The impact of describing someone as being in recovery from alcohol problems on the general public’s beliefs about their life, use of treatment, and drinking status

Abstract Purpose: The general public’s attitudes toward former heavy drinkers can impact on the wellbeing of these individuals. The current study sought to determine if describing a former heavy drinker as ‘in recovery,’ and varying the amount they drank, impacts the general public’s perceptions of how the person is functioning (both personally and as a member of society), their need for treatment, and the possibility of a moderate drinking recovery. Materials and methods: An online panel survey (n = 4450; adults from multiple countries) asked participants to read a brief vignette describing a former heavy drinker (i.e. John). Participants were randomized to receive a vignette in which John was described as ‘in recovery’ (vs. no mention of recovery) and as having consumed heavy (vs. very heavy) amounts of alcohol prior to seeking help. Participants were then asked to rate John on how he is functioning, and to also rate the possibility of his recovery with or without treatment and as abstinent or a current moderate drinker. Results and Conclusions: Participants who read the vignette in which John was described as being in recovery rated him as being more likely to be functioning well compared to those where no mention of ‘recovery’ was made. However, this manipulation did not impact ratings regarding the likelihood of untreated and moderate drinking recoveries. Varying the amount of drinking described did not impact ratings of how John was functioning but very heavy (compared to heavy) drinking reduced ratings of the likelihood of untreated and moderate drinking recoveries.


Introduction
The term, 'in recovery,' is sometimes adopted by people who previously consumed alcohol in a hazardous fashion and have now stopped (or reduced) their drinking (White 2007;Kelly et al. 2018). The phrase is closely associated with sobriety and the 12-step movement (Betty Ford Institute Consensus 2007). An online panel survey examining the prevalence of individuals who identify as being in recovery in the USA found that roughly 60% of those who endorsed having had a problem with alcohol identified themselves as currently being in recovery or as having been in recovery in the past (Kelly et al. 2018).
In other population health and epidemiology research, the focus has been largely placed on patterns of recovery, where findings demonstrate that there are both abstinent and moderate drinking remissions from alcohol dependence (Dawson 1996;Dawson et al. 2005). It is unclear whether the terms remission and recovery should be used interchangeably. Alternatively, natural history research largely targets the in-depth exploration of people who have recovered from alcohol or drug problems (Klingemann et al. 2001). This research emphasizes that there are multiple pathways to recovery, both abstinent and moderate use recoveries, as well as self-guided versus treatment-led Klingemann et al. 2010). The natural history tradition also emphasizes the importance of social capital as part of the recovery process Cloud 1996, 2001). These concepts are also present, in combination with sobriety, in the 'new recovery' initiative (Groshkova et al. 2013; Thomas et al. 2019). Finally, recovery-oriented policy work is informing the structure of treatment provision and can, in turn, be co-opted as a means to justify reductions in government spending (Humphreys and Lembke 2014;Best et al. 2017).
Another area of research relevant to the study of the 'in recovery' term is the examination of the general public's attitudes and beliefs. There is a long tradition of this type of work in the addictions, with one focus examining the impact of labels such as alcoholism, or disease on the general public's attitudes toward the individual described (Sobell and Sobell 1975;Wallston et al. 1976;McKirnan 1977;Stafford and Petway 1977;Dean and Poremba 1983;Cash et al. 1984;Tootle 1987;Kelly and Westerhoff 2010). More recent work in the area focuses on the impact of terms such as, 'chronically relapsing brain disease' (Kelly et al. 2021) or compares ratings of stigma resulting from describing someone as having an alcohol use disorder versus a mental health disorder (Pescosolido et al. 2010;Schomerus et al. 2011;Rundle et al. 2021).
There appears to be no similar work, to-date, examining the general public's attitudes to people described as being in recovery. The term appears widely known among the general public; perhaps because of its depiction in the media (Room 1989;Sulkunen 2007;Baker 2016). Nonetheless, it remains a relevant topic as it might inform the potential experience of someone who regards themselves as in recovery when interacting with others in their everyday life. The general public's reaction also has implications on the delivery of services, and treatment funding decisions by governments and policy makers. The hypotheses guiding this research are tentative as the authors can find no existing literature directly related to this topic. However, given that the academic literature generally describes the recovery identity as a helpful goal (White 2007;Kelly et al. 2018), we assume that the general public will also view those identifying as being in recovery positively. Thus, it is predicted that people in the general public will be more likely to endorse that a former hazardous drinker will be functioning well, both personally and as a member of society, when that person is described as being in recovery compared to when a former hazardous drinker has not been labeled as in recovery. Further, it is also predicted that former hazardous drinkers described as having been a very heavy drinker (compared to a less heavy drinker) will receive more negative ratings regarding their recovery.

Procedures and participants
An online panel survey was conducted using the Prolific website. All registrants on the website were shown a study advertisement (i.e. no restrictions were placed on geographical location or other participant characteristics; Prolific restricts registration to those 18 years or older). The advertisement asked people to take a short survey (approximately 15 minutes) about drinking alcohol, regardless of their current or past drinking behavior, to receive an honorarium of £2.70 (UK pounds sterling). Potential participants who clicked on the link to the survey were taken to an electronic consent form and then, once consenting to participate, were provided the survey to complete.
The survey started by asking participants for their country of residence (United Kingdom, Canada, USA, or Other; a required text field was provided for those who checked 'other'). Participants were then asked to read a brief vignette ) and to answer some questions about the person described: 'John Smith used to drink a lot. On weeknights, he usually had from 5-10 beers. On weekends, John had from 10-15 beers a day. About five years ago, John sought help for his drinking and is in recovery. John is the manager of a small department store. He is married with two children, aged six and ten.' Four versions of the vignette were assigned at random in a 2 by 2 designthe italicized text (not italicized for participants) indicates sections that were varied: 1) the volume of alcohol consumed (higher quantity described John Smith as drinking 10-15 drinks on weeknights and 20-25 drinks a day on weekends); and 2) the statement that John is 'in recovery' was included or excluded.

Measures
After reading the vignette, participants were asked a series of questions about John Smith. First, participants were asked a modified version of the brief addiction recovery questionnaire (Iverson-Brown and Raistrich 2016). This questionnaire asked, in the last month, how often the participant thought that John had a reasonable quality of life (e.g. money to live reasonably well, suitable accommodation), was likely avoid risky situations (e.g. is abstinent, mixing with people who are problem drinkers), and had a positive attitude to life. The response options were 'not at all,' 'rarely,' 'often,' and 'all the time.' The items were modified to ask participants' ratings about John, and to refer solely to alcohol consumption (instead of alcohol or drug use). Due to a transcription error by the first author, one item from the original 12 item survey was not included in the version used in this study. The remaining 11 items were combined to form a composite score (possible score range: 0-33). Given the content of the questionnaire, we are interpreting this modified recovery questionnaire as a measure of the extent to which the participant believes John Smith is functioning well, both personally and as a member of society. Finally, two other questions were asked about John Smith: 1) 'Do you think it is possible for someone like John to fix their drinking problem on their own, without getting any treatment?' (Yes, No, Don't know); and 2) 'Do you think it is possible for someone like John to reduce their drinking to that of a social drinker or do they need to quit all together?' (Response options: 'I think John can reduce his drinking to that of a social drinker,' 'John needs to quit,' Don't know) (Cunningham et al. 2007).
Next, participants were asked about their own alcohol use. Those who said that they drank in the last year were asked the Alcohol Use Disorders Identification Test (AUDIT) (Babor et al. 1989;Saunders et al. 1993) with expanded categories for the quantity and frequency of drinking items taken from the Alcohol Toolkit Study, a series of general population surveys conducted in the United Kingdom (Beard et al. 2015). Participants from the United Kingdom were asked about the frequency of consuming 6 or more drinks on one occasion in the third AUDIT question, while all other participants were asked about the frequency of consuming 5 or more drinks. A standard drink chart was provided, with participants from the United Kingdom having a standard drink defined as 8 grams of alcohol and participants from all other countries being provided with a standard drinking chart reflecting the approximate quantity of alcohol in a standard drink in the USA and Canada (13.6À14 grams of alcohol) (Kalinowski and Humphreys 2016).
The survey continued with questions about alcohol consumption and recovery among the subsample of former heavy drinkers identified from the survey. Results from these items are reported in a separate manuscript (Cunningham and Godinho 2021). At the end of the survey, participants were asked a series of demographic characteristics. Finally, participants were asked to rate their agreement with whether they answered all questions truthfully (7-point scale, with 1 ¼ 'strongly disagree' and 7 ¼ 'strongly agree') (Godinho et al. 2016). In addition, the survey contained an attention check question (nested in the recovery questions at the beginning of the survey), 'I want to indicate that I have read this question by checking all the time?' Only participants who answered the attention check question correctly and rated 'strongly agree' on the question about how truthfully they answered the question were included in the analyses (all participants were paid regardless of their responses to these items).

Results
Of the 5,002 participants who completed the survey, 4,450 correctly answered the attention check question and strongly agreed to answering truthfully. Of these 4,450 participants, 13.4% (n ¼ 597) reported that they did not drink alcohol in the past year. A total of 22.5% (n ¼ 1,002) scored eight or more on the AUDIT, indicating current unhealthy drinking. Table 1 provides a summary of the demographic characteristics. Notably, only 45% of the current sample said that they lived in the UK or the USA (in contrast, the Prolific website reports that approximately 70% of registered participants currently live in one of these two countriesfor the remaining sample, 29 different countries were reported with samples sizes ranging from one to 643). Also notable, almost a third of participants (32%) listed student as their employment status (again, in contrast to the Prolific website which indicated a lower proportion of studentsalthough the method of categorization was different; other commonly endorsed employment categories: 40.1% full-time employed; 14% part-time employed).

Ratings on the recovery questionnaire
A 2 by 2 Analysis of Variance (ANOVA) compared participants' ratings of the vignette character, John Smith, on the modified recovery questionnaire between the experimental conditionsvolume of drinking (heavy versus very heavy) and recovery label (described as being in recovery versus no mention of recovery). A main effect was found of recovery label (F(1, 4446) ¼ 155.40, p < .001, d ¼ 0.07) with ratings on the modified recovery questionnaire being higher when John was described as being in recovery compared to when recovery was not mentioned in the vignette (M ¼ 20.6, SD ¼ 5.8 versus M ¼ 18.4, SD ¼ 6.2, respectively). There was no significant main effect (p ¼ .14) between volume of drinking conditions, and no significant interaction (p ¼ .99) between volume of drinking and recovery label, on ratings on the modified recovery questionnaire. The 11-item version of the questionnaire employed in these analyses had a Cronbach's alpha of .88. It is also possible to subdivide the modified recovery questionnaire into three subscalesabstinence, normality, and positivity. However, as these subscales are all highly correlated with scores on the full questionnaire (r ! .87 for all subscales) and between the subscales themselves (r ranged from .66 to .69), separate analyses for each subscale did not appear merited. Finally, a hierarchical logistic regression was conducted to predict participants' responses on the one item in the modified recovery questionnaire directly asking about frequency of John being abstinent ( … 'been abstinent from alcohol'; recoded as 'all the time' versus all others) between participants in the volume of drinking conditions, and between those in the recovery label conditions. The main effects of drinking volume (very heavy versus heavy) and recovery label (labeled as in recovery versus no mention of recovery) were entered in Step 1. The interaction term between drinking volume and recovery label was entered in Step 1 (see Table 2). There was no significant relationship between the volume of drinking conditions and ratings of likelihood to be abstinent from alcohol (p ¼ .57). However, there was a significant relationship between the recovery label condition and ratings of the likelihood of John currently being abstinent, with those participants being told that John was currently in recovery rating him as more likely to be abstinent compared to those where being in recovery was not mentioned (p < .001; EXP(b) ¼ 1.965 [95% CI 1.708, 2.261]; summary from cross tabulation resultabstinent all the time ¼ 30.7%; other ¼ 18.4%). Further, there was no significant interaction between the volume of drinking and recovery label conditions (p ¼ .09) on ratings of the likelihood that John was abstinent from alcohol.

Ratings on likelihood of recovery without treatment
Participants were also asked whether they thought it was possible for someone like John to fix their drinking problem on their own, without getting any treatment (Yes ¼ 42.6%; No ¼ 43.1%; Don't know ¼ 14.3%). As with the question regarding whether John was currently abstinent, a hierarchical logistic regression was conducted to predict agreement that an untreated recovery was possible (yes versus all others; i.e. no and don't know combined) between participants in the volume of drinking conditions, and between those in the recovery label conditions see Table 2). When John's drinking was described as at a heavier volume before dealing with his problem, participants were less likely to believe that he could recover without treatment compared to when John's drinking was described as comparatively less heavy (p < .001; EXP(b) ¼ 0.782 [95% CI 0.694, 0.881]; summary from cross tabulation resultself-change possible: 39.6% versus 45.6%). There was no significant relationship between the recovery label conditions and ratings of the possibility of recovery without treatment (p ¼ .51). Further, there was no significant interaction between the volume of drinking and recovery label conditions (p ¼ .15) on ratings of the possibility of self-change.
Finally, participants were asked if they thought it is possible for someone like John to reduce their drinking to that of a social drinker or if they need to quit altogether (can reduce drinking to that of a social drinker ¼ 42.9%; needs to quit ¼ 51.7%; Don't know ¼ 5.4%; note: needs to quit combined with don't know for the following analyses). As with the previous two questions, a hierarchical logistic regression indicated that the amount John was described as drinking prior to dealing with his problem had an impact on ratings of the possibility of being able to be a social drinker, with higher volumes leading to ratings of lower likelihood of success at social drinking compared to lower volumes (p < .001; EXP(b) ¼ 0.809 [95% CI 0.718, 0.911]; summary from cross tabulation resultssocial drinking possible: 40.3% versus 45.5%). There was no significant main effect of the recovery label on ratings of the possibility of social drinking after John dealt with his problem (p ¼ .46). Further, there was no significant interaction between volume of drinking and recovery label conditions (p ¼ .75) on ratings of the possibility of being a social drinker.

Discussion
Describing a former hazardous drinker as being in recovery resulted in a small positive increase in participants' estimates on a modified recovery questionnaire which measured the extent to which a person is believed to be functioning well, both personally and as a member of society. Examination of the item specifically addressing being abstinent on the modified questionnaire also found that labeling the person as in recovery resulted in a higher proportion of participants estimating that the person was not drinking any alcohol at all, compared to when participants were presented with the same vignette without the person being described as in recovery. Given the close association of the 12-step movement with the recovery concept, and with how recovery is depicted in the media, this belief is understandable (Betty Ford Institute Consensus 2007; Baker 2016). Thus, alongside terms such as 'alcoholism' or 'chronically relapsing brain disorder' (Kelly et al. 2021), the term, 'in recovery,' also appears to have an impact on the general public's attitudes toward someone who used to drink heavily.
There did not appear to be any impact of varying the amount a person used to drink in the vignette on ratings on the recovery questionnaire. This is possibly because the amount described was substantial in both conditions, resulting in a ceiling effect on participants' ratings. Varying the quantity of drinking in the vignette did, however, impact on ratings of the possibility of the person being able to deal with their alcohol consumption by themselves (without seeking treatment) and on the possibility that they could now drink in a social manner (as opposed to needing to quit altogether). Previous research has indicated that people are generally skeptical of untreated recoveries and moderate drinking recoveries (Cunningham et al. 1998), despite clear evidence that both are prevalent among former heavy drinkers (Dawson 1996;Dawson et al. 2005). In the current study, describing the person as a heavier drinker led to a lower proportion of participants believing that untreated, or social drinking recoveries were possible, compared to when the person's drinking was described as less heavy (but still substantial). There was no significant association (p > .05) between describing the person as in recovery (versus no mention of recovery) on ratings of the need for treatment and of the possibility of a social drinking recovery. However, there is some ambiguity in interpreting the lack of an observed relationship between the recovery label and beliefs about the need for treatment. Specifically, with the close association of the recovery term with the 12-step movement, and the ongoing lack of clarity as to whether a program like Alcoholics Anonymous (AA) is a treatment or a self-help program, some participants might assume that treatment is not needed because the person attended AA. Disentangling these relationships is merited but would require further research on this topic.

Limitations
There are a number of limitations with the study, including that the modified recovery questionnaire employed in this study was missing one item from the original and that several of the outcomes of interest were measured with single items (belief that John Smith is abstinent, need for treatment, need for abstinence). Further, it is unclear the extent to which participants' rating of a brief vignette would generalize to how they might react to someone in-person. There are many people who regard themselves as having resolved a drinking problem who do not tell others about this experience, primarily because of concerns about stigma (Cunningham et al. 1998). Thus, it is unclear exactly how likely it is that someone ever hears about people who are in recovery (beyond what they see in the media). Nevertheless, it does appear that the recovery term has some sort of positive association, even if it is a small one. Important to recognize, the sample size for this study was chosen to allow for the recruitment of sufficient numbers of people who had reduced or quit their heavy drinking (results not reported here), resulting in a large sample where small differences on the recovery questionnaire appear quite significant. We had also hoped to be able to recruit a sample that was stratified to have its demographic characteristics similar to that of the general population of the UK and the US. However, this was not possible because the sample size we needed was too large for the Prolific website. The resultant sample was unrestricted as participants could come from any country, resulting in some strengths and weaknesses. The strength is the diversity of the sample. One weakness is the apparent overrepresentation of students in the sample and the relatively young age of participants. Another weakness is that the standard drink size defined in the study was restricted to either be relevant to someone from the UK, or to those from the USA or Canada (and a few other countries with similar standard drink definitions). In addition, it is unclear exactly how much participants thought the person in the vignette was drinking, given that, for example, 10 standard drinks contains a different amount of absolute alcohol in different countries (assuming the participant is thinking about a standard drink for their respective country) (Kalinowski and Humphreys 2016). Lastly, it is unclear whether the same results would have been observed if the individual in the vignette had been described as a woman, so the results are not generalizable to how 'recovered' females may be perceived by the general public.

Conclusion
This exploratory study indicates that people who used to drink heavily and who are now described as being in recovery receive more positive ratings on a scale measuring how they are functioning, both personally and as a member of society, compared to those where being in recovery is not mentioned. Further, when someone is described as in recovery, they appear more likely to be assumed to be abstinent than those not described as in recovery. While the results of this paper alone cannot recommend whether clinicians and policy makers should adopt the term 'in recovery' when treatment planning, these preliminary findings suggest clinicians should be aware that adopting this term will have an impact on how the general public views their patients. As evidence suggests that individuals who have not sought formal or informal treatment for alcohol use rarely identify with the term 'in-recovery' (Cunningham and Godinho 2021;Kelly et al. 2018), and cite stigma and public attitudes as barriers to seeking treatment , it becomes further unclear whether the term 'recovery' should be used. While the general public may view this term more positively, it may conversely dissuade individuals from seeking treatment. Future quantitative and qualitative research on this topic would benefit from the recruitment of samples that are more clearly representative of the general population of interest in order to replicate these findings and to gain a clearer picture of the influence of the recovery label on both those potentially seeking treatment and peoples' judgements about those who used to drink alcohol in a hazardous fashion.