Intervention programme for fathers who use domestic and family violence: Results from an evaluation of Caring Dads

This article presents findings from an evaluation of a trial of the Caring Dads programme in an Australian jurisdiction. Caring Dads is a nonmandated group-based programme for fathers who have been physically and/or emotionally abusive towards their children, their children's mothers or both. The programme aims to engage men around their fathering and is ultimately vested in and focused on children's safety, well-being and the intrinsic link this has with the safety of the mother. Fathers under-taking the programme ( n = 40), and associated mothers ( n = 17), participated in a mixed-methods self-report questionnaire administered at programme commencement and conclusion. This article presents findings related to change in fathers' use and mothers' experiences of domestic and family violence, shared parenting experiences, mother safety and levels of psychological distress between the two points of data collection. Results align with those produced by other national and international Caring Dads evaluations and show overall positive outcomes for mother experiences of violence, psychological well-being and safety, along with fathers' increased awareness and understanding of what constitutes abusive behaviour. Implications for father-focused interventions around domestic and family violence and the role of victim-survivor voices in delivering and evaluating father and family-focused interventions are discussed.


| INTRODUCTION
In Australia, where this research is based, domestic and family violence (DFV) was declared a national emergency in 2015 (The Australian, 2015).Globally, DFV has repeatedly been identified as a major public health crisis, disproportionately affecting women and children (World Health Organization, 2017).As DFV is a gendered issue, children are disproportionately exposed to violence perpetrated by their father or a father-like figure against their mother or a motherlike figure (Australian Institute of Health & Welfare [AIHW], 2018; Brown et al., 2016;Campo, 2015).Prevalence estimates suggest that between one in four and one in six children grow up experiencing parental DFV (Devaney, 2015;McTavish et al., 2016).Even when families separate, violence may persist or worsen, and children may continue to be affected by DFV and have ongoing contact with the abusive parent (Areán & Davis, 2006;Campo, 2015;Holt et al., 2008).
Fathers, like mothers, play a critical role in promoting child wellbeing and development (Holt et al., 2008;Panter-Brick et al., 2014).
Research finds that positive father involvement in children's lives is important for reducing children's psychosocial and criminogenic risk and for enhancing their social, emotional and cognitive functioning (Allen & Daly, 2007;Marshall et al., 2001).Inversely, abusive and controlling fathering techniques have been linked to a range of adverse developmental outcomes for children (McTavish et al., 2016).Systematic review evidence indicates fathers who perpetrate DFV may use inconsistent, controlling or authoritarian parenting styles (Holt et al., 2008;Humphreys & Campo, 2017).Further, they may have poor understandings of child development or role modelling of positive relationships (Holt et al., 2008;Humphreys & Campo, 2017).
Fathers may not intend the consequences their behaviour has on children's short-and long-term development.However, their overcontrolling behaviour, a sense of entitlement or the implicit belief that one's child's love is unconditional often have adverse effects on child outcomes (Humphreys & Campo, 2017;Stover & Morgos, 2013).
Supporting abusive fathers towards non-violent co-parenting and parent-child relationships has the capacity to improve short-and long-term outcomes for children who have been exposed to DFV (Featherstone & Fraser, 2012;Humphreys & Campo, 2017).
In addition to the impact of fathers' use of DFV against the other parent or carer on children's development and well-being, fathers' use of DFV has far-reaching consequences for the victim-parent/carer.Aside from the well-established range of adverse effects of DFV on victims' immediate safety and well-being (AIHW, 2018;Wong & Mellor, 2014), evidence suggests that experiences of DFV can negatively affect a victim's parenting capacity (Fogarty et al., 2019).This may be due to a number of factors, ranging from being physically incapacitated due to the need to recover from injuries to lacking emotional presence where victims' day-to-day lives are preoccupied by coping with trauma and minimizing ongoing risk to themselves and their children (Meyer, 2018).While DFV does not affect all victimparents equally, many experience the strategic undermining of the parent-child relationship as part of the 'perpetrator toolbox' (Fogarty et al., 2019;Meyer & Stambe, 2020).Research has therefore begun to highlight the need to address perpetrator accountability within the co-parent and parent-child relationship (Featherstone & Fraser, 2012;Humphreys & Campo, 2017;Labarre et al., 2016;Stover & Morgos, 2013).
Despite evidence suggesting some fathers have a poor understanding of child development, nonabusive parenting techniques and the impact of DFV on children more broadly, fathers who perpetrate DFV may be willing to improve their parenting (Meyer, 2017;Rothman et al., 2007).Here, it is noteworthy that evidence-based Men's Behaviour Change Programs (MBCPs) are often required for men to recognize and acknowledge the impact of their behaviour in the first place (Humphreys & Campo, 2017;Meyer, 2017).Multiple studies report father motivation to attend and succeed in MBCPs, as well as motivation to change DFV behaviours, often centres around improved parenting and relationships with their children (Pennell et al., 2014;Stanley, Fell, et al., 2012;Stanley, Graham-Kevan, & Borthwick, 2012).Indeed, while research on father engagement is relatively scarce (Pfitzner et al., 2017), evidence suggests that tapping into a man's desire to be a better father may be an effective means through which to address his violence and motivate intervention programme attendance (Carlson & Casey, 2018;Humphreys & Campo, 2017;Meyer, 2017).
MBCPs are a mainstream approach to reducing male perpetration of DFV.Some authors suggest that these programmes are limited due to their one-size-fits-all approach that is somewhat inflexible in tailoring approaches to suit individuals (Meyer et al., 2020;Stover & Morgos, 2013).Indeed, MBCPs typically cater for a general DFV perpetrator population rather than subgroups who share specific characteristics, such as the unique experience of fatherhood (Brown et al., 2016;Feder & Wilson, 2005;Humphreys & Campo, 2017).Rivett (2010) argues that the practice of delivering interventions for this population is contested and there are many different perspectives regarding what should be done, and how and by whom it should be implemented.Due to the key role fathers play in their children's development, the heightened risk for families when their fathers perpetrate DFV and the purported positive impact specialized programmes may have for families with fathers who perpetrate DFV, there is a need to understand the effectiveness of specialized DFV programmes that include a focus on fathers as perpetrators.
These specialized programmes come in a range of forms.They may be stand-alone programmes run independently from traditional MBCPs, delivered as supplemental content within traditional programmes, or interventions for couples and/or families (Labarre et al., 2016).Additionally, they have varying objectives (e.g., increasing accountability while decreasing violence or fostering positive father-child relationships) and are underpinned by a range of theoretical frameworks (e.g., feminism, restorative justice, family system theory, motivational interviewing and psychoeducation) (Labarre et al., 2016;Rizo et al., 2011).One of the most prominent fatherfocused MBCPs internationally is the Caring Dads (CD) programme, which commenced in Canada 20 years ago and is now being utilized in a number of Canadian, US, UK, European and Australian jurisdictions (CD, 2017).

| Caring Dads
The CD programme caters for men who have been physically and emotionally abusive towards their children, their children's mother or both (Scott et al., 2006).The programme is not mandated, and fathers may be referred to attend from a myriad of places, including child protection, probation and parole, mental health agencies, facilitators of traditional MBCPs, community organizations, police, courts or selfreferral (Scott, 2010;Scott et al., 2006).CD is run in a group setting, ideally by one male and one female trained facilitator (Brady & Scott, 2012).Participants attend a 2-h session every week for 17 weeks (McConnell et al., 2017).
The programme consists of four modules, which aim to (1) engage men in examining their fathering, (2) teach awareness and application of child-centred parenting, (3) encourage awareness of and responsibility for abusive fathering behaviours and (4) address rebuilding trust with children and plans for the future (Scott, 2010).It uses components of traditional MBCPs, motivational interviewing and child development but remains vested in and focused on children's safety, well-being and the intrinsic link this has with the safety of the mother.
CD therefore incorporates systematic outreach to mothers to ensure their safety and well-being.The programme concentrates on placing the responsibility for being an appropriate parent on men and acknowledges that being non-violent towards women and children is critical to healthy fatherhood and safe children.CD harnesses the potential motivation of 'being a good dad' to encourage participation and behaviour change but also utilizes motivational interviewing to increase engagement.For the mother contact component, facilitators aim to contact mothers at three times during the programme to provide information about the programme, refer mothers to applicable support services and ensure that mothers and children are safe (Scott, 2010).
Previous evaluations of the programme show promising results.
A preliminary analysis from Canada found that postprogramme attendance, fathers experienced lower levels of hostility, rejection and denigration of their child, as well as anger towards family situations (Scott & Crooks, 2007).Further evaluations have found positive results for parenting and co-parenting (Scott & Lishak, 2012) as well as emotional responsiveness (Hood et al., 2015), father contact with child protection workers and re-referral (Scott et al., 2021) and awareness of the impact of their behaviour on their children (Diemer et al., 2020;McCracken & Deave, 2012).Qualitative research suggests that the programme can offer a safe space for fathers to discuss sometimes uncomfortable details of their lives and also equip them with skills for deescalating DFV situations (Parker, 2015).A recent evaluation from Victoria, Australia, found that participation in CD equipped fathers with a greater ability to reflect on their harmful fathering practices, alongside recall and application of the parenting skills that they had learned (Diemer et al., 2020).A number of evaluations have also found that mothers who were involved with CD participants may feel safer and report fewer instances of DFV (Diemer et al., 2020;McConnell et al., 2017).While past evaluations are particularly promising with regard to fathering and father-child relationships, some also note limitations around improvements in fathers' hostile attitudes towards mothers in the context of DFV (cf.Diemer et al., 2020).Similar limitations have been observed in other research with men using DFV, which reveals that attitudes and behaviours around DFV related to the impact on children may be easier to achieve than a shift in attitudes and behaviours related to the intimate (ex)partner (Meyer, 2017).

| The CD programme evaluation in an Australian jurisdiction
In the current study setting, 20 CD groups were delivered over five trial rounds between July 2017 and June 2019 and incorporated the following components in line with the licenced programme guidelines: 1. Group Work: the programme provider delivered a 17-week group programme to fathers, comprising of 16 weekly 2-h closed group sessions and one individual session, held in Week 10, to review progress and share information with relevant individual referral partners.Fathers were not mandated to attend a certain number of weeks in the programme to be considered completers or noncompleters and were invited to participate in subsequent weeks if they missed a session.Additionally, they were not mandated to attend make-up sessions to 'complete' the programme.This is different to CD implementation in other jurisdictions and has been discussed in the full evaluation report (Meyer, Hine, & McDermott, 2019).
2. (Ex)partner Engagement: the programme provider agreed to make at least three contacts with mothers over the course of the programme to provide information about the programme and assist with safety planning and referral pathways to relevant support.
Contact with mothers was integral to ongoing safety, risk assessment, management and monitoring progress.

| Psychological well-being
Participants were asked to report global psychological distress using the Kessler10 Psychological Distress Scale (K10).This 10-item scale is a commonly used assessment tool in public health treatment and intervention approaches and measures a person's agitation, depression, nervousness and psychological fatigue over the last 4 weeks (Australian Bureau of Statistics, 2012).As an example, the scale asks, 'During the last 30 days, about how often did you feel tired out for no good reason?'For each question, respondents report on a 5-point Likert scale whether they experienced those feelings none of the time (1) through to all of the time (5).The scores across all measures are summed, producing a minimum of 10 and a maximum of 50.Thus, higher scores on the K10 indicate higher levels of psychological distress, which are grouped into four bands indicating levels of global psychological distress: Low (10-15), Moderate (16-21), High (22-29) and Very high (30-50).

| Mother safety
Mothers answered one general question about their perception of safety: 'Overall, how safe do you currently feel in relation to your (ex) partner?'.Participants responded on a Likert scale from not safe at all (0) to very safe (1).

| PARENTING PROBLEMS CHECKLIST
Participants were asked about issues around child rearing that may arise for parents.This was assessed using the 16-item Parenting Problems Checklist.Specifically, this tool assesses conflict concerning parenting, such as 'in the last four weeks, to what extent has inconsistency between parents been a problem for you and your partner?'These items were measured on a 5-point Likert scale from not at all (0) to very much (4).Scores across all items were summed to create  Despite this limitation in the statistical power of the analyses, findings that are not statistically significant still contribute meaningful insights into the experiences of mothers and fathers within this study.Further, while mean changes between Waves 1 and 2 may not be statistically significant, they may be in the expected direction (i.e., showing an improvement in scores at Wave 2 measurement).We produced effect sizes with Hedge's g correction to further examine potential change between Waves 1 and 2 of the survey.Some short response qualitative feedback is presented as supplemental to the quantitative analyses.Given our low sample sizes, we also provide Table S1 showing the N and %.Fathers were informed during their intake process into CD that the programme was funded as a trial with associated evaluation research conducted by CQUniversity.All programme participants were provided with a participant information sheet, outlining the nature and purpose of the evaluation, and asking them to provide preliminary consent to be contacted by a researcher to receive more information about the evaluation.The female evaluation lead and a male research assistant liaised with facilitators in each region to attend a session in Week 1 or 2 of the programme to provide more detail about the evaluation and answer any questions.After attending this session, consenting fathers were contacted via phone by the male research assistant to organize a time to complete the $45-min Wave

| Recruitment of mothers and related data collection
The evaluation aimed to incorporate the voices of (ex)partners who shared parenting responsibilities for one or more children with fathers involved in CD.This was seen as a crucial element of the evaluation due to the role of victims' voices in examining the effectiveness of perpetrator-focused interventions highlighted in other evaluation research (e.g., Kelly & Westmarland, 2015;Nicholas et al., 2020).
The programme provider assisted in the recruitment of mothers relevant to the evaluation.Facilitators informed the (ex)partners of programme participants about the evaluation during initial contact with mothers and obtained consent for them to be contacted by the research team to find out more about the evaluation.At the time of preliminary telephone contact between mothers and a researcher from the evaluation team, mothers were fully informed about the research, were offered an email copy of the participant information sheet and were asked to identify a suitable time to complete the $45-min telephone survey.All participating mothers were asked to provide audio-recorded verbal informed consent to complete the survey and to be contacted again at Wave 2 to establish whether they would be willing to complete the follow-up survey.At the end of each survey, mothers received information about available support should this be required.Mothers were further asked at the conclusion of the survey whether they would like the CD facilitator providing the victim advocacy work and support component to follow up for an opportunity to debrief.
To compensate research participants for their time and contribution, both mothers and fathers received a $30 shopping voucher for each follow-up survey they completed.

| Survey response rates
Of the 109 fathers who participated in CD, 40 participated in the initial survey.Table 2 shows the survey response rates for both this population and participating mothers (n = 17).There was around 50% attrition at Wave 2 for both mothers and fathers.While the mean number of contact attempts ranged from 2 to 3.3 for fathers and 3.1 to 5.1 for mothers, most dropouts in this study were due to participants no longer being contactable (e.g., due to relocation and/or change in phone numbers).

| Demographic characteristics
Table 3 denotes the demographic information for the CD population and a comparison between research and nonresearch participants.
Overall, the average participant in CD (regardless of their participation in the evaluation) was in his thirties, non-Indigenous, had three children and was in a relationship with and coresided with the children's mother.Independent samples t tests and chi-square tests showed no difference between the research and nonresearch participants for most variables in We did not obtain the same information for mothers who chose not to participate in the research.However, the average mother who did complete the Wave 1 survey was 29.68 years old (SD = 7.49), non-Indigenous (n = 12, 70.6%) and had an average of 2.2 (SD = 0.83) children.
In general, demographics were similar across the two genders for the research samples (see Table 4).Over a third of both samples were in de facto relationships, and many lived with a partner and/or Results indicate that while mothers reported experiencing less harassment and abuse over time, fathers reported a slight increase in self-reported behaviours between Waves 1 and 2. Table 5 shows the effect sizes for these relationships, with the direction of the effect favouring Wave 2 for mothers and Wave 1 for fathers.Both effect sizes are small, and the confidence intervals include zero.It is important to note here that fathers underreported the use of harassment and abusive behaviours at Wave 1 compared with mothers' selfreported experiences.Fathers' increase in self-reported harassment and abusive behaviours by Wave 2 did not exceed mothers' selfreported experiences of fathers' behaviours and is likely a reflection of fathers' greater insight into and acknowledgement of these behaviours after completion of the program.
Positively, for the Sexual and Physical Violence scale, the direction of the effect favours Wave 2 for both mothers and fathers, with small effect sizes, although the confidence intervals include zero (see Table 5).Both mothers and fathers reported a decrease in mean scores across survey waves.Similar to reports of Harassment and Abuse, fathers reported lower levels of physical and sexual violence than mothers' self-reports of their own experiences across the two waves.
This is likely the result of low levels of physical and sexual violence reported by mothers and fathers to start with compared with other forms of violence, thus offering limited room to decrease significantly.
Mothers were also asked to identify their perceived space for action within their relationship, measuring the degree of coercively controlling behaviours experienced by mothers.The t test for group differences between Waves 1 and 2 showed a statistically significant decrease in scores identifying the presence of coercively controlling behaviours, suggesting that mothers were experiencing fewer problematic behaviours by their (ex)partner over time (see Table 5).Similarly, the effect size for this analysis was large, indicating that the effect favours Wave 2. Figure 3 shows the decrease in mean scores from Waves 1 to 2.
Figure 4 and Table 5 show mothers' perceptions of the level of respectful communication in their relationship.This figure shows an upward trend in reported respectful communication, with scores increasing from Waves 1 to 2 for mothers who had ongoing contact with their (ex)partner.The direction of the effect favours Wave 2, with a medium effect size, but the confidence intervals included zero.

| Mothers' safety
Mothers experienced a positive change in their self-perceived level of safety as it related to their (ex)partners.Results show that mothers reported feeling safer at Wave 2, with an increase in mean scores from 2.11 to 2.56.While this result is not statistically significant, the effect size does favour Wave 2, albeit this is a small effect (see Table 5).

| Psychological well-being
With regard to psychological distress, mothers reported a decrease at Wave 2, with scores moving from the 'High' psychological distress band to 'Moderate' (see Figure 5).While result is not statistically significant, the effect size, albeit small, favours Wave 2 (see Table 5).
Figure 6 and Table 5 show the results of the t test examining the difference in fathers' scores on the Kessler10 between Waves 1 and 2. The graph and effect size indicate that fathers reported less psychological distress at Wave 2, moving from the 'High' band to 'Moderate' between programme intake and exit.

| PARENTING PROBLEMS CHECKLIST
Both mothers and fathers completed the Parenting Problems Checklist (see Figure 7 and Table 5) to examine problems around coparenting in present couples as well as separated co-parents.While overall, mothers reported higher problem scores than fathers, both reported less conflict at Wave 2. Mothers' mean scores on the checklist decreased from 29.16 at Wave 1 to 18.83 at Wave 2. Similarly, there was a decrease in fathers' mean scores from 22.16 at Wave 1 to 17.89 at Wave 2. The direction of the effect favours Wave 2 for both mothers and fathers, with a small effect size for fathers and a medium effect size for mothers, although the confidence intervals included zero for both groups.Similar to other self-report findings presented in this paper, the difference in problem scores reported by mothers and fathers is likely the result of the victim parent (in this case mothers) experiencing parental conflict and inconsistencies in parenting as more problematic than the parent using DFV (in this case, fathers).

| Qualitative feedback
The open-ended survey questions provide some interesting findings to complement the statistical analyses.First, when asked why they agreed to participate in the programme, the top two responses provided by fathers were that they (a) wanted (or were mandated) to satisfy involved agencies and (b) desired to improve their parenting techniques and skills.Other common responses included improving their own behaviours (e.g., better self-management of anger and stress and improved coping skills) and wanting more access to their children.
Second, when women were asked whether they felt any negative effects or unintended consequences from the fathers' participation in CD, none were identified.This is a positive observation as past research has highlighted the need to ensure men's participation in DFV-focused interventions does not have adverse effects on victimsurvivors (Chung et al., 2020;Meyer et al., 2020).

| DISCUSSION
While the small sample size limits wider generalizability of our findings, this CD evaluation aligns with those from previous evaluations of the programme (Diemer et al., 2020;McConnell et al., 2016;Scott & Crooks, 2007;Scott & Lishak, 2012;Scott et al., 2021).Overall, fathers' participation in CD appeared to have positive outcomes for mothers, with mothers reporting improvement across all measures.
These indications from mothers are encouraging given that some fathers may provide socially desirable answers to questions around While mothers reported feeling safer in relation to their (ex)partner, mothers reported higher prevalence rates of experienced physical and sexual abuse along with harassment compared with fathers' acknowledgement of having used these forms of DFV.This discrepancy has also been observed in other men's programme evaluations around DFV (cf.Meyer et al., 2021), which reflects the point above where fathers may have underestimated the impact of their behaviour on mothers' safety.

| Fathers' self-reports
Unsurprisingly, fathers' self-reports of DFV behaviours against their (ex)partner was lower than that reported by mothers over the study period.Underreporting of abusive behaviours was anticipated and can be a common occurrence in perpetrator-focused research (Kelly & Westmarland, 2015;Nicholas et al., 2020), particularly at the commencement of an intervention.Following programme completion, fathers appeared to develop an increased awareness and understanding of what constitutes abusive behaviour, potentially providing more realistic or insightful measures of their abusive behaviour as the study progressed.
Observations around fathers' underreporting of abusive behaviours and underestimation of their impact on other family members highlight the need to regard any self-reported perpetrator behaviour, F I G U R E 7 Mother and father scores on the parenting problems Checklist especially at intervention onset, with care.More importantly, this highlights the importance of including victim-survivor voices in evaluation research centred on perpetrator behaviour to capture a more accurate reflection of both adult and child experiences of DFV and its impact on individual family members (Chung et al., 2020;Day et al., 2019;Kelly & Westmarland, 2015;Meyer et al., 2020).As MBCPs are designed to prioritize the safety of women and children, future programme evaluations must ensure the inclusion of their voices (Nicholas et al., 2020).

| Co-parenting relationships
Over the course of fathers' engagement in CD, inconsistencies and problems reported between existing and separated parent couples decreased.This observation is particularly positive in the context of DFV-affected couples, as both mothers and fathers in the current study reported a decrease in parenting problem behaviours.Coparenting in the context of DFV has repeatedly been identified as problematic and at times unsafe, especially for mothers in their role as primary victims and carers (Fogarty et al., 2019;Meyer, 2018;Meyer & Stambe, 2020).The improved co-parenting experiences observed here are likely to be closely tied to a decrease in abuse, harassment and controlling behaviours experienced by mothers in this study over time.This is important to note that parenting-focused programme content alone is unlikely to address problem behaviours associated with co-parenting unless it is integrated into a wider focus on accountability around abusive fathers' use of violence and coercion in their co-parenting relationships.

| Study limitations
There are a number of limitations to the findings presented here.First, the small sample sizes lack the statistical power needed for inferential and predictive methods and thus limit the generalizability of our findings and make it difficult to establish correlational claims between variables.The attrition rate across the waves of the survey (especially for mothers) shows that this is a difficult population to reach and engage in longitudinal research.It is not uncommon for research with highly vulnerable and hard-to-reach populations to encounter high attrition rates, especially when relying on telephone follow-up contact as the mode of data collection (Diemer et al., 2020;Meyer, Hine, McDermott, & Eggins, 2019).Mothers and families previously or currently affected by DFV may be facing a variety of demands on top of everyday life, including trauma, housing stress, financial hardship and the need to comply with statutory interventions.In addition, those who have separated may feel that his programme participation and behaviour change is his responsibility, and thus, making time for research participation can play a subordinate role (Diemer et al., 2020;Kelly & Westmarland, 2015).
Second, the evaluation is unable to establish causal relationships between the intervention and mothers' improved experiences of safety and co-parenting.In the absence of a control group, it is impossible to determine whether the circumstances of mothers have improved due to fathers' participation in CD and/or for reasons not captured within the scope of this study.Given the extensive recent DFV policy and practice reforms in the study locations (cf.Special Taskforce on Domestic and Family Violence Queensland, 2015), CD families may have also benefitted from wider improvements in community and service responses to families affected by DFV.
Further, the small sample size did not allow for an analysis of specific subgroups, including the small number of First Nations programme participants.Cultural and linguistic diversity was not specifically explored in the small sample size and did not emerge as a theme in the limited qualitative data presented here.Future research on the CD programme should therefore consider the exploration of culturally sensitive experiences and support needs of First Nations and culturally and linguistically diverse programme participants and family members.
Finally, children's voices did not form part of the evaluation.
While the research team received ethical clearance to include children aged 7 years and older in the research, we were unable to recruit children for a number of reasons, including because a majority of children were aged under 7 years, children were identified as too traumatized to be involved in the research and parental objection to child participation, preventing them from being interviewed.

3.
Case Coordination: the programme provider agreed to have open and regular communication with referral partners and other service providers relevant to the delivery of the programme.This included joint stakeholder meetings and established protocols for risk and safety management.Data presented in this article forms part of the wider independent process and outcome evaluation of the implementation of this CD trial (seeMeyer & Hine, 2018; Meyer, Hine, & McDermott, 2019).The overarching aim of the evaluation was to inform future implementation of the programme by identifying for whom the programme works, under what circumstances and how.This article presents key findings from the outcome evaluation.2 | METHODSThis article presents findings from 40 fathers and 17 mothers who agreed to participate in the evaluation of CD in two trial regions.A self-report questionnaire was used allowing for both quantitative and qualitative response formats.The questionnaire asked mothers and fathers to report on a range of items including demographic characteristics, living arrangements, use and experiences of DFV, child wellbeing and shared parenting experiences.In this article, the focus of analysis is on understanding mothers' and fathers' experiences in relation to use and experience of abusive behaviours, their psychological well-being, mothers' safety and co-parenting across the two survey time points.Data for mothers and fathers are analysed separately and provide group comparisons of outcomes and responses at Week 1 or 2 of the programme (Survey Wave 1) and again at Week 17 (Survey Wave 2). 1 Some comparisons are also drawn between characteristics of CD participants who were involved in the evaluation and those who were not to establish whether evaluation participants were representative of the wider CD programme participant population.

|
Experience and use of DFVTo measure the use (fathers) and experience (mothers) of abusive behaviours within the surveys, the Safety and Freedom from Violence and Abuse for Women and Children measures(Kelly & Westmarland, 2015) were used.The 30 individual items from this measure are used to construct four subscales targeting different aspects of DFV.All items were measured on a Likert scale from 1 to 5 and scores for each subscale were created by taking the mean of each participant's answers.The subscales are Physical and Sexual Violence, Harassment and Abuse, Expanded Space for Action and Respectful Communication.Mothers and fathers both responded to the questions on the Physical and Sexual Violence and Harassment and Abuse subscales, reporting on their experiences (mothers) and use (fathers) of these types of violence.Higher mean scores equate to greater levels of violence for these scales.Only mothers responded to the questions for the latter two subscales.Lower mean scores on the Expanded Space for Action subscale are indicative of mothers having more space for action, or of fathers using fewer controlling tactics and behaviours.The Respectful Communication subscale measures mothers' experiences of (ex)partner's communication with them; for this subscale, higher mean scores equate to greater respectful communication experienced by mothers.It should be noted that, to our knowledge, there has been no previous reports on the items or subscale construction for the Safety and Freedom from Violence and Abuse for Women and Children measures.
a total Parenting Problems score for each participant with a minimum possible value of 0 and maximum of 64.The original Parenting Problems Checklist measure uses a 7-point Likert scale, whereas surveys conducted for this evaluation used a revised 5-point Likert scale.Due to this, caution should be exercised if relying on published psychometric data (see Parent Problem Checklist, 2013).
5 | ANALYSISPaired samples t tests were employed to analyse change in survey responses between Wave 1 and Wave 2 using the 95% confidence threshold whereby P ≤ 0.05 denotes statistical significance.Chisquare tests of independence were used where comparisons were made between fathers who participated in the research and those who did not.Given the small sample sizes, caution should be taken when making causal claims based upon the paired samples t tests.

5. 1 |
Participant recruitment, survey response rates and demographic characteristics 5.1.1 | Recruitment of fathers and related data collection One hundred and nine fathers participated in CD during the evaluation timeframe, with 40 of these ultimately engaging in the evaluation component.In terms of referral pathways into the programme, a majority came to the programme through referrals from Child Safety or Probation and Parole (see Table 1).A chi-square test found no difference between the research participants versus nonresearch participants for referral pathway (χ 2 (4, n = 108) = 5.801, P = 0.215).

1
telephone survey prior to the third programme session.Before commencing each survey, the researcher obtained the father's informed consent to participate and to be contacted again with regard to Wave 2 data collection.Participants were given the opportunity at each follow-up to opt out of the research.For fathers who attended the programme up until Week 17, the programme provider assisted with the administration of Wave 2 data collection by allowing the research team to administer these surveys face to face during the final week of the programme.To minimize impact on programme time, the evaluation team worked with the facilitators to schedule some participants to arrive early and complete the survey prior to commencement of the final session.Remaining fathers were interviewed during programme time in a private office or counselling space on the premises.The evaluation team attended Week 17 sessions, with up to three researchers, to conduct up to three surveys in parallel to minimize impact on men's attendance of Week 17. Fathers who did not attend Week 17 of the programme were contacted via the phone number they had provided at Wave 1.If contactable via phone, the researcher organized a mutually suitable time to complete the Wave 2 survey via telephone.
their behaviour towards their (ex)partner.Incorporating mothers' data in evaluations of father-focused interventions and MBCPs more broadly is therefore critical to ensure a more accurate reflection of father behaviour as it relates to use of DFV and related behaviour change over time.It is noteworthy that change reported for fathers appears to be more prominent in relation to co-parenting than attitudes and behaviours directly related to DFV.Similar observations have been made in other recent CD evaluations (cf.Diemer et al., 2020), suggesting the programme may have more positive effects on attitudes and behaviours related to DFV and its impact on children as opposed to the hostile attitudes that often underpin violence directed at an (ex)partner.8.1 | Safety and well-being of mothersThe programme aims to increase the safety and well-being of children and that of their mothers.Over the study period, there were improvements in mothers' perceptions of safety, experiences of DFV (with less harassment and abuse, sexual and physical violence and problematic behaviours displayed by the father) and with more respectful communication.These improved outcomes corresponded with mothers experiencing less psychological distress and qualitative feedback from mothers around programme benefits.Here, most mothers reported that their circumstances had improved since their (ex)partner's participation in CD, particularly around family interactions and communications.
While limited to a small sample size of programme participants and their (ex)partners, findings from this evaluation indicate some benefits of father-focused perpetrator interventions also observed in larger CD programme evaluations (cf.Diemer et al., 2020;McConnell et al., 2017), including a reduction in abusive behaviours (identified through victim and perpetrator accounts of experiences and use of violence respectively) and improved psychological well-being among programme participants and (ex)partners.These observations highlight the impact of DFV on parental psychological well-being and demonstrate that both victim and perpetrator psychological well-being increases as violence perpetrated against women and children in DFV-affected families decreases.This observation may be particularly useful to consider in the engagement and motivation of men who use DFV in MBCPs.In addition to promoting the benefits for adult and child victims of DFV, promoting improved individual well-being may be a useful component of the initial motivation and engagement process.Further, some promising results were observed for shared parenting practices.While mothers appeared to be more adversely affected by parenting problems, both fathers and mothers reported a reduction in overall parenting problem behaviours.It is important to note here that this observation may primarily apply to parents who remain together, as these couples made up almost two thirds of the evaluation sample.To ensure safe and improved co-parenting relationships between separated couples affected by DFV, CD-involved fathers may require more intensive education and support to recognize the benefits of supportive and respectful co-parenting relationships, even where the couple relationship is irreparable.Separated couples may further require more intense risk and safety monitoring by the programme.This may in part be addressed through a dedicated family safety contact worker who is in a position to ensure regular and ongoing support to mothers, both separated and coresiding.This evaluation was unable to capture the voices of children; however, the strong focus on improved co-parenting relationships along with improved child well-being as a key objective of the programme suggests a need for future evaluations of CD (along with other father-focused DFV intervention programmes) to include the voices of children to better understand their experiences and support needs as victims in their own right.This has also been raised in other CD evaluations (cf.Diemer et al., 2020) and will be equally critical in assessing the benefits of CD for children with coresiding parents and those subject to shared parenting arrangements.While this study focused specifically on the CD programme, it does provide social workers and human service professionals with an understanding of the potential benefits of MBCPs that aim to motivate men's behaviour change around DFV through their father-child relationships.With the shift in social work practice over the past few years towards increasing the focus and accountability on fathers who use violence(cf.Featherstone & Fraser, 2012;Humphreys & Campo, 2017;Pfitzner et al., 2017), practitioners have the opportunity to engage, educate and support abusive fathers in redressing harm to children and their mothers through father-focused DFV interventions.

Table 3
Paired samples t tests and effect sizes for outcomes a Three missing responses for this measure.