Making fathers relevant: How practitioners include both parents in talk about parenting programmes. Child and Family Social Work

Parenting programmes are recommended as an effective means to support parents in promoting positive relationships with, and managing the behaviour of, their children. One barrier that impedes their successful implementation is that partners, especially fathers, are less frequently recruited by child welfare services. This article reports on a study that investigated how both parents were engaged with parenting services. Direct recordings were made of initial telephone conversations between six practitioners and 28 parents referred to those services and investigated for evidence of how the other parent was recruited. Conversation analysis was used to identify how participants introduced the possibility of both parents being included in the service, how these possibilities were negotiated, and what eventual agreements were made for both parents to be included in future arrangements. Implications for practice, training, and future research are considered.

not to disaggregate attendance and engagement figures by gender that can obscure any differences between mothers and fathers (Philip & O'Brien, 2017).
Some evidence suggests that there may be benefits for developing programmes specifically for groups of fathers in particular circumstances such as fathers in prison (Hayes, Butler, Devaney, & Percy, 2018;Langston, 2016) or fathers of at-risk children (Scourfield, Allely, Coffey, & Yates, 2016). However, when there are more general difficulties with children's behaviour or the parental relationship, there is evidence of improvements to paternal behaviour when parents attend co-parenting programmes together (Pilkington, Rominov, Brown, & Dennis, 2019). The potential benefit to the parental relationship is supported by evidence from a trial conducted by Cowan, Cowan, Pruett, Pruett, and Wong (2009) that reported significantly reduced parenting stress and increased satisfaction with the marital relationship when parents attended together. Conversely, mothers who completed a programme on their own reported difficulties in implementing strategies at home because of resistance from their partners, resulting in diverging parenting practices and increased parental tension (Mockford & Barlow, 2004). This is important because a primary influence of children's well-being is the quality of the parental relationship (Harold, Acquah, Sellers, & Chowdry, 2016), the implication being that when it is safe to do so, services should engage fathers as well as mothers to attend parenting programmes.
Despite the widespread acceptance that fathers should be engaged as part of a whole family approach, achieving this in practice represents an ongoing challenge for services and practitioners. Some fathers have told researchers that they avoid parenting services for fear of being "dictated to" (Bayley, Wallace, & Choudhry, 2009) or "told what to do" (Butt, 2009). Although some practitioners are committed to working with fathers (Scourfield, Cheung, & Macdonald, 2014), others have, in the past, conceptualized fathers as "not relevant" (Scourfield, 2003). In Ireland, fathers were excluded from services because stories about them would "float around the system" through case files and organizational communication (Ferguson & Hogan, 2004).
In a Canadian study, mothers reported withholding information about the father for reasons related to their safety from domestic violence or concern over loss of financial benefits, resulting in practitioners remaining unaware of his presence (Dominelli, Strega, Walmsley, Callahan, & Brown, 2010). Some practitioners have told researchers that they believe mothers have the right to determine whether the father should be involved in a service or not (Parent, Saint-Jacques, Beaudry, & Robitaille, 2007). Given that practitioners are more likely to have first contact with the mother, asking questions about the involvement of the child's father might also be avoided because of concerns about risking the relationship they are building with the mother.
Strategies to improve the engagement of fathers have focused on working face to face with them. Featherstone and Peckover (2007) argued that practitioners should assume a father's "desire for cooperation." In Australia, fathers have told researchers that they prefer practitioners to take a strengths-based approach (Berlyn, Wise, & Soriano, 2008), and solution-focused work has been claimed to be appreciated, although specific data on this are limited (Huebner, Werner, Hartwig, White, & Shewa, 2008). Identifying who the father is represents a challenge to practitioners, and Ferguson and Hogan (2004) recommended that referrals should always include the father's name and contact details. In a U.S. study, Malm, Murray, and Green (2006) went further and found that when there was no information about fathers, some practitioners searched for it through other available official records.
The importance of including fathers from the first point of contact was highlighted by Sandstrom et al. (2015) in a study of father engagement in home visiting programmes. One finding was that successful practitioners talked about the father in their first conversations with mothers, emphasizing that fathers would be "welcome to participate in home visits" (p. 37). This is supported by a study of fathers' experiences of the child protection system in which Brandon, Philip, and Clifton (2017) argued for the importance of practitioners developing "opening gambits" as a strategy to engage fathers. Decisions about engaging with services are sometimes made on the basis of a single conversation (Coulter, 2007), and this suggests that what happens during those initial conversations may have material consequences for the engagement of fathers in child welfare services. The current article reports on a study that investigated parental engagement during the initial telephone conversations made by practitioners to parents who had been referred for parenting support. The study approached the topic by making recordings of such initial conversations and analysing the occasions where the speakers talked about the other parent being included in the service.

| METHODOLOGY
Services in three local authorities in England agreed to take part in the study, identified because they delivered or coordinated parenting programmes in their local area. From these services, six practitioners made audio recordings of their initial telephone conversations with parents who had been referred to the service. Practitioners began the recording before the start of the call, explaining the study to parents after introducing themselves and asking for consent to continue recording. If the parent declined, the recording was ended immediately and subsequently deleted. At the end of the call, the practitioner asked the parent for consent again to see if the parent was still willing for the recording to be used. For further details of the ethical considerations, see Symonds (2018). The study received ethical approval from the University of Bristol Research Ethics Committee.
The practitioners who took part were all female and made between one and nine recordings each. In total, 31 recordings were successfully made and available for analysis, but in three of these, the speakers agreed that the service was not relevant to them (for example, because the parent was already attending a parenting programme). This article considers the conversations with the 28 remaining parents, 25 of whom were women and three men. Although no additional data were collected about participants, it was evident from the recordings that two of the men were in cohabiting relationships with a partner and one had separated (details about the women's living situations were too inconsistent to make comparable observations).
Analysis was conducted using conversation analysis, an approach that identifies the interactional resources that people use to pursue and achieve particular social actions (Schegloff, 2007). Interaction is recognized to be fundamental to practise in child welfare settings, whether by communicating child protection concerns (Forrester, Kershaw, Moss, & Hughes, 2008), engaging fathers in domestic violence interventions (Pfitzner, Humphreys, & Hegarty, 2017), or the formation of successful working relationships (De Boer & Coady, 2007). By taking into account the sequential production of interaction, conversation analysts have shown how it is possible to trace the emergence of particular social actions through the specific utterances of talk between the speakers (Schegloff, 2007). It is therefore an appropriate approach to take when considering interactional processes in child welfare practice such as the engagement of parents with parenting programmes.
Analysis began by transcribing the recordings verbatim and anonymising all names of people, places, and services. The transcripts were inspected to identify occasions when other parents were invited to be included in the service. These invitations were worded in a variety of designs, at different sequential locations in the conversations, and with different outcomes. To manage this variation, the entire dataset was analysed to trace how the speakers progressed the interaction turn by turn, over longer sequences of action to pursue common interactional tasks. This made it possible to map discrete phases of talk across the whole dataset and identified in Figure 1.
Having identified the overall structure of the conversations, the invitations (and any preceding talk about the other parent) could be located within particular phases of the conversation, providing the possibility of linking the relevance of that talk for the eventual outcome of the conversation. All sequences of talk that related to the other parent were then identified and transcribed in detail according to the conventions of conversation analysis. This form of transcription includes specific interactional features such as overlapping talk, hesitations, and pauses because they have been shown to be consequential for the speakers' understanding of their interaction. A full list of the meaning of transcription features can be found in Jefferson (2004).
An overall trajectory of engagement of the other parent was then traced over three interactional events: the first reference to the other parent; establishing the relevance of the other parent in the family; and inviting the other parent to be involved in the service. These are considered in the findings below.

Referring to the other parent
References to the other parent were made in 22 out of the 28 recordings, but they appeared in different places depending on who made them. When practitioners made the reference, it was usually after they had introduced themselves at the beginning of the call. An example is included in Extract 1, which occurred 59 s into the recording ("Wor" is used to denote "worker," and "Par" denotes the "parent"). This practitioner's use of Andrew's full name reveals that she has prior knowledge of him, presumably from the referral form (although these data were not collected). There is a slight delay before the parent responds, but when she does, she confirms the accuracy of the practitioner's information. Parents typically did this in other calls, and when this was achieved, the name of the other parent (in this case, Andrew) became a potential resource that could be used later in the call.
When parents made the first reference to the other parent, the references occurred in sequences of talk when the parent was describing their difficulties. Extract 2 is a typical example. FIGURE 1 Phases of initial calls to parents referred to a parenting programme (Symonds, 2015)  Parents only described their difficulties when they had been invited to do (Symonds, 2018), but when they were, parents faced the task of presenting their difficulties as serious enough to warrant support but also of presenting themselves as a responsible parent who was managing as well as could reasonably be expected. The parent in Extract 2 achieves this by positioning herself as more able than most to manage her son's difficulties (which were "always elsewhere"). It is only "now" that she is having trouble and needs support. The reference to "even his dad" supports her action of persuading the practitioner that her son's behaviour is serious enough to warrant professional support, supported by her use of "even" (positioning her son's father as especially able to cope) and the use of reported speech "he's hard work on the weekends," which claims access to the father's own perspective rather than simply asserting her own views (Holt, 1996).
The practitioner responds with "right, okay." This response aligns with the parent's overall action of justifying support. Although this parent has revealed the existence of the father in her son's life, the form of reference is "his dad" (in contrast to the practitioner's reference in Extract 1 to "Andrew Green"). This suggests that the parent does not expect the practitioner to know the father's name and is finding the most efficient means of referring to him so that the practitioner can identify who is being talked about (Sacks & Schegloff, 1979).
The mother is also modulating the relative distance between the people in the family by referring to the father's position in the family in relation to their son, rather than in relation to herself (in contrast with other calls, where references were made to "my partner" and "my husband"; see Enfield & Stivers, 2007). Combined with the reference to the father and son seeing each other at the "weekends," this suggests that the parents have separated. The practitioner does not pursue the topic of the father at this point in the conversation, which would have initiated a move away from the topic of her difficulties and possibly had consequences for their emerging relationship. The absence of the father's name means that the practitioner had more limited resources to refer to him later in the call. In fact, she did not return to the relevance of the father in the remainder of the call, and he was not invited to be involved in the service.
Although parents referred to other parents in seven of the calls, none of them did so to make them relevant to the service. In this dataset, it was only when practitioners took specific action that the other parent became relevant to the service and the means by which they achieved this was by establishing whether the other parent had an ongoing relationship in the family. directly "who's Andrew." The parent treats this question as seeking to confirm his relationship in the family and goes on to explain that he is "the dad." In making the reference to "the dad," the parent is also marking his parental relationship with his teenage child as most relevant, rather than his relationship with his partner (she later describes herself as "stepmum"). In another call, a practitioner sought confirmation that a referral was for the parent and "Lizzie your wife," which both introduced her name and proposed her position in the family. Even though this action might be very brief, it was important in these calls because it provided a basis for seeking the involvement of the other parent in the service.

Inviting the other parent
The third interactional feature was when practitioners solicited agreement that the other parent should be involved in the service.
These turns were designed in three different formats: yes-preferred questions with names; invitation to no specific person; and invitations to "partners," each of which will be considered here.
i. Yes-preferred questions with name In six recordings, the practitioner invited the other parent by using a closed question that required a yes/no response. In a study of doctorpatient interaction, Boyd and Heritage (2006) showed that doctors design their yes/no questions in ways that are "tilted" towards an anticipated response (for example, the question "is your father alive?" anticipates a confirmation). Pomerantz and Heritage (2013) argued that when respondents shape their answers to conform to the action that the questioner is proposing, their answers can be described as matching the "preference organization" of the question and are therefore "preferred." When practitioners in the current study had established the name of the other parent, they included names in their invitations and designed them in such "yes-preferred" formats, as in Extract 4. In this question, the practitioner includes the name of the other parent. Earlier in the call, Neil had picked up the receiver and had revealed his relationship in the family by suggesting that he pass the phone to "my wife." When the name of the other parent had been established earlier in the call, practitioners always referred to them directly in their invitations (rather than, for example, "your husband" or "his dad"). In fact, there was evidence that practitioners made particular efforts to use the person's name in these turns, even when their name had not already been established in the conversation. In two calls, the parent had discussed the situation in the family, but there was no evidence that the practitioner knew the name of the other parent. When it came to the point of inviting them, the practitioners introduced the name in the turn design, revealing their prior knowledge.
In Extract 4, the parent responds with "yes," but there is evidence that she is also aligning with what she understands the worker to be anticipating in her question. There is a gap of only 0.2 s between the end of the question and the start of the answer. When it is produced, the answer is in a format that conforms with the shape of the question and can therefore be described as "preferred." This is in contrast to "dispreferred" answers that are characterized by longer delays, delaying utterances such as hesitations "uhm," mitigations such as "well," or appreciations such as "that's awfully nice of you" (Pomerantz, 1984). Another example of an invitation that includes a name is given in Extract 5, which occurs after the parent and practitioner have been discussing the family's circumstances. The design of this question also includes the name of the other parent and is formatted to anticipate a "yes" response. There is a longer delay before the parent responds, which might relate to this being the first time his name has been mentioned, but when the response comes, the parent delivers a straight agreement "yeap" without any markers of dispreference.
Practitioners were not always successful when they used these turn designs. Extract 6 provides an example of a question being answered with a flat rejection. Earlier in this call, the practitioner had explained that the reason for the call was because she understood that the parent and "Owen Marsh" were interested in a "parenting group." At that point, the parent did not respond directly, pausing for 1.2 s before giving the downwardly intoned continuers, "right, yeah." At the beginning of the extract, the practitioner seeks to identify the problem in the interaction so that she can establish whether the parent and then Owen are still interested in the service. Given that the question about Owen in line 9 is tilted towards a "yes" response, the flat "no" in Line 11 strongly rejects the invitation.
Speakers commonly mark their dispreferred responses with markers of politeness, for example, providing an account for why they are declining, but the absence of such an account makes this rejection In both calls, the relevance of the other carer had not been established prior in the conversation. This supports the proposition that establishing whether the other parent is involvemed in the SYMONDS family is an important preparatory step before attempting to recruit them to the service.
ii. Invitations to no specific person formats. Both preface their question by stating their current understanding that it is "just yourself" or that the practitioner has "only got your name." Having done this, each practitioner enquires about another adult, but without a form of reference to use, there is no specific person to refer to. In Extract 7, it is "another adult," whereas in Extract 8, a range of possible options are described in the phrase "another parent, another carer, another partner." If the practitioner has no information about other parents, the approaches taken in these extracts display sensitivity to different family structures and could refer to step-parents, separated parents, or grandparents. However, the responses of each parent in these extracts orient towards the proposed existing circumstances, confirming that it is "just" me or "only myself." The absence of any markers of dispreference provides evidence that in discounting the involvement of another parent, the parent is aligning with what they understand the practitioner is proposing.
This alignment is further supported by the recycling of words "just" and "only" from the practitioner's turns, which work to strengthen the alignment between the two speakers.
Although this might be understood simply to reflect the reality of the parent's circumstances, there is evidence in some calls that suggest otherwise. The parent in Extract 8 later revealed the presence of the father in their daughter's life by explaining that he might be able to provide childcare while she attends the programme. There are two implications of this. The first is that parents designed their responses to conform to the preference design of the question (and confirmed what the practitioner seemed to be anticipating). As a consequence, when practitioners designed their questions in this way, they may inadvertently have led to the exclusion of other parents from services.
This analysis is further supported by Extract 9, in which the parent supports the involvement of the father but includes several features of a dispreferred response because the design of the question suggests the practitioner was anticipating that he would not be involved. In this extract, the practitioner's question includes "just yourself" and "anybody," both of which anticipate a "no" response. However, in this extract the parent does want the father to be involved and uses many features of dispreferred responses to propose this, delaying her response considerably by more than a second, prefacing her response with the hesitation "ehm," and introducing repeated mitigations of "might be" and "possibly" (Pomerantz, 1984). All of these features suggest that this parent is doing a considerable amount of additional interactional work to propose the father's involvement in spite of the format of the practitioner's question.
iii. Inviting a partner The third format that practitioners used in these data did include a reference to a specific person, that of a "partner." Like the invitations to no specific person, invitations to "partners" were used in interactional environments where there had been no prior reference to another parent and no relevance established of their involvement in the family. In these calls, the use of "partner" could refer to either a birth parent who was still living in the family or a new partner following the separation of the child's birth parents. In this way, the use of "partner" orients to the possibility of a parent being in a relationship (whether more longer term or more recently formed) but does not incorporate the possibilities of lone parenthood or parental separation. Given this uncertainty, practitioners could not design their turn as a yes-no question and had to introduce an element of conditionality, as in Extract 10. you've got a partner." This is followed by a declarative statement "they're welcome to come along as well" that proposes the involvement of the other parent without making it a requirement. Such a design does not constrain the range of relevant responses in the same way that a yes-no question does, and there is no requirement to make a firm commitment. The appreciation given by the parent, "at'd be good," is therefore able to conform to the format of the worker's turn without providing such a commitment (she goes on to explain how his "awkward shifts" might mitigate against his attendance). At this point, there is a risk that the parent's own commitment to the service might be understood as incomplete, and she moves to address this by emphasizing that she is "definitely interested." In a study of proposals in interaction, Houtkoop-Steenstra (1987) argued that proposals can be treated as fully accepted in the following turn if they relate to events in the immediate environment. For more "distant" proposals about action at some point in the future, proposals tend to be treated as fully accepted only when they are ratified by an additional confirmation. The invitations that targeted "partners" were designed as proposals about some future action, but in none of these calls was there an additional confirmation of their involvement, such as in the arrangements for future contact, resulting in uncertainty about whether or not the other parent would be included in the service.

| The relationship between interaction and outcomes
The analysis presented above identified three interactional tasks pursued by practitioners. The presence of these tasks was then compared with the outcome of each call, defined as whether or not the speakers agreed that the other parent should be involved in the service. The trajectories of different calls were then mapped on to a diagram (see  When these tasks were not completed or were out of sequence, as in Extract 6, invitations were less successful.

| DISCUSSION
The results of this study suggest, for the first time, that there may be a correlation between specific practitioner utterances and the recruitment of both parents to parenting programmes. This has particular relevance for the engagement of fathers who are under-represented in parenting programmes and shows how their involvement might be negotiated in the very first conversation with the mother. In doing so, this article contributes to evidence about good practice in engaging fathers with child welfare services.
The study is limited by the absence of conversations with child protection concerns or of practitioners successfully identifying a previously unknown father. Working successfully with involuntary clients, especially unidentified fathers, has long been recognized as a challenge for social work (Brandon et al., 2009), and further research would be required to identify successful ways of managing these interactions.
Because the data are based on recordings of practice, they have a very strong validity and had real-world consequences by the participants. Although data were not collected about which parents eventually attended the parenting programme, identifying small outcomes represents a first step towards achieving larger ones in the future.
Conversation analytic research is increasingly used to support practitioners to reflect on their professional interactions (Stokoe, 2014), and this article suggests that there may be scope for further application of this approach in social work.
The fact that invitations were attempted in only 14 calls will be a concern for those interested in whole family engagement. Developing tools based on conversation analytic findings of what actually happens in practice may be one way of improving the recruitment of both parents in the future. One tool developed from this study would recommend including the following tasks, in sequence.

Refer to the other parent by name
This was easiest when practitioners already knew the name of the other parent, such as though the referral information. If the name was not known, practitioners might enquire about home circumstances and pick up on subsequent references to the other parent (or other caregivers).

Establish the relevance of the other parent in the family
Practitioners only achieved this when they asked directly about it, but it was an important precursor to inviting the other parent. If the parents have separated, then respectfully asking about their ongoing relationship with their child might ensure that they are still included by the service in a child-centred way.
3. Invite the other parent by name When practitioners were able to invite other parents by name, and in "yes-preferred" formats, these were most likely to be accepted. This might be one example by which practitioners can demonstrate the recommendation to "assume cooperation" (Featherstone & Peckover, 2007).
This tool is inevitably incomplete but may be of use to practitioners committed to working with the whole family. Further studies could develop and evaluate its efficacy in practice as well as exploring other potential applications in other areas of social work. Because it was so relevant to have the name of the other parent, these findings also support recommendations to require these details to be included on referral information (Ferguson & Hogan, 2004;Malm et al., 2006).
The participants in this study reflect the fact that most referred parents (and practitioners) continue to be women. However, the interactional pattern identified in the data was consistent across all calls, regardless of whether the referred parent was male or female. This raises questions about the relationship between interaction, practice, and gender. As other studies have argued, engagement needs to be understood as more nuanced than simply viewing it as a problem located in fathers or practitioners (Brandon et al., 2017), although the circumstances of families continues to be gendered, for example, the influence of residency status in contact with child protection services (Laird, Morris, Archard, & Clawson, 2017). Further research would be needed to investigate other examples of practice at an interactional level, such as directly with fathers or with male practitioners.
In doing so, it may be possible to assess how interaction has a mediating influence on the ways that practice can address the gendered circumstances of both parents in order to improve outcomes for children and their families.