File 20170723 21223 1e3vuj0.jpg?ixlib=rb 1.1

Domestic violence and Australian churches: why the current data have limitations

None of the studies on domestic violence rates among church-goers are Australian. AAP/Tracey Nearmy

Domestic violence and Australian churches: why the current data have limitations

An ABC report that Australian churches are not only “failing to sufficiently address domestic violence, it is both enabling and concealing it” has generated an outpouring of responses.

Many church and Christian leaders, as well as abuse survivors and their advocates, have praised the highlighting that Christians are not immune from domestic violence, churches are often ill-equipped to respond, and they have done so in ways that perpetuate – rather than relieve – harm.

But others have condemned the report as “selective”, “inaccurate”, and part of a broader “war on Christianity”.

So what do the data and evidence actually say?

What research was used?

Much criticism of the ABC’s report has centred on a finding using US data that:

… evangelical men who sporadically attend church are more likely than men of any other religious group (and more likely than secular men) to assault their wives.

Critics claim the ABC report gave insufficient attention to another finding that:

Regular church attenders are less likely to commit acts of intimate partner violence.

These findings are drawn from a 2007 paper that cited five supporting empirical studies. The most recent of these studies, from 2004, uses data collected between 1992 and 1994.

The researcher, University of Virginia sociologist W. Bradford Wilcox, reported that men who infrequently attended conservative protestant churches were more likely to perpetrate violence, compared to men unaffiliated with religion and those who attended frequently.

However, the analysis shows no statistically significant difference in domestic violence perpetration by men frequently attending conservative protestant churches compared to the unaffiliated population.

The five studies have some key limitations.

  • The studies are all more than 20 years old. Data this old regarding religion and gender relations clearly has limited applicability today.

  • The studies collect data only on physical violence. They do not capture other common forms of domestic violence – including sexual, emotional, psychological, social, financial and spiritual abuse. As a result the studies are likely to underestimate domestic violence substantially.

  • Two of the studies did not include women’s reporting of experiences and relied only on men’s reports of their own violence perpetration. This is also well-recognised as leading to under-reporting.

  • All but one of the studies are cross-sectional, and only capture experiences in the past year – rather than lifetime exposure.

  • None of the studies is Australian.

Inference about any protective effects of regular conservative protestant church attendance on domestic violence perpetration in contemporary Australia is therefore highly problematic. It also runs the high risk of shifting blame and drawing attention away from listening to the experiences of those who have spoken out.

How research in Australia could be done

Comprehensive, independent Australian data regarding domestic violence within churches are long overdue. Analysis of existing administrative and ongoing longitudinal studies is a less expensive and quicker option than commissioning new data collection.

For example, the Household, Income, Labour Dynamics in Australia study collects data on religious affiliation, gender attitudes, and violence – though not specifically domestic violence.

Other longitudinal studies such as Ten to Men, the Australian Longitudinal Study of Women’s Health, and the Longitudinal Study of Australian Children also provide opportunity for analysis.

Using longitudinal data is particularly important. It allows for more detailed analysis of patterns and predictors of exposure over time, and supports exploration of causal pathways and mediating and moderating factors not possible in cross-sectional data. The downside is they are subject to sample attrition, and therefore may become unrepresentative through time.

National population data sources on domestic violence such as the ABS’ Personal Safety Survey and Crime Victimisation Data do not currently collect data on religious attendance or affiliation. But they could easily do so.

The National Community Attitudes Survey on Violence Against Women is another opportunity that, with a few additional demographic questions and a targeted boosted sample, could provide essential data on these issues within faith communities – including those who have left.

The 2016 National Church Life Survey includes questions on domestic violence. However, it is collected via surveys completed during church services, when women are likely to be sitting beside their partners – and so highly open to under-reporting.

This is not to say that new data collection is not warranted. Independent, academically robust data to examine questions not answerable using existing data is also essential.

Ideally, this would include multi-level analysis of institutional factors, such as policies and practices, as well as individual level factors, related to experiences of – and responses to – domestic violence. It must include both those within, and those who have left, churches.

Critically, any research must also identify areas for action, and rigorously evaluate the effectiveness of policies and responses.


The National Sexual Assault, Family & Domestic Violence Counselling Line – 1800 RESPECT (1800 737 732) – is available 24 hours a day, seven days a week for any Australian who has experienced, or is at risk of, family and domestic violence and/or sexual assault.

The Conversation is a non-profit + your donation is tax deductible. Help knowledge-based, ethical journalism today.