“We Were Totally Supportive,of Course”: How People Talk About Supporting a Significant Other to Stop or Reduce Their Drinking

Participants

For this study, we sought to recruit people with a partner, close friend, or family member who had recently stopped or substantially reduced their alcohol consumption. We initially recruited participants via a media release as well as snowball sampling. We contacted people who had previously participated in a study about stopping or reducing alcohol consumption (Bartram, Eliott, & Crabb, 2017) and asked if they would be willing to forward details about the current study to a partner, friend, or family member with whom they (previously) drank alcohol. We similarly approached other people who had stopped or reduced their alcohol consumption and expressed an interest in participating in future research. People expressing an interest in the study were provided with a plain-language information sheet containing further information about the study, and those meeting inclusion criteria were invited to an interview at a time and place convenient for them. Inclusion criteria for the study were that participants were aged 18 years or above, lived in Australia, spoke English, and had a partner, close friend, or family member who stopped or substantially reduced their alcohol consumption for at least 3 months within the past 2 years.

Our aim was to recruit for diversity with respect to gender and relationship to the person who stopped or reduced their drinking (partner, friend, or family). As our initial recruitment efforts yielded an insufficient number of male participants, we developed a flyer targeting male participants only. This flyer was posted around the researchers’ community, and we asked men’s health and alcohol-related organizations to circulate the flyer to their membership groups.

Procedure

This research was approved by the University of Adelaide Human Research Ethics Committee (Approval number H-2016-035). Ashlea conducted, audio-recorded, and transcribed all interviews. At the beginning of each interview, Ashlea reiterated study aims, emphasized that the interview was confidential and that no information would be shared with their significant other (or anyone else outside the research team), and obtained participants’ written consent. One interview was conducted by telephone; the remainder, face-to-face. Interviews were semistructured, with open-ended questions covering topics such as their experience of their significant other making a change to their drinking, their observations of any changes to the use of alcohol in shared social occasions, or to their own drinking, following their significant others’ change, and the role of significant others in providing support for people to stop or reduce their alcohol consumption. The interview schedule was structured to minimize leading questions and foreground participants’ experiences, and we made minor revisions to it following initial interviews to further improve this structure. Nonetheless, we acknowledge that the dynamic of interaction between interviewer and participant will have had some influence on the interview and, consistent with a social constructionist paradigm, consider the interview as a coconstruction between participant and interviewer, rather than a “true” account of the participant’s experience. At the end of the interview, participants completed a short demographic questionnaire. All names, places, and other potentially identifying information were changed during transcription. Transcripts were entered into the NVivo™ 11 software package (QSR International, 2015) for analysis.

Analysis

Drawing on discourse analysis and positioning theory, the aim of our analysis was to identify ways in which participants accounted for providing support to a significant other who stops or reduces their drinking as reasonable or otherwise, and how such accounts positioned the participant and their significant other. Given our social constructionist perspective, we were less interested in whether participants reported providing support to their significant others, but rather how they accounted for the provision of support as a reasonable thing to do, both in relation to their significant other and more generally for people trying to reduce their alcohol consumption. Our approach to analysis was also informed by the guidelines for thematic analysis proposed by Braun and Clarke (2006).

Ashlea initially read and reread the transcripts, then identified extracts of relevance to the research question, in which participants accounted for the provision, or nonprovision, of support. These included extracts that explicitly addressed support, as well as extracts where participants talked about actions that could be implicitly understood as supportive, such as modifying a social practice in response to the significant other’s change in drinking. Ashlea then systematically coded these extracts into categories of similar types of accounts. To collate these codes into broader themes, we considered how accounts within each code positioned both the participant and their significant other, and grouped together codes, which offered similar positions. Ashlea developed an initial thematic structure based on these groupings, and all authors then met to discuss this structure. In this meeting, the coauthors reviewed the extracts allocated to each code and theme to ensure their agreement with the coding and provide input into the overall structure. In reviewing the initial thematic structure, Jaklin noted a similarity between the themes and theories from moral philosophy. Thus, we reviewed the literature around these theories to name and further refine the themes.

The final thematic structure and related extracts were reviewed and discussed by all authors, to enhance the dependability of our findings by ensuring our interpretation was grounded in the data (Liamputtong, 2013). Nonetheless, we acknowledge this is our interpretation, and that others may be possible. To allow the reader to assess our interpretation, we have provided excerpts to illustrate each theme. Excerpts were selected as the most concise and/or archetypical example (Brinkmann & Kvale, 2015). To aid readability, we have added punctuation and removed speech elements such as false starts and repetitions (Bailey, 2008). After each excerpt, we have indicated the participant’s gender (M = male, F = female) and relationship to the person who changed their drinking, to provide additional context. As is common in the reporting of qualitative studies, we have presented our results and discussion concurrently (Liamputtong, 2013).

Participant Characteristics

Participants were 13 adults (10 female, three male) aged from 20 to 68 years (median = 35 years). For six participants (two male), the significant other who had changed their drinking was a romantic partner; for four participants (one male), a family member; and three participants (all female), a friend. All participants described themselves as of European/Australian or European ethnicity.

Thematic Structure

Through our analysis, we identified three types of accounts of supporting someone who stops or reduces their drinking: deontological accounts relating to duty; consequentialist accounts relating to a balance of costs and benefits; and relational accounts, which emphasized the importance of inclusion and shared experiences. These account types were prevalent across the data set, with no systematic differences apparent by gender or relationship type, with the exception of relational accounts, as we discuss below. Each of these types of accounts offered distinct positions to supporters and recipients of support.

Deontological Accounts

This first theme encompassed accounts in which participants justified supporting someone who stops or reduces their drinking with reference to a moral obligation: support was something significant others should provide. These accounts were deontological in character, in that they suggested that an action, such as providing support, must be performed as a matter of general duty or principle, regardless of the particular situation (Holland, 2007). The accounts invoked principles such as respect for autonomy, including a duty not to undermine others’ decisions, as well as a duty to assist others whose autonomy has been constrained, for example through illness or addiction. Consistent with a deontological focus on general duties, in these accounts, participants tended to talk in impersonal or universal terms. For example, when Ashlea asked one participant if she thought partners, friends, and family had a role to play if someone changes their alcohol consumption, she said “Absolutely” and went on to explain:

It’s about supporting people if they make a decision to change their lifestyle. And I guess particularly trying to you know make them feel comfortable and maybe not put it back in their face, or, not that but, or tempt them. . . . Particularly if they don’t mind drinking in the first place but have chosen not to then it is a particularly hard decision for people to make and if you’re surrounded by people who are drinking then that’s a lot harder on that person. (F, friend)

In this quote, the participant’s use of impersonal, universal language (“it’s about,” “people,” “they,” “you’re”) worked to convey that this duty to provide support was a general principle, not just a personal philosophy or in response to the particular needs of her own significant other. As she described it, this duty can be invoked when someone “make(s) a decision”: out of respect for their autonomy, one then has a duty not to interfere with that decision. The participant also described what support involved as part of this duty: “mak[ing] them feel comfortable” and not “tempt[ing] them.” These tasks emphasize that significant others have influence over each other’s drinking, and thus that a lack of support may actively undermine someone’s choice by making it “a lot harder.” This description of support resembles the notion of autonomy support, which involves creating an environment in which an individual’s autonomous pursuit of their goals is enabled, rather than thwarted (Powers et al., 2008). The provision of autonomy support by significant others has been found to facilitate behavior change; for example, perceiving family and friends as autonomy supportive of weight loss efforts was associated with significantly greater reported weight loss among U.S. college women (Powers et al., 2008).

Some participants acknowledged a general duty to support someone changing their drinking, but then worked to justify a lack of active support for their own significant other by positioning their significant other as particularly strong-willed and independent. An example came from a participant who continued to drink around her daughter after her daughter stopped:

Sometimes the other person I think has that influence as well. But no, it’s harder [to avoid alcohol when others are drinking], and this I admire [my daughter] for, because she said no when the rest of us were having a glass. I mean we were never drinking a lot but, a glass or two max probably, and she could say no and stick to that. And I think that was really good. (F, mother)

This participant, similar to the previous example, drew on abstract language (“the other person”) to acknowledge that significant others could “influence” or undermine another’s resolve not to drink, but then switched to personal language (“I,” “she”) to position her daughter as “admir[able]” for sticking to her decision not to drink even when the participant and the rest of her family were drinking. The description in turn positioned the participant and her family as having behaved appropriately, despite not having acted on the general obligation to offer support to someone whose behavior you have influence over: the responsibility was lessened because they “never dr[ank] a lot,” and support was unnecessary in this specific case because her daughter was able to “say no.”

This positioning of the participant’s daughter raises a note of caution with regard to accounting for support in terms of duty—if this person was admirable for changing her drinking without support, does that mean that those people who do receive support are less than admirable? Such an undesirable positioning was evident in accounts in which participants framed support as contingent on a significant other having a “problem” in relation to alcohol, such as addiction or another health condition exacerbated by alcohol. Consider, for example, the following exchange:

This participant had previously heard Ashlea speak on the radio about research with people who had stopped or significantly reduced their drinking, in which participants had reported feeling pressured by others to drink (Bartram, Eliott, & Crabb, 2017). Note that this research was not specifically with “problem” drinkers—most participants did not present themselves as having had an alcohol problem. Nor did Ashlea refer to someone with an “alcohol problem” in her question. Yet, the participant’s response presumed, at least initially, that if someone is stopping drinking, they must have an alcohol problem. Where this is so, support is then obligatory—the participant “really can’t understand” how someone could fail to offer this support, as someone with an alcohol problem “need[s]” others to be supportive, not obstructive. She then oriented to the possibility that some people stopping drinking may be “social,” not problem, drinkers, but suggested this is “different”—presumably, in this case, the same obligation does not apply. As in the previous excerpts illustrating this theme, the obligation was described in abstract language (“a person,” “they”), suggesting that it applies universally. Such an account thus effectively positioned anyone who receives support from significant others when stopping drinking as having an alcohol problem, inadvertently reinforcing some of the stigma attached to choosing not to drink in Australian or other Western countries. This stigma may relate to an assumption that nondrinkers are former problem drinkers (Bartram, Eliott, & Crabb, 2017; Posner, 1976; Romo et al., 2015), itself a highly stigmatized position (Room, 2005).

Overall, the accounts under this theme framed support in terms of universal obligations or duties invoked by the action of a significant other changing their drinking. Support was portrayed as one-directional, flowing from the supporter to the receiver, and thus positioned supporter and receiver asymmetrically. The needs or characteristics of the person who changed their drinking, the receiver, might modify the obligation to provide support, but no allowance was made for the supporter’s personal needs or desires, nor did it matter how providing support may affect them. As we discussed earlier, such asymmetric support provided out of a sense of duty or obligation can have negative implications for those involved, leaving the recipient feeling indebted or less competent, as well as affecting the provider (Rook et al., 2011), leading them to characterize the relationship as difficult (Offer & Fischer, 2018). For example, drug and alcohol professionals have identified the challenge of engaging family in the treatment of drug and alcohol abusers when these family members have become “hopeless” or “burned out” from their previous attempts to provide support (Misouridou & Papadatou, 2017).

Interestingly, this asymmetric type of account most commonly occurred in response to Ashlea asking a general question about whether significant others have a role to play when someone changes their drinking. Elsewhere in interviews, particularly where participants talked more specifically about their personal actions with regard to their significant other, accounts offering more symmetrical positioning were common, as we will discuss in the following themes.

Consequentialist Accounts

In contrast to the portrayal of support as a universal obligation, triggered by the actions or the requirements of the recipient, in this theme, the appropriateness of providing support was accounted for with reference to the impact on the supporter. These accounts were consequentialist accounts, in that the appropriateness of providing support derived from the consequences of that support—the end justified the means. Thus, this appropriateness could vary depending on the particular circumstance under consideration (Holland, 2007). Within this theme, we identified two subtypes of accounts: accounts in which participants emphasized positive or minimized negative impacts on themselves, and accounts in which participants framed impacts as balanced between themselves and their significant other.

Some participants portrayed their provision of support (for example, by modifying social occasions or cutting back their own drinking) as reasonable because it also brought benefits for them. For example, the following participant, who stopped drinking to support her husband, described how this had also improved her own health:

I haven’t done it because I’ve wanted to, I’ve done it because I’ve just seen the need [to support my husband], but I feel so much better for it. I mean health-wise, weight-wise, I’ve obviously not realizing, I’d put on a lot of weight through drinking, but I was on Diet Coke [and bourbon], it shouldn’t have been making me fat, but it’s just your whole, your eating habits change, everything changes. But not that I was a full on, I was, you know, just on the weekends and stuff, but yeah just a healthier feeling about yourself. (F, wife)

In this quote, the speaker framed her change in drinking as driven by support for her husband’s change, rather than being something she “wanted to” do herself. Nonetheless, she framed it as a positive change, not (just) because it had helped him (as the excerpts in the first theme oriented toward), but because it had helped her to “feel so much better” herself. By positioning herself as having gained health by stopping drinking, she risked also being positioned as someone who was unhealthy because they were a “full on,” or problem, drinker. However, she countered this by working to position herself as a responsible drinker, who drank “just on weekends” and used “Diet Coke” to reduce calories.

Not all participants identified personal gains from supporting their loved ones—given the abovementioned risk in positioning oneself as having benefited from reducing alcohol consumption, this is perhaps unsurprising. However, many worked to minimize their experience of negative impacts. In the following quote, one participant, whose friend stopped drinking while trying for a baby, and who avoided drinking in turn when socializing with this friend, talked about her initial response upon learning about this change:

I was like “Cool!,” excited for her I guess, that that’s why she’d chosen to stop drinking, and I don’t think, it wasn’t a big shock to me, and I guess we’d been through stages in our life when we were younger and did have lots of drinking times, partying and stuff, going out, and naturally those times had petered out a bit. And so when we did catch up it wasn’t like we were going out drinking a lot on a Saturday night anyway so it wasn’t a massive change to how we interacted and socialized. (F, friend)

Although this participant acknowledged that her friend stopping drinking led to changes in “how we interacted and socialized,” she minimized the significance of these changes by describing them as not “massive” or “big,” particularly in comparison with the reductions in “drinking times” from when the two were younger. In referring to this larger reduction as “natural,” the participant worked to counter a potential positioning of her friend as having forced a change in how they socialized, an asymmetrical positioning that would be inconsistent with norms of friendship, which emphasize symmetry, mutuality, and reciprocity (Hall, 2011). In contrast, a “natural” change is not instigated by either party—thus, both were positioned as equally affected by the passage of time. Other participants used similar strategies to minimize the impact of their significant other’s change; for example, attributing their own changes in drinking to factors other than their significant other, such as “doing more running” (M, husband); or framing the occurrence of negative impacts as rare, the “only time” (F, wife).

Other participants acknowledged that they were affected by their significant other’s change in drinking, but accounted for this as acceptable, as long as the impact was balanced between themselves and their significant other. Participants used terms such as “reciprocity,” “equilibrium,” “mutual respect,” and “fairness” to describe how they had worked to support their significant other’s change in drinking, and how their significant other in turn had supported them. Alternatively, they described a lack of such balance as leading to conflict. For example, one participant positioned herself throughout the interview as “accommodating” of her friend who had stopped drinking, by trying to come up with alternative ways for them to socialize and sometimes moderating her own drinking. When asked if there was anything that might stop someone from being accommodating, she said,

Maybe if they were really quite rude about it, or all superior and not trying to be accommodating to the fact that you’re going to want to drink as well. You want it to be a mutual respect of each other’s decision. I think if someone was really “Oh you shouldn’t drink because it’s going to kill you and it’s bad and you’re a bad person,” it’s like, “Well I don’t want to hang out with you, I’m going to go drink over here and not talk to you anymore.” (F, friend)

This participant portrayed accommodation as inappropriate where it was not “mutual.” This account positioned a significant other who does not show reciprocity in accommodation as someone who lacks “respect,” acts “superior,” and judges the other person as “bad” for drinking. Thus, this (hypothetical) nonreciprocating significant other was positioned as having violated key expectations of friendship (Hall, 2011), making it appropriate for one to elect not to “hang out with” the significant other, rather than provide support. In contrast, reciprocating significant others both work to find a way to respect or balance both parties’ needs and desires.

Unlike the accounts in the previous theme, these accounts positioned the participant and their significant other more symmetrically, either by framing any impacts as neutral or positive for the participant, or emphasizing that impacts were distributed evenly between the two parties. Here, the distinction between supporter and receiver was less relevant, as both parties provided and received respect from each other. As we have already noted, the concept of symmetry underpins key expectations of friendships (Hall, 2011). Symmetry is also considered important to family relationships; family therapies such as contextual therapy emphasize that justice, equity, and fairness in family relationships derives from a balance of give and take (Goff, 2001). Ogle, Park, Damhorst, and Bradley (2016, p. 183) described the importance of being able to both receive and provide support as “a fundamental feature of supportive transactions.” Indeed, a transaction is a good metaphor for the characterization of support under this theme, with this support incurring costs and benefits for each individual. However, some authors have criticized such transactional depictions of relationships as overemphasizing individuals and their rights (Goff, 2001).

Relational Accounts

In contrast, in the accounts captured in our third theme, the participant and significant other were positioned less as two individuals engaged in a transaction and more as parties in the same relationship. Supportive actions were those which maintain the relationship between the two, by ensuring that they continue to share experiences. These accounts were commonly characterized by the use of “we” language. Consider the following two exchanges:

In both examples, Ashlea asked about the actions of an individual person, the participant and the significant other respectively. Yet, both participants responded using first person plural forms (“Let’s do it,” “we don’t”), rejecting Ashlea’s implied positioning of participant and significant other as two individuals who might behave independently, instead positioning the participant and significant other as unified, a team who acts jointly—at least in relation to reducing drinking. Research on couple-based therapy for alcohol addiction has found that the use of such we-language is predictive of successful treatment outcomes, perhaps because it reflects a greater sense of togetherness and a communal orientation (Hallgren & McCrady, 2016; Rentscher, Soriano, Rohrbaugh, Shoham, & Mehl, 2017). Such an orientation might also be helpful for significant others working to change other health behaviors. For example, young adults attempting to lose weight identified coengagement as the most beneficial form of social support they received from their significant others (Faw, 2014); similarly, for women who had undergone bariatric surgery, significant others who served as “joint collaborators” were an important source of support (Ogle et al., 2016). Consistent with this, Romo and Dailey (2014) explored how weight loss can influence communication and behavior in romantic relationships in which only one member loses weight, finding that couples reporting positive consequences tended to position themselves as a team with equal power, whereas mixed consequences were more likely where responses reflected a vying for power or conflict.

In contrast to this communal orientation, participants also described times where they had failed to support their significant other, clearly portraying this as wrong because it positioned the significant other as excluded from the social group. For example, one participant described a particular party where she (and others) failed to include her friend:

Our whole group was there and she was there and she obviously wasn’t drinking and we all drank quite a bit and she just said she had a rotten time. We probably took it a bit far because it was free alcohol. We weren’t sort of, in hindsight probably should have been a bit more, I guess respectful of the fact that she wasn’t taking part in it. (F, friend)

In this excerpt, the participant’s friend was distinctly positioned outside of the “whole group” because she was not drinking: the participant described those drinking with the collective “we,” but her friend as “she.” By not drinking, the friend was positioned as not “taking part” in the collective activity, and consequently as having a “rotten time.” Drinking at similar levels can be important for relationship quality: Couples with discordant levels of drinking report higher levels of relationship dissatisfaction, which researchers have suggested may partly be because people who drink at similar levels take part in more shared activities (Birditt, Cranford, Manalel, & Antonucci, 2018; Homish & Leonard, 2007). Consistent with this, qualitative research has highlighted that nondrinkers tend to avoid social occasions where they do not feel included (Bartram, Eliott, & Crabb, 2017; Harris, 2010; Herring, Bayley, & Hurcombe, 2014). Here, the participant positioned the drinkers as having some responsibility to foster this inclusion; their failure to do so was not “respectful” of their friend and they “should” have behaved differently.

In addition to drinking at similar levels to their significant other, participants also talked about fostering inclusion in other ways. For example, one participant described supporting her husband by ensuring that their house was stocked with nonalcoholic drinks:

I would make sure that we, if there was alcohol in the house, that we had loads of nonalcoholic drinks, and that rather than him getting forgotten, because it’s almost like if you’re having some champagne everyone gets the alcohol and nobody gets a drink for the other person, like it’s not important. So just making sure that he was included with that too, and that maybe he got to drink something that was really nice, rather than just something boring. So I used to think a little bit ahead, and so that he didn’t feel left out. (F, wife)

The participant framed ensuring this availability as an “important” thing to do, as it meant her husband would be “included” in a drinking occasion, rather than “forgotten” or “left out.” Her use of “we” to describe the ownership of the drinks worked to characterize these drinks as belonging to the two of them, rather than being a special accommodation required for her husband. To be inclusive, the drinks also needed to be “really nice,” not “boring.” This is consistent with research indicating that replacements for alcohol are more readily accepted in a social group where they carry a similar meaning (Bartram, Eliott, Hanson-Easey, & Crabb, 2017). Champagne is traditionally drunk on special occasions, or to celebrate; thus, by ensuring her husband could have a similarly special drink, this speaker was able to present herself as supporting not just his change in drinking but his continued inclusion in shared activities.

Overall, accounts in this theme positioned participant and significant other not just symmetrically, but as part of a partnership or social group. Support was framed as appropriate because it worked to maintain or reaffirm that group membership. Thus, these accounts reflected ethical assumptions consistent with a feminist ethics of care, which views “doing good” as acting in a way that works to value and maintain personal relationships (Rachels & Rachels, 2012). Note that, unlike the previous account types, we identified examples of relational accounts only among partners and friends, not family members. This absence might reflect the less voluntary nature of family ties: as these ties are defined by kinship and are not easily dissolved (Offer & Fischer, 2018), participants may have felt less need to actively assert their shared group membership to Ashlea, treating it as “taken-for-granted.”

These relational accounts also reflected a more collectivist conception of self than in the previous two themes: a self that is fundamentally connected and defined through relationships. This is in contrast to an individualist self who is defined through the pursuit of individual goals, and has responsibility to support the pursuit of other’s goals (Taylor et al., 2004). Western countries such as Australia are usually understood as individualist cultures, and this has been reflected in health promotion campaigns around alcohol, which have been criticized for positioning people in highly individualistic terms (Cherrier & Gurrieri, 2013). The use of collectivist, as well as individualistic, conceptions of self by participants in this study suggests there may be value in considering how health promotion might also draw on these alternative conceptions of self, particularly where seeking to engage with social groups.