, any implementation needs to consider data infringement risks very carefully.

9) NGOs, social work services and medical services are important sources of data and can provide information about the incidence and impact of domestic violence across different populations. Consideration should be given to the use of the definition of domestic violence adopted by the Istanbul Convention in data recording, and the utilisation of de-identified and aggregated health or social work services data to identify and respond to domestic both individual and community levels.

The health and social care needs of domestic violence victims can inform interventions that can improve a victim’s quality of life and prevent future abuse; however there are of course significant issues of confidentiality which must be respected in relation to health and social care data.

10) The needs and demands placed upon FLRs should be a key consideration development, implementation and operation of administrative data collections systems.

Adequate support, resources and GDPR knowledge should be provided for FLRs as they progress their case work and data recording responsibilities. To minimise the data gathering burden placed upon FLRs such as the police, it should be recognised that surveys are best placed to elicit data from victims on issues such as impact, nature and extent of abuse.

11) The unidirectional flow of data from FLRs to data gathering systems should be addressed by ensuring that FLRs are data recipients rather than just providers.

‘Closing the loop’ for FLRs will allow FLRs to locate and understand their actions in relation to managing and mitigating domestic violence. This too is something for consideration within the training tasks of IMPRODOVA as the consortium works to consider the learning needs of FLRs in the related tasks of WP3.

12) Administrative and survey data analyses should not only be made available to the public (and FLRs), it should be made accessible to them.

National (anonymised) domestic violence data should be publicly available without request. Accessibility should also be considered in relation to the format and presentation of statistical information.

13) Raw data should be made available for further analyses.

Making (anonymised) raw data available to relevant agencies and researchers facilitates analytical insight beyond the headline analyses that are published as standard, and enhances the utility of the data gathered.

14) The EU and Member States should promote and fund surveys that can be repeated every few years to measure developments over time.

This recommendation concurs with the FRA (2014) recommendation on this issue and its adoption would signify a concerted effort to uncover information on the extent and nature of domestic violence.

15) Alternative methods of gathering and utilising data about the ‘bigger picture’ of domestic violence should be considered, in addition to the use of administrative and conventional survey data.

In addressing this recommendation, consideration should be given to the important data gathering undertaken by NGOs. Inspiration here might be taken from the innovative methodology illustrated in the Day to Count first used in the UK in September 2000 (Stanko, 2001). This 24-hour snapshot of domestic violence audited every police service, women’s refuge and the national helpline in the UK, as well as some local authority services in Scotland, to request a simple tally of the total number of who had asked for help, assistance, support or advice about domestic violence. A similar approach has since been used by Women’s Aid England in the Day to Count and Week to Count statistics, and by Scottish Women’s Aid (SWA) in their 24-hour census of the number of women, children and young people who were supported that day. Importantly, the SWA census also documents the number of people that they were unable to provide safe accommodation for due to a lack of resources. Such counts offer a way to communicate to the public at large, not simply the scale of the problem, but the demands upon FLRs.