Domestic and Family Violence: Responses and Approaches across the Australian Churches

Abstract

Domestic and family violence (DFV) is a serious and widespread problem in Australia and across the world, including in faith communities. There are calls for research to assist churches to better recognize, respond to and prevent violence. This study draws on data from the 2016 Australian National Church Life Survey (n = 883 senior local church leaders, n = 1270 churchgoers) to provide the first Australia-wide cross-denominational statistics on Christian clergy responses to DFV. Two-thirds of leaders had previously dealt with DFV situations in their ministry, primarily responding to victims of abuse by referring them to specialist support services and by counselling them. The findings suggest a particular depth of experience with DFV situations and strength of awareness of the needs of victims for safety and specialist support among Salvationist leaders. While, overall, a substantial majority of churchgoers felt that they could approach their church for help if they were experiencing DFV, just half of Catholics felt that they could do so. Future research should explore responses to DFV in specific denominations and culturally and linguistic diverse contexts in more detail and seek to understand the practices used by the large minority of clergy who are dealing with perpetrators.

1. Introduction

Domestic and family violence (DFV) is a serious and widespread problem both internationally (WHO 2010) and in Australia (AIHW 2018), with enormous individual and community impacts and social costs, predominantly affecting women and children. DFV constitutes violence in the context of intimate partner relationships (domestic violence, domestic abuse or intimate partner violence) as well as between family members more broadly (family violence). DFV includes physical, sexual, psychological and emotional violence and abuse, as well as the sustained patterns of intimidation, isolation and control known as coercive control (AIHW 2018). According to the 2016 Australian Personal Safety Survey, 23% of women and 8% of men in Australia have experienced violence from an intimate partner since the age of 15 (ABS 2017) and women are much more likely than men to be hospitalized and killed as a result of violence from a current or former partner (AIHW 2018). Domestic violence is the greatest health risk factor for women aged 25 to 44 and is the single largest driver of homelessness for women, a common factor in child protection notifications, and results in a police callout on average once every two minutes across the country (Our Watch n.d.). There are limited data on the nature, extent and impacts of family violence in relation to children, but the effects of DFV on their health and development can be severe (AIHW 2018).

Violence can occur in all contexts and communities. Efforts to raise awareness of the reality of DFV in Christian churches in Australia date back decades (e.g., Last and Gilmore 1994) and recent testimonies by survivors of abuse show that the violence continues (Baird and Gleeson 2017a, 2017b, 2018; Lim 2015; Pepper et al. 2021; STDFVQ 2015, p. 104). Initiatives are underway to develop policies, training and resources for churches to better recognize, respond to and prevent DFV, some longstanding and some more recent. These include efforts by denominations (e.g., ABM 2018; Anglican Church of Australia n.d.; Anglican Diocese of Melbourne n.d.; PSU n.d.), ecumenical and parachurch initiatives (e.g., Common Grace 2018; Queensland Churches Together n.d.; NCCA n.d.), and interfaith projects (Vaughan and Sullivan 2019). While there is a growing body of international research on DFV in Christian contexts (e.g., Nason-Clark et al. 2017), there has been less investigation undertaken in Australia (Priest 2018; Truong et al. 2020). Calls for current research to support efforts to better prevent and respond to DFV have come from DFV workers and experts, the academic community, government, and churches themselves (e.g., Bartels 2010; Hamence 2018; Moore and Dickson 2015; Priest 2018; Vaughan et al. 2020).

In this paper, we report on actions that church leaders have taken when responding to DFV situations and on views about the approachability of Australian churches for those experiencing DFV. This is the first Australia-wide, cross-denominational survey study to do so. First, we outline how DFV is present in Christian contexts and roles that churches play in the dynamics of and in response to abuse, with a particular focus on the contributions of clergy. Next, we present research into the responses of church leaders to DFV situations and the approachability of churches, drawing on data from the 2016 National Church Life Survey across multiple denominations. We conclude with a discussion of our findings in conversation with the literature on DFV in church contexts, a consideration of the limitations of this study, and recommendations for future research.

1.1. Domestic and Family Violence and the Australian Churches

Factors underlying and contributing to DFV lie in a range of environments, with “faith-based institutions” cited as one such environment (VicHealth 2007). While it can be difficult to accurately record the extent of DFV in a population (ABS 2018; AIHW 2018), internationally, the evidence indicates that in English-speaking countries, intimate partner violence (IPV) occurs in Christian contexts at rates similar to those seen in general populations (Lock 2018; Knickmeyer et al. 2010; Nason-Clark 2009; Levitt and Ware 2006). The first study of the prevalence of IPV in Australian Christian churches was conducted in late 2019 for the Anglican Church and suggests that the prevalence is similar or greater among Anglicans than in the Australian community at large (Pepper and Powell 2021).

At the same time, attitudes of communities and bystanders are significant both in the support of those experiencing DFV and ongoing preventive frameworks (VicHealth 2007), with faith communities having an unrealized potential to further violence prevention and response (Vaughan et al. 2020). As part of supporting community-driven initiatives to prevent and respond to violence against women and improve community awareness of violence, the Third Action Plan of Australia’s National Plan to Reduce Violence Against Women and their Children included a priority to “build the capacity of community and faith leaders to reject, prevent and respond to violence” (DSS 2016, action 3.9).

Church contexts can both contribute to and, conversely, help to prevent violence (Horne and Levitt 2003; Pepper et al. 2021; Truong et al. 2020). Church communities are embedded in the broader Australian context and churchgoers come from diverse communities and socio-economic backgrounds (Powell et al. 2017). The same social and cultural risk factors which impact on the severity and frequency of violence operate in faith communities as elsewhere. While causes of violence are complex and require greater understanding, a common argument is that gender inequality is a key determinant of violence against women. In this framework, gender inequality refers to the social condition where men and women have unequal value and access to power and resources. The structure, norms and practices of gender inequality underpin and produce specific drivers of violence against women, such as condoning violence, men’s control of decision making, stereotyped constructions of masculinity and femininity, and disrespect towards women and male peer relations that emphasise aggression. In the context of other social inequalities, these gendered drivers support the tolerance of violence against women (e.g., Our Watch et al. 2015; Vaughan et al. 2020). Against this framework which links gender inequality to dynamics of abuse, the gendered leadership hierarchies and norms within some faith traditions has come under scrutiny (e.g., Vaughan et al. 2020).

Research on the experiences of domestic violence survivors shows that, at their best, churches can provide trusted relationships and various forms of support to help victims and survivors to meet their needs within and beyond life in an abusive relationship (Pepper et al. 2021). However, there are particular factors in churches that can contribute to the dynamics of abuse. These include cultural factors such as norms about marriage, forgiveness and gender roles, and complexities to do with the presence of both partners in a church community (Drumm et al. 2018; Nason-Clark et al. 2017; Pepper et al. 2021; Wendt 2008; Westenberg 2017). For example, Christian victims of domestic abuse often feel that they are called by God to endure their suffering, to repeatedly forgive their abuser, and to fulfill their marriage vows until death, while abusers may use religious justifications for their violence and manipulate religious leaders (Nason-Clark et al. 2017, p. 1). Research on violence in culturally and linguistically diverse communities also indicates the complex role that religion plays in the continuation of violence and in barriers to seeking help (Truong et al. 2020; Vaughan et al. 2020). For example, fear of stigmatization and of law enforcement authorities may lead faith communities to manage problems with violence internally (Vaughan et al. 2020) and uncertain residency status increases the risk of DFV and is a barrier to seeking support (Vaughan et al. 2016).

The experiences of those who have been abused show that clergy and other church leaders can play a particularly important role in responding to abuse in church contexts or conversely in failing to do so (Pepper et al. 2021). Research on the ways that clergy are responding to domestic violence exists internationally, in North America especially, including for specific ethnic groups where DFV levels are elevated and for particular Christian traditions (e.g., Ames et al. 2011; Bent-Goodley et al. 2015; Choi 2015a, 2015b; Davis and Johnson 2021; Nason-Clark et al. 2017; Rotunda et al. 2004; Williams and Jenkins 2019). Little such research is available in an Australian context, with faith communities lacking capacity, resources and funding to document and evaluate the measures they are taking to address DFV (Vaughan et al. 2020). An exception is a single-denomination study of clergy responses undertaken for the Anglican Church of Australia (Powell and Pepper 2021). Other Australian research has involved a small number of clergy from a range of religions and culturally and linguistically diverse communities (Truong et al. 2020), and a Victorian multifaith participatory action research project undertaken in the period 2018–2020 known as “Faith Communities Supporting Healthy Family Relationships”, which aimed to build the capacity of faith leaders to prevent and respond to DFV and violence against women and in which Anglican and Uniting churches participated (Multicultural Centre for Women’s Health n.d.). In studies conducted internationally and in Australia, a minority of clergy have reported feeling well equipped to respond to abuse (Nason-Clark et al. 2017, p. 31; Powell and Pepper 2021; Skiff et al. 2008; Zust et al. 2018).

Based on more than 25 years of research into domestic violence in Christian families in North America, including a large number of studies involving religious leaders, congregations, victims and perpetrators, and workers who assist families impacted by abuse, Nason-Clark et al. (2017) assert that there are a number of ways that pastors can effectively respond to the needs of a person who has experienced violence from their partner. These include understanding what domestic violence is and how it can occur within a faith community, knowledge of available resources in the community and how to make referrals, providing spiritual counsel for victims, and supporting victims and their families and others affected by the abuse in the long-term process of healing (pp. 58–59). Pastoral support for abusers on the other hand is particularly complex, challenging and risky. Such support should follow a clear plan, including referrals to trained professionals, measures to ensure the safety of the victim, and accountability by the abuser (Nason-Clark et al. 2017, pp. 94–96).

Faith-based contexts influence social norms and beliefs, and religious leaders can play a key role in building respectful relationships (Our Watch et al. 2015). As well as responding to specific domestic violence situations when and as they occur, steps can be taken with a view to preventing abuse in Christian contexts. In preaching, biblical interpretation, liturgy marriage programs and more, religious language that asserts egalitarianism in relationships and links forgiveness to justice and dignity are important to correct harmful discourses that signify toleration of abuse (Westenberg 2017). Moreover, awareness raising within the church community about the nature of abuse, the reality of its occurrence in Christian families, the unacceptability of abuse, and the supports that are available to those who are experiencing abuse is important for creating a culture of abuse prevention (Pepper et al. 2021). The Faith Communities Supporting Healthy Family Relationships project developed a set of principles for what works for faith settings to prevent and respond to DFV and violence against women including prioritize the safety of women and children; strengthen relationships between secular organizations with DFV expertise and faith leaders; co-design and co-deliver initiatives between faith leaders, communities and sectoral experts; understand the central role of gender inequality and how this intersects with other inequalities; and engage senior leadership early and in a sustained way (Vaughan and Sullivan 2019).

However, a review of the evidence available internationally about prevention and responses to DFV and violence against women in faith settings indicates that there is limited evidence on what works to build capacity in faith communities to prevent violence and to bring about long-term change. There is also little evidence on effective prevention and response when violence is perpetrated by people other than intimate partners. Further, more research is needed about what approaches are effective in holding religious men to account and in stopping their violent behaviour from reoccurring (Vaughan et al. 2020).

The Victorian Royal Commission into Family Violence recommended that “faith leaders and communities establish processes for examining the ways in which they currently respond to family violence in their communities and whether any of their practices operate as deterrents to the prevention or reporting of, or recovery from, family violence or are used by perpetrators to excuse or condone abusive behaviour” (State of Victoria 2016, p. 89).

In this paper, we provide baseline, quantitative data on the ways in which clergy and church leaders across multiple Christian denominations have responded to DFV situations with a view to informing the churches and those who work with them to improve their practices.

1.2. Research Questions and Hypotheses

We seek to answer three broad research questions.

First, what proportion of clergy have dealt with DFV situations, and what responses have they used?

Second, because forming bridges between churches and secular support services is an important way to equip clergy to better respond to DFV (e.g., Dyer 2016; Nason-Clark 2009; Vaughan et al. 2020), how familiar are clergy with support services in their local communities?

Third, given the increased focus on faith communities with respect to their potential to contribute to DFV prevention and response (e.g., DSS 2016; Vaughan et al. 2020), to what extent do churchgoers feel they could go to a church for help if they or someone they know were experiencing DFV?

We also explore demographic and denominational variations in our findings. While this is largely an exploratory study, we make the following hypotheses about demographic differences. First, older clergy will be more likely to have dealt with DFV situations than younger clergy, because older clergy will typically have had more years in ministry contexts with a longer-term opportunity for contact with DFV situations. Second, female clergy will be more likely to prioritize victim safety in their responses. Given women are much more likely to be victims of domestic violence than men (AIHW 2019; ABS 2017), they may choose to disclose more often or in more detail to female clergy, who may, in turn, be more attentive to another woman’s presenting needs.

2. Materials and Methods

2.1. Data Collection and Samples

Data from the 2016 Australian National Church Life Survey (NCLS) were used to answer the research questions. The NCLS is a five-yearly quantitative survey of thousands of Christian churches, hundreds of thousands of churchgoers and thousands of church leaders in approximately 20 Australian denominations (Catholic, Anglican and other Protestant). The 2016 survey wave received ethics clearance from Australian Catholic University (Ethics Register Number 2016-186E).

The NCLS is based in local churches (congregations and parishes), with approaches to recruitment and sampling varying across the denominations (attempted census, random sampling, or opt-in; paid for by the local church or by the denomination) (Pepper et al. 2018). Participating local churches complete several types of survey forms, including the “NCLS Attender Survey” and the “NCLS Leader Survey”. All churchgoers aged 15 years and older are requested to fill out an NCLS Attender Survey form, usually during or after a service of worship. NCLS Leader Survey forms are available to anyone involved in leadership (including clergy and pastors, elders, church councillors and other leaders), with the most senior leader at the local church, at minimum, encouraged to participate. Both hard copy and online survey forms are available.

Catholic NCLS data are random samples; however, in Protestant denominations, there are self-selection biases in church participation related to church size, locality and theological tradition, with larger urban churches of an evangelical flavor over-represented in the datasets (Pepper et al. 2018). Nevertheless, the datasets have national coverage and denominational diversity, and churches from a wide diversity of traditions participate. To correct for differences in participation, results are typically weighted to adjust for variations in survey participation levels between denominations and between churches of different sizes within the Protestant denominations (Pepper et al. 2018). In the present case, 14 denominations and movements were sampled well enough for inclusion in the analysis, accounting for some 95% of the weekly churchgoers in Australia (not including Orthodox, independent or house churches) and 89% of the estimated number of senior local church leaders, according to NCLS estimates (Powell et al. 2021). For a more detailed explanation of the NCLS methodology, participation rates, and the strengths and limitations of the NCLS datasets, see Pepper et al. (2018).

2.1.1. Leader Survey Participants

The 2016 NCLS Leader Survey (total n approximately 8200) comprised two 12-page variants. The questions used in the present paper were included in one of these variants and in a section of the survey which was completed by pastoral and ministry staff only. The final number of respondents for analysis was n = 883 “senior local leaders” (ministers, pastors or priests who are the primary or senior leader of a local church or who have equal standing in the ministry team, as well as laypeople serving as the principal leader in a local church). Almost all of these leaders were clergy—just 3.5% indicated that they were a layperson serving as principal leader. We restrict our analysis to senior leaders only rather than clergy more generally because of the NCLS sampling approach, which particularly targets senior leaders. Compared with churchgoers more broadly, the leaders were predominantly male (78%), Protestant (reflecting the relatively large size of Catholic parishes) and highly educated. Some 29% of leaders had completed a postgraduate theological qualification, while for 35% a bachelor degree was their highest level of formal theological qualification. Some 21% were born outside of Australia, including 11% from a country where English is not the main language spoken (primarily a country in Asia).

2.1.2. Attender Sample Survey Participants

The 2016 NCLS Attender Survey (total n of approximately 260,000 people from 3000 congregations and 20 denominations) consisted of a four-page main survey of demographics, Christian faith and practice and church health, which was completed by most individual participants and a series of smaller four-page surveys, each of which was a random sample of the total participants. The 2016 NCLS Small Sample Attender Survey L (“2016 Attender L”) covered the majority of the questions from the Main Attender Survey, together with a suite of questions on personal wellbeing, social inclusion and a range of social issues in relation to the church. The final sample size was n = 1270, some 59% of whom were female and 36% were born outside of Australia, including 23% from a country where English is not the main language spoken (primarily a country in Asia).

The demographics of the leader and churchgoer samples are shown in

Table 1. Demographics of NCLS samples.

Variable	Percentage
	NCLS Leader	NCLS Attender
Age
15–29 years	3	16
30–49 years	37	24
50–69 years	52	33
70+ years	8	27
Gender
Female	22	59
Male	78	41
Country of birth
Australia	79	64
Other English-speaking	10	12
Non-English-speaking	11	23
Educational attainment
School	8	37
Trade certificate/diploma	19	25
University degree	73	38
Staff position
The minister/pastor/priest of the congregation/parish	59	-
The senior minister/pastor/priest of ministry team	27	-
Minister/pastor/priest of equal standing in ministry team	9	-
Interim minister/pastor/priest	2	-
Layperson serving as principal leader	3	-
Denomination
Anglican	18	12
Baptist/Churches of Christ	18	11
Catholic	13	46
Pentecostal	20	16
Salvation Army	6	-
Uniting Church	16	8
Other Protestant	9	8 1

Source: Leaders—Powell et al. (2016c). Attenders—Powell et al. (2016a). ¹ Includes Salvation Army. Data are weighted.

.

2.2. Measures

The following measures from the 2016 NCLS Leader Survey were used.

Responses when dealing with DFV situations: The question was modified from a survey of Protestant pastors in the USA in 2014 (Sojourners, and IMA World Health 2014). “Of the following responses, which have you used when dealing with domestic and family violence situations? (Mark ALL that apply)”. Response options were “Provided counselling to the victim”; “Provided counselling to the perpetrator”; “Provided marriage or couples counselling”; “Conducted a safety risk assessment with the victim”; “Referred the victim to a service agency”; “Referred the perpetrator to a service agency” “Other response” and “I have not dealt with domestic violence situations”.

Familiarity with support services: The question was “How familiar are you with support services available in your local community to assist victims and perpetrators of domestic and family violence (e.g., crisis centre, shelter, hotline)?” Response options were “Very familiar”; “Somewhat familiar”; “Not very familiar” and “Not at all familiar” (order reversed in analysis).

The 2016 NCLS Attender Sample Survey included the following question.

View of church as a source of help: “If you or someone you knew were experiencing domestic or family violence, do you feel that you could go to someone from the church for help?” Response options were “Yes”, “No” and “Not sure”.

Denomination was the affiliation of the church where the respondent completed the survey. There were sufficient unweighted cases to allow Anglican, Baptist/Churches of Christ, Catholic, Uniting Church, Salvation Army (Leader Survey only) and Pentecostal (Australian Christian Churches, C3 Church, International Network of Churches, Christian Revival Crusade) respondents to be separated in the analysis, leaving an Other Protestant category comprising the remaining denominations (Lutheran, Presbyterian, Christian Reformed, Fellowship of Independent Evangelical Churches, as well as Salvation Army in the Attender Survey).

Demographic variables were age, gender, highest level of formal education (school/trade certificate/diploma/associate diploma, university degree, postgraduate qualification) and country of birth (Australia; other country where English is the main language spoken, comprising the UK, Ireland, North America, New Zealand and South Africa; and other non-English-speaking county).

2.3. Analysis

Patterns in leaders’ responses to DFV situations were explored by looking at the number and types of responses and correlations between different types. Cluster analysis (TwoStep Cluster procedure) was also used to determine whether leaders could be grouped according to the way that they responded. The basis for analyses of the different types of responses is leaders who had dealt with DFV situations rather than all leaders.

Bivariate analyses were undertaken to determine whether responses differed by denomination and demographics, using chi-square tests to determine statistical significance (p < 0.05). Effect sizes were calculated using Phi. Results for Catholic and Pentecostal leaders should be treated with caution due to the low number of unweighted cases upon which they are based.

Because there were significant differences in the leaders’ DFV responses by both demographics and denomination, multivariate analyses were then conducted to see which effects remained when the other variables were controlled. Binary logistic regressions were performed with leader responses to domestic and family violence as the dependent variable, and with demographics (age, gender, education and country of birth) and denomination as the independent variables. The proportions reporting DFV responses were relatively low among Uniting Church leaders, so the Uniting Church was treated as the reference category.

Demographic and denominational breakdowns on leaders’ assessment of their familiarity with local DFV support services were conducted, with eta-squared used for effect size. Linear regression was used to predict familiarity.

Demographic and denominational breakdowns on churchgoers’ feelings about approaching the church for help if they or someone they knew were experiencing DFV were conducted, with Cramer’s V used for effect size. Results for Pentecostal respondents should be treated with caution due to the low number of unweighted cases upon which they are based. Because there were no statistically significant demographic differences in the responses, multivariate analysis was not undertaken on the attender sample.

Analyses were conducted using IBM SPSS statistics package version 27. All univariate and bivariate analyses are weighted, while multivariate analyses (regressions) are unweighted.

3. Results

3.1. Clergy Responses to DFV

In 2016, two-thirds of senior local church leaders (67%) reported that they had dealt with domestic and family violence situations. Of those who had done so, the most common response was to refer the victim to a service agency (77%), followed by provided counselling to the victim (70%); providing marriage or couples counselling (41%); providing counselling to the perpetrator (34%); referring the perpetrator to a service agency (28%); and conducting a safety risk assessment with the victim (23%). A further 10% had made some other (non-specified) response (

Table 2. Type and number of responses to DFV situations among senior local church leaders.

Type of Responses Used	Percentage of Leaders Using the Number of Types of Responses Shown
	0 Types	At Least 1 Type	1 Type	1–2 Types	At Least 3 Types	Couples Counselling
Counselling victim	0	70	20	46	93	83
Counselling perpetrator	0	34	1	10	57	48
Couples counselling	0	41	15	15	66	100
Safety risk assessment	0	23	2	7	39	30
Referred victim	0	77	56	64	89	71
Referred perpetrator	0	28	0	6	48	38
Other response	0	10	6	8	12	9

Source: Powell et al. (2016c). Data are weighted.

). Almost all (94%) indicated that they had dealt directly with a victim through referral, counselling and/or safety risk assessment. In contrast, under half, 44%, had dealt directly with a perpetrator through counselling or referral.

While one-third of senior local leaders had not responded to DFV situations, some 15% reported using one of the seven types of responses listed (including “other response”), while approximately one-third (32%) reported using up to two of the responses. A further one-third (34%) reported at least three sorts of responses.

As well as being the most commonly reported response, referring the victim to a service agency was the response that was most often reported in isolation; a little over half of those who reported only one type of response had made such a referral (56%, Table 2). In contrast, only one-fifth of those who reported just one type of response had counselled the victim.

Those who reported one or two types of responses had most commonly referred victims to service agencies (64%) and counselled victims (46%). Almost all of those who reported three or more types of responses undertook these actions (93% had counselled victims, 89% referred victims to service agencies).

Responses that were focused on the perpetrator—namely counselling the perpetrator and referring the perpetrator to a service agency—tended not to be reported in isolation, but rather as part of a set of responses. Very small numbers of clergy who reported one or two responses had focused on perpetrators. In contrast, among those who reported three or more types of responses, 57% had counselled perpetrators while 48% had referred perpetrators to a service agency. Almost all leaders who referred perpetrators had also referred victims (97%), and almost all who had counselled perpetrators had also counselled victims (98%).

Referring perpetrators and counselling them were associated with each other. However, this was not the case for victims—referral and counselling of victims were independent of each other (

Table 3. Pearson’s correlation coefficients between responses to DFV (among leaders who had responded to DFV).

	1	2	3	4	5	6
1 Counselling victim
2 Counselling perpetrator	0.435 **
3 Couples counselling	0.238 **	0.254 **
4 Safety risk assessment	0.116 *	0.185 **	0.125 *
5 Referred victim	−0.017	−0.078	−0.109 #	0.131 *
6 Referred perpetrator	0.116 *	0.377 **	0.182 **	0.214 **	0.292 **
7 Other response	−0.111 *	−0.011	−0.026	0.010	−0.072	−0.017

Source: Powell et al. (2016c). # p < 0.05; * p < 0.01; ** p < 0.001. Data are weighted.

). Still, given the prevalence of referring and counselling victims, most who had undertaken one of these actions had also undertaken the other.

Similarly, marriage/couples counselling tended to be used as part of a suite of measures, rather than in isolation. Some two-thirds (66%) of leaders who reported three or more types of response had provided this counselling, compared with 15% of leaders for whom it was their only type of response or one of two sorts of responses (Table 2). Most of those who had counselled couples had also counselled victims (83%) and approximately half had counselled perpetrators (48%, Table 2). However, those who engaged in marriage/couples counselling were a little less likely to refer victims to service agencies than those who did not use such counselling, but more likely to refer perpetrators and to conduct safety risk assessment with a victim (Table 3).

Cluster analysis indicated two groups of leaders, who differed in their types of response. A large number or leaders (approximately 80% of the group) undertook few sorts of interventions (median of 2 types) and which were focused primarily on victims (referral to services and counselling). A smaller number of leaders (approximately 20% of the group) focused on both victims and perpetrators, undertaking a variety of interventions (median of 5 types). It was not possible to predict these two groups from demographics or denomination.

There were small demographic and denominational differences in the responses made (see

Table 4. Senior local church leader responses when dealing with DFV situations by demographics and denomination.

	“Of the Following Responses, Which Have You Used When Dealing with Domestic and Family Violence Situations?”
	% of All Leaders	% of Leaders Who Have Dealt with DFV
	Dealt with DFV	1 1	2	3	4	5	6	7
Overall	67	70	34	41	23	77	28	10
Age
Under 40 years	49	67	24	44	26	69	23	10
40–49 years	67	67	31	27	20	78	22	14
50–59 years	71	73	40	43	26	75	28	9
60+ years	72	71	35	47	22	79	34	9
Effect size (Phi)	0.172 **	0.053	0.102	0.158 *	0.062	0.074	0.103	0.061
Gender
Female	65	64	30	29	31	81	29	7
Male	67	72	35	44	21	75	28	11
Effect size (Phi)	0.017	0.067	0.050	0.122 *	0.095 #	0.055	0.011	0.054
Education
Certificate/diploma/school	72	71	39	53	21	84	30	10
Bachelor degree	61	71	29	37	21	70	24	11
Postgraduate	68	70	36	36	26	77	29	10
Effect size (Phi)	0.089 #	0.012	0.087	0.156 *	0.059	0.133 *	0.052	0.004
Country of birth
Australia	67	71	33	41	25	80	28	10
Other English-speaking	63	61	34	41	22	79	30	9
Non English-speaking	67	71	40	45	10	50	23	14
Effect size (Phi)	0.027	0.065	0.041	0.025	0.111 #	0.210 **	0.036	0.040
Denomination
Anglican	75	69	35	39	18	69	20	6
Baptist/Churches of Christ	58	64	33	35	25	76	22	10
Catholic	63	68	23	39	13	81	42	13
Pentecostal	72	79	46	60	33	79	39	9
Salvation Army	88	72	30	21	28	93	35	15
Uniting	61	61	23	28	19	77	25	16
Other Protestant	57	83	41	53	24	72	11	9
Effect size (Phi)	0.181 **	0.158 #	0.170 #	0.249 **	0.160 #	0.147	0.217 **	0.111

Source: Powell et al. (2016c). ¹ 1 Counselling victim, 2 Counselling perpetrator, 3 Couples counselling, 4 Safety risk assessment, 5 Referred victim, 6 Referred perpetrator, 7 Other response. Chi-square tests were used to evaluate if any differences are statistically significant, # p < 0.05; * p < 0.01; ** p < 0.001. To quantify the magnitude of the effect, the effect size was calculated using Phi.

). Those aged under 40 years were less likely to have dealt with DFV (49%) than those over 40 years (70%), confirming our first hypothesis. Those in their 40s were less likely to have used marriage/couples counselling than younger senior leaders and leaders aged at least 50. Female leaders were less likely to offer marriage/couples counselling (29% of women who had responded to DFV situations, 44% of men) and more likely to do a safety risk assessment with the victim (31% of women, 21% of men). These findings supported our second hypothesis. The responses of leaders also differed by level of formal education, with those with non-degree education most likely to have dealt with DFV situations, used marriage counselling and to have referred victims to a service agency. Leaders from countries where English is not the main language spoken were less likely than other leaders to have conducted a safety risk assessment with the victim and to have referred the victim to a service agency.

Regarding denomination, Salvation Army leaders were the most likely to have dealt with DFV situations (88% of leaders) by some margin. When it came to the specific responses used, Pentecostal leaders reported relatively high usage of most of the responses. However, Salvation Army leaders reported the highest level of referring victims to support services (93%), although the difference across the seven denominational groups for this type of response was just outside of statistical significance. Denominational differences were most pronounced for marriage/couples counselling, where 60% of Pentecostal leaders and 53% of “Other Protestant” leaders who had dealt with DV situations used this, compared with at most four in 10 leaders from other denominations. Referral of perpetrators was reported by 4 in 10 Pentecostals and Catholics and 35% of Salvation Army leaders, compared with up to one-quarter of leaders from other denominations.

However, few statistically significant relationships remained in the multivariate analysis, and prediction was weak (

Table 5. Prediction of having dealt with DFV situations from demographics and denomination.

	B	Standard Error	Exp(B)
Constant	0.667 #	0.298	1.949
Age (centered, in years)	0.021 *	0.007	1.021
Age (centered) squared	−0.002 *	0.000	0.998
Female	−0.359	0.205	0.699
Education: reference school/trade
Degree	0.053	0.248	1.055
Postgraduate	0.179	0.251	1.196
Country of birth: reference Australia
Other English-speaking	−0.17	0.245	0.844
Non-English-speaking	0.297	0.284	1.346
Denomination: reference Uniting
Anglican	0.242	0.236	1.273
Baptist/Churches of Christ	−0.014	0.255	0.986
Catholic	−0.061	0.362	0.941
Pentecostal	0.675 #	0.342	1.965
Salvation Army	1.785 **	0.349	5.957
Other Protestant	−0.469	0.279	0.626

Source: Powell et al. (2016c). n = 835 (all senior local church leaders). Chi-sq = 70.7, df = 13, p < 0.001. Pseudo R2 (Nagelkerke) = 0.113. # p < 0.05; * p < 0.01; ** p < 0.001. Data are unweighted.

,

Table 6. Prediction of marriage/couples counselling from demographics and denomination.

	B	Standard Error	Exp(B)
Constant	−0.266	0.361	0.767
Age (centered, in years)	0.014	0.009	1.014
Age (centered) squared	0.000	0.001	1.000
Female	−0.851 *	0.264	0.427
Education: reference school/trade
Degree	−0.545	0.288	0.580
Postgraduate	−0.607 #	0.287	0.545
Country of birth: reference Australia
Other English-speaking	−0.345	0.308	0.708
Non-English-speaking	0.378	0.318	1.460
Denomination: reference Uniting
Anglican	0.454	0.306	1.574
Baptist/Churches of Christ	0.638	0.328	1.892
Catholic	0.293	0.462	1.341
Pentecostal	1.134 *	0.399	3.108
Salvation Army	−0.300	0.379	0.741
Other Protestant	0.975 *	0.378	2.650

Source: Powell et al. (2016c). n = 560 (senior local church leaders who have dealt with DFV situations). Chi-sq = 60.6, df = 13, p < 0.01. Pseudo R2 (Nagelkerke) = 0.140. # p < 0.05; * p < 0.01. Data are unweighted.

,

Table 7. Prediction of counselling perpetrator from demographics and denomination.

	B	Standard Error	Exp(B)
Constant	−0.961 *	0.358	0.383
Age (centered, in years)	0.01	0.009	1.010
Age (centered) squared	−0.001	0.001	0.999
Female	−0.747 *	0.258	0.474
Education: reference school/trade
Degree	−0.045	0.283	0.956
Postgraduate	0.344	0.278	1.411
Country of birth: reference Australia
Other English-speaking	0.009	0.294	1.009
Non-English-speaking	0.093	0.323	1.097
Denomination: reference Uniting
Anglican	0.645 #	0.306	1.905
Baptist/Churches of Christ	0.413	0.331	1.511
Catholic	−0.232	0.510	0.793
Pentecostal	1.079 *	0.396	2.943
Salvation Army	0.634	0.359	1.886
Other Protestant	0.545	0.381	1.725

Source: Powell et al. (2016c). n = 560 (senior local church leaders who have dealt with DFV situations). Chi-sq = 30.5, df = 13, p < 0.01. Pseudo R2 (Nagelkerke) = 0.073. # p < 0.05; * p < 0.01. Data are unweighted.

and

Table 8. Prediction of referred victim to service agency from demographics and denomination.

	B	Standard Error	Exp(B)
Constant	1.512 **	0.428	4.535
Age (centered, in years)	−0.002	0.011	0.998
Age (centered) squared	0.001	0.001	1.001
Female	0.14	0.309	1.150
Education: reference school/trade
Degree	−0.498	0.349	0.608
Postgraduate	−0.068	0.356	0.934
Country of birth: reference Australia
Other English-speaking	−0.351	0.335	0.704
Non-English-speaking	−1.34 **	0.327	0.262
Denomination: reference Uniting
Anglican	−0.354	0.329	0.702
Baptist/Churches of Christ	0.019	0.368	1.019
Catholic	0.473	0.569	1.605
Pentecostal	−0.094	0.458	0.911
Salvation Army	1.073 #	0.482	2.923
Other Protestant	−0.322	0.416	0.724

Source: Powell et al. (2016c). n = 560 (senior local church leaders who have dealt with DFV situations). Chi-sq = 47.8, df = 13, p < 0.01. Pseudo R2 (Nagelkerke) = 0.125. # p < 0.05; ** p < 0.001. Data are unweighted.

). Age increased the likelihood of having responded to DFV overall (plateauing in the 50s—a curvilinear relationship), as did belonging to Pentecostal and Salvation Army churches (Table 5). The odds of having responded were twice as high for Pentecostals and six times as high for Salvationists as they were for Uniting Church leaders. In terms of the specific responses, prediction was strongest for marriage/couples counselling (14% of pseudo variance explained), with males, Pentecostals and “Other Protestant” leaders more likely to use this response, and those with postgraduate education less likely to do so (Table 6). Being female decreased the odds of counselling perpetrators (Table 7) and couples (Table 6), by 53% and 57%, respectively. Leaders born in countries where English was not the main language spoken where less likely than those born in Australia to refer victims to service agencies (74% lower odds of having referred them, Table 8). Salvationists were more likely than Uniting Church leaders to refer victims to service agencies. Anglican and Pentecostal leaders were more likely than their Uniting Church counterparts to have counselled a perpetrator (Table 7). Regressions were not statistically significant for the prediction of counselling victims, conducting safety risk assessments with victims, referring perpetrators to service agencies, and “Other response”.

3.2. Clergy Familiarity with Support Services

When asked “How familiar are you with support services available in your local community to assist victims and perpetrators of domestic and family violence (e.g., crisis centre, shelter, hotline)?”, only 17% of local church leaders indicated that they were very familiar with support services available. A majority were somewhat familiar (57%) and approximately one-quarter were not familiar (26%). Those with the lowest levels of formal education expressed greater familiarity than those with higher levels of formal education. There was no relationship between familiarity and country of birth. Familiarity varied by denomination, with Salvationists reporting the highest levels of familiarity. Male clergy were less likely to be familiar than female clergy and familiarity increased with age (

Table 9. Senior local church leader level of familiarity with DFV support services by demographics and denomination.

	Not at All Familiar %	Not Very Familiar %	Somewhat Familiar %	Very Familiar %	Effect Size (Eta-sq)
Overall	4	22	57	17
Age					0.021 **
Under 40 years	7	26	54	13
40–49 years	3	27	57	14
50–59 years	6	23	54	17
60+ years	2	15	61	22
Gender					0.009 *
Female	1	19	58	22
Male	5	23	56	16
Education					0.024 **
Certificate/diploma/school	1	18	60	22
Bachelor degree	6	25	57	12
Postgraduate	5	22	55	18
Country of birth					0.000
Australia	4	21	58	17
Other English-speaking	3	24	55	18
Non English-speaking	4	27	51	18
Denomination					0.030 **
Anglican	8	20	61	11
Baptist/Churches of Christ	4	23	59	14
Catholic	6	22	47	25
Pentecostal	1	18	62	18
Salvation Army	1	7	62	31
Uniting	3	28	50	19
Other Protestant	6	31	53	10

Source: Powell et al. (2016c). ANOVAs were conducted to determine statistically significant differences, * p < 0.01; ** p < 0.001. Effect size Eta-squared is shown. Data are weighted.

).

In multivariate analysis, familiarity was positively predicted by being female and being a Salvationist (compared with being from the Uniting Church), and negatively predicted by holding a bachelor degree. Age did not predict familiarity. Prediction was poor, with only 7% of variance explained (

Table 10. Prediction of senior local church leader level of familiarity with DFV support services by demographics and denomination.

	B	Std Error	Beta
Constant	2.674 **	0.157
Age (in years)	0.004	0.002	0.060
Female	0.137 #	0.064	0.078
Education: reference school/trade
Degree	−0.174 #	0.076	−0.110
Postgraduate	−0.006	0.076	−0.004
Country of birth: reference Australia
Other English-speaking	−0.006	0.080	−0.003
Non-English-speaking	−0.100	0.090	−0.038
Denomination: reference Uniting
Anglican	−0.109	0.079	−0.060
Baptist/Churches of Christ	0.018	0.085	0.009
Catholic	0.129	0.120	0.041
Pentecostal	0.119	0.108	0.046
Salvation Army	0.323 **	0.094	0.154
Other Protestant	−0.140	0.096	−0.058

Source: Powell et al. (2016c). # p < 0.05; ** p < 0.001. F (12, 848) = 6.107, p < 0.001. R2 (adjusted) = 0.067. Data are unweighted.

).

3.3. Approachability of Churches for Those Experiencing DFV

When churchgoers were asked “If you or someone you knew were experiencing domestic or family violence, do you feel that you could go to someone from the church for help?”, nearly two-thirds (64%) indicated in the affirmative. A total of 3 in 10 were not sure (29%) and 7% would not go to the church for help (

Table 11. Churchgoer views about approaching the church for help.

	Yes (%)	No (%)	Not Sure (%)
Overall	64	7	29
Age
15 to 29 years	67	3	31
30 to 49 years	69	8	24
50 to 69 years	63	7	30
70 years and over	62	8	31
Gender
Women	66	6	28
Men	63	8	29
Education
School	63	7	30
Diploma/certificate	65	8	27
Degree	64	7	30
Country of birth
Australia	63	8	29
Other English-speaking	67	5	28
Non English-speaking	66	6	28
Denomination **
Anglican	76	3	21
Baptist/Churches of Christ	73	6	21
Catholic	52	11	36
Pentecostal	79	1	20
Uniting	64	2	33
Other Protestant	73	4	23

Source: (Powell et al. 2016a). ** p < 0.001. Chi-square tests were used to evaluate if differences were statistically significant. Cramer’s V for denomination = 0.183. All other differences were not statistically significant. Data are weighted.

).

The gender of the respondent made no statistically significant difference to churchgoers’ responses and neither did age nor education nor country of birth. However, denomination did. Catholics were least likely to agree they could make an approach (52%), 64% of Uniting churchgoers indicated likewise, while approximately three-quarters or a little more of Anglican, Baptist/Churches of Christ, Pentecostal and “Other Protestant” churchgoers felt that they felt that they could do so (Table 11).

4. Discussion

This study aimed to explore clergy responses to domestic and family violence across Australian churches, clergy familiarity with support services, and churchgoers’ views about the approachability of churches in relation to DFV situations.

What proportion of clergy have dealt with DFV situations and what responses have they used? Most senior local church leaders—two-thirds—indicated that they had dealt with DFV situations. Unsurprisingly, as hypothesized, this increased with age, most rapidly in younger cohorts, with leaders more likely to be exposed to DFV the longer they were in ministry. Overwhelmingly, leaders who had dealt with DFV situations had dealt directly with a victim, typically by referring the victim to a service agency and/or by counselling the victim. Referral to service agencies is a key component in supporting victims of abuse, and is a principle recommended for faith leaders to effectively respond to and prevent DFV (Nason-Clark et al. 2017; Vaughan et al. 2020). A relatively high degree of referral of victims to service agencies was observed across denominations in this study. Approximately one-quarter of leaders had undertaken a safety risk assessment with the victim.

In comparison to responding to victims, far fewer leaders—although still a large minority—had dealt directly with a perpetrator, by means of counselling and to a lesser degree through referral to support services. Almost all of these leaders had also dealt with a victim. Male leaders were more likely than females to counsel perpetrators, as were Anglican and Pentecostal leaders.

A strong majority of Australian churchgoers in an intimate relationship have a partner who attends the same church as them (Powell et al. 2016b). The discrepancy between leaders responding to victims versus perpetrators might reflect the greater tendency of victims to seek help. Leaders might also be reluctant to deal with perpetrators, due to concern for the safety of victims, as well as awareness of the complexity involved in dealing with perpetrators. The discrepancy is likely to also reflect the gendered nature of both DFV and church attendance, given that women are much more likely to be victims of domestic violence than men (AIHW 2019; ABS 2017) and that some 60% of adults who attend church in Australia are female (Powell et al. 2017).

The propensity to offer marriage/couples counselling, by 41% of senior leaders who had responded to DFV situations, is problematic. Studies conducted in other countries have indicated that this practice is also common elsewhere (e.g., Choi 2015a; Horne and Levitt 2003; Sojourners, and IMA World Health 2014), with many clergy reluctant to see marriages end, even when they are abusive (Nason-Clark et al. 2017, p. 34). Prioritization of safety at all times is the first principle recommended by the Faith Communities Supporting Healthy Family Relationships project for preventing and responding to DFV (Vaughan and Sullivan 2019). Victims of domestic violence, services that support them and other experts express concerns about the safety and efficacy of couples counselling in that it fails to address the unequal power in an abusive relationship and can place the victim at increased risk. Standards and guidelines in many jurisdictions therefore discourage the use of couples counselling as a primary intervention for domestic violence (Mackay et al. 2015; Vaughan and Sullivan 2019). That most leaders who reported using couples counselling had also used other responses—particularly counselling the victim and/or perpetrator individually—suggests that couples counselling might not be commonly conducted in isolation. Nevertheless, the concerns remain—particularly the weak negative relationship between reporting couples counselling and victim referral to a service agency. Males (who are the majority of clergy), Pentecostals and “Other Protestant” leaders were particularly likely to have counselled couples, although the results for Pentecostal leaders should be understood as tentative because of the small number of survey participants. Education is needed to ensure that leaders understand the dangers involved in couples counselling.

How familiar are clergy with DFV support services in their local communities? While approximately three-quarters of senior leaders in this study considered themselves to be familiar to some degree, just one in six considered themselves to be very familiar. Being female and holding low levels of formal education (below bachelor degree) was related to higher levels of familiarity. These results suggest that more is needed to strengthen relationships between secular organizations and faith leaders, which is a key recommendation for faith communities to effectively respond to and prevent DFV, including regular contact and relationship between leaders and local specialist violence services, establishing referral pathways and ensuring that faith leaders have access to current information about local services (Vaughan et al. 2020). Although leaders born in countries where English is not the main language spoken felt that they were no less familiar with DFV support services than leaders born in Australia and other English-speaking countries, these leaders were less likely to refer victims to support services. More research is needed to understand why this is the case.

The results for Salvationist leaders are worthy of particular attention. The Salvation Army is renowned for its mission to people facing hardship, with its local church structure (corps) integrated with its social service work, with its clergy (known as officers) strong involved in welfare programs (Daniel 2009). In the present study, Salvationists were the most likely among the denominations to have dealt with DFV situations and to have referred victims to specialist support services. They were the least likely denomination to have counselled couples and were the only one with higher use of safety risk assessments than couples counselling. Levels of familiarity with local DFV support services were also highest in this denomination. Taken together, these results suggest a depth of experience with DFV and strength of awareness of the needs of DFV victims for safety and specialist support. More detailed research on the practices of Salvationist leaders may well be beneficial for churches across the denominations.

How confident were people in the pews that they could go to someone at their church for help if they or someone they knew were experiencing DFV? Nearly two-thirds felt that they could do so while approximately three in 10 were unsure. There was, however, a significant denominational divide in this sentiment. Only half of Catholics, nearly two-thirds of Uniting Church people, and three-quarters or more from other denominations felt they could make such an approach. This result suggests that attention is needed in Catholic parishes particularly, if Catholics facing DFV are to seek out and receive appropriate support from their church.

People from culturally and linguistically diverse backgrounds are at increased risk of DFV (AIHW 2019) and face barriers to seeking help (Vaughan et al. 2016, 2020). However, churchgoers from countries where English is not the main language spoken were not less likely to say that they would approach their church for help.

There are limitations of this study that relate to sampling, survey question wordings, and the currency of the data. We consider each of these issues in turn.

Sampling limitations associated with the NCLS include the non-probabilistic nature of Protestant samples and the under-representation of certain groups. Catholic samples are random. In contrast, there is a likely participation bias among Protestant churches related to church health (Pepper et al. 2018). While it is unclear whether or how this bias relates to DFV matters, that the DFV survey questions were such a small part of the survey forms is a strength of this study compared with other dedicated opt-in survey studies on religion and violence (e.g., Aune and Barnes 2018; Powell and Pepper 2021). The topic is unlikely to have affected participation in the NCLS.

While weighting addresses the low participation of Pentecostal and Catholic churches in the NCLS for overall statistics, small unweighted numbers of Pentecostal respondents in the leader and attender samples and of Catholics in the leader sample used for this study means that the results for these groups should be treated with caution. Larger samples would be required for a stronger analysis in these denominations as well as in the smaller Christian denominations, either as collected by the NCLS (by running the questions in more than one survey variant in future) or in dedicated studies.

Various commentators have highlighted the complexity and challenges in addressing DFV in culturally and linguistically diverse communities of faith (Truong et al. 2020, Vaughan et al. 2020). The present study was constrained in its capacity to examine DFV in culturally and linguistically diverse contexts and found few differences in the results for country of birth. The survey respondents from countries where English is not the main language spoken were mainly from multiethnic churches and were confident with English. Migrants are integrated into territorial Catholic parishes in Australia (McEwan et al. 2020; ACBC 2007), and multiethnic parishes are common (NCLS Research 2018). Random samples of Catholic parishes participate in the NCLS. “Mono-ethnic non-Anglo” churches are, in contrast, a part of the Protestant church landscape, but their participation in the NCLS is low (Pepper et al. 2018). Moreover, the Leader Survey and Attender Sample Surveys are offered in English only, which poses a barrier to participation for those less confident with English. Better, larger samples would be required to examine relationships between particular ethnic backgrounds and responses to DFV. The low number of cases afforded only general country of birth groupings in the present study.

Another set of study limitations concerns the framing and wording of the survey questions. The questions used the terms “domestic and family violence” (Leader Survey) and “domestic or family violence” (Attender Survey) without providing definitions. Leaders were asked to indicate the responses they had used when dealing with these situations. One response option referenced couples, and another “domestic violence” only. No other sorts of family relationships were explicitly referenced. Moreover, the question did not specify a timeframe, nor did it ask leaders how frequently they had used these responses, or which they had used concurrently for a particular family/couple/situation. Further, the term “counselling” is an ambiguous one. It might be understood by clergy to include pastoral care, and/or it could invoke the notion of professional counselling, which some might view to be the remit of specialist services rather than of clergy. A later dedicated DV study on Anglican leaders found that almost all clergy who had dealt with specific DV situations provided “pastoral or spiritual care” to a victim while approximately half had provided “counselling”, though that study was likely to be biased towards those clergy with a greater interest in the topic (Powell and Pepper 2021). The present research can only provide a broad-brushed picture of the sorts of responses church leaders have used. More detailed research, including more specific survey questions and qualitative research, would be required to explore these nuances.

Finally, the survey was conducted in 2016, in the early stages of an upsurge of attention given to DFV in the Australian media, by the government, and in the churches. Although the results are now dated, they do provide rare quantitative data on DFV responses in churches and form a baseline against which later waves of the NCLS or other studies can be compared.

In addition to research on the specificity of clergy responses to DFV in particular cultural contexts, including different denominations and culturally and linguistically diverse communities, future investigations should examine responses to perpetrators particularly. A large minority of clergy appear to be working directly with perpetrators. Given the complexities and risks involved in doing so (Nason-Clark et al. 2017), as well as the lack of evidence for what interventions with religious men are effective for accountability and behaviour change (Vaughan et al. 2020), understanding current practices in dealing with perpetrators is a priority.

In conclusion, this study has found that, in 2016, two-thirds of local church senior leaders had previously dealt with domestic and family violence situations in their ministry. They mainly responded to victims of abuse by referring them to specialist services and counselling them. A large minority had also counselled and/or referred perpetrators and had counselled couples. The findings suggest a depth of experience among Salvationist leaders with DFV and strength of awareness of the needs of DFV victims for safety and specialist support. While, overall, a substantial majority of churchgoers felt that they could approach their church for help if they were experiencing DFV, just half of Catholics felt that they could do so. Future research should explore responses to DFV in specific denominations and culturally and linguistic diverse contexts in more detail and seek to understand the practices used by clergy who are dealing with perpetrators.