Non-alcoholic beverage consumption among US adults who consume alcohol

Abstract

Background and Aims:

Non-alcoholic beverages (NABs) that mimic alcohol without inducing intoxication, such as non-alcoholic beers, non-alcoholic wines and spirit-free drinks, are increasing in popularity. It is unknown whether NABs help to mitigate or stimulate alcohol use. The present study aimed to describe NAB consumption practices among US adults consuming alcohol, characterize who is likely to consume NABs and examine whether NAB use influences desire for and perceived consumption of alcohol.

Design and Participants:

The survey study used data collected June–July 2023 from an on-line convenience sample. The first survey (n = 1906) assessed frequency of NAB consumption among US adults who consume alcohol. A second more detailed survey on use patterns was conducted with 466 respondents who reported past-year NAB consumption, of whom 153 (32.83%) screened positive on the CAGE questionnaire for alcohol use disorder (AUD).

Setting:

This study took place in the United States.

Measurements:

NAB consumption measures included type of NAB consumed, frequency, quantity, first consumption age, consumption reasons, consumption contexts and perceived effect on desire for and consumption of alcohol. Alcohol use measures included frequency, quantity and first consumption age.

Findings:

Past-year NAB use was endorsed by 28.44% of respondents (61.70% ever used). Non-alcoholic liquor/’mocktails’ were the most common NAB type consumed (83.69%). Compared with respondents without AUD, those who screened positive for AUD were significantly more likely to consume NABs in an effort to decrease or abstain from drinking alcohol [adjusted odds ratio (AOR) = 3.54, 95% confidence interval (CI) = 2.24–5.58] and 67.97% endorsed less alcohol consumption (3.23% endorsed more) due to their NAB use. NAB consumption frequency and quantity were significantly positively predicted by alcohol consumption frequency (AOR = 1.46, 95% CI = 1.17–1.83) and quantity (β = 0.25, 95% CI = 0.15–0.35), respectively.

Conclusion:

Adults who consume alcohol and screen positive for alcohol use disorder report drinking non-alcoholic beverages as a harm reduction strategy.

INTRODUCTION

Alcohol misuse can result in significant harm [1]. Appealing non-psychoactive beverages consumed instead of alcohol could have significant population health effects. Non-alcoholic beverages (NABs) are those that seek to mimic alcohol without inducing intoxication. The Food and Drug Administration defines NABs as having an alcohol by volume < 0.5% [2]. Common NABs in the United States include non-alcoholic beers, wines, liquor and ‘mocktails’ (mixed beverages imitating cocktails, without alcohol).¹

NABs have recently surged in availability and popularity [3], with US sales increasing 15–30% yearly since 2018 [4]. A 2022–30 World Health Organization action plan recommended that the alcohol beverage industry replace full-strength alcohol with NABs (and low-alcohol products) whenever possible to mitigate harmful alcohol use [5]. However, a more recent report underlined the dearth of knowledge on NABs’ public health effects and called for additional research to inform policies [3].

While extant research is scarce, market research and focus groups highlight a variety of reasons for NAB consumption, including health, taste, curiosity, sobriety and alcohol reduction; such data have not differentiated NAB use reasons based on alcohol use history [6, 7]. Household data from Great Britain and Finland reveal that NAB purchases are associated with being male, higher education, higher income and purchasing full-strength alcohol [8, 9]. Household data from Spain highlight that 12% of all beer purchases and 4% of all wine purchases were non-alcoholic, and these purchases were associated with subsequent reduction in full-strength alcohol purchases [10]. A recent UK field trial found that regular-strength beer sales decreased when pubs added a non-alcoholic beer option (with no loss of profits) [11]. However, household purchase data are limited by not capturing all purchases a household makes and, as with the field trial, not directly assessing who is consuming NABs or for what purpose.

An on-line study found the odds of selecting an alcohol-free drink (NABs and soft drinks combined) increased when there was a greater proportion of alcohol-free options [12]. A recent randomized controlled study found that social drinkers who received free NABs across 3 months had greater reduction in alcohol consumption than control group participants (who did not receive NABs) [13]. However, NAB use has also been associated with increased alcohol use and future intentions to consume alcohol [9, 14]. Competing theories offer differing predictions about whether and for whom NABs are likely to reduce or potentiate alcohol use.

Theoretical considerations

Regular alcohol use has been conceptualized as a habit influenced by cues that can reliably prompt alcohol consumption [15–17]. Habit theory identifies substitution of a new response to cues as one way to interrupt undesired habits [18]. According to habit theory, NABs may help people to transition away from alcohol by serving as a plausible substitute and protective behavioral strategy for reducing use [19].

Behavioral theories of choice offer predictions about whether individuals will consume NABs in place of alcohol. The contextualized reinforcer pathology model posits that the availability of alternative reinforcers—reinforcing activities that do not involve substances—is critical to whether or not people decide to consume substances (see [20]). NABs may be an alternative reinforcer that reduce alcohol use, if accessible, and if their consumption is sufficiently rewarding in the short- (e.g. social acceptance) and/or long-term (e.g. no hangover).

While NABs may be used as a substitute or alternative reinforcer to alcohol when exposed to drinking cues, NABs may also be drinking cues themselves. The incentive salience sensitization theory of addiction poses that, in response to reward cues, the brain’s mesolimbic dopamine systems facilitate wanting for a substance and urge to obtain it [21, 22]. This theory would predict that NABs, which contain numerous alcohol-related cues (e.g. appearance, taste, sipping behavior), may prompt craving and alcohol-seeking, perhaps especially among people with a significant alcohol use history [23].

The present study

Data on US NAB consumers have relied upon market research and alcohol industry surveys [4, 24]. Given the high rates of alcohol use and alcohol use disorder (AUD) within the United States, formal research is needed to advance insights on NAB consumption practices and their relation to alcohol use to guide public health recommendations and inform clinical practice guidelines for treating AUD. Critically lacking are guidelines as to whether NABs should be encouraged as a harm reduction strategy.

The present study aimed to provide descriptive data on NAB use among US alcohol consumers and identify demographic and alcohol use characteristics associated with NAB consumption. While this study was not designed to establish or disprove either potentially applicable theory, we sought to examine whether NAB use influenced consumers’ desire for and perceived consumption of alcohol to guide future theoretically oriented inquiry. Novel in assessing specific outcomes related to NAB practices, tests of associations were exploratory. This study was pre-registered prior to data collection (https://osf.io/a7gc6).

METHODS

Sample

Prolific, an on-line research platform, was used to reach a convenience sample of n = 2500 registrants. Registrants aged 18–89 years, consuming alcohol and residing in the United States were invited to complete an on-line survey on drink consumption. Survey 1 was used to identify a subsample of past-year NAB consumers, who were invited to participate in survey 2 (n = 466)—both surveys were hosted on Qualtrics.

Procedure

Data collection occurred during June–July 2023. Prolific provided interested users with a link to survey 1. After viewing a research information sheet, respondents continued with the survey (written consent was not obtained due to minimal risk). Survey 1 assessed past-year beverage consumption. Respondents received 20 cents for completing this 1-minute survey. Survey 1 respondents confirming past-year alcohol and NAB consumption were invited to survey 2, which assessed more demographic characteristics and in-depth detail on NAB and alcohol consumption. Respondents received $1 for completing this 5-minute survey.

Measures

Survey 1

Respondents reported how recently they had consumed the following beverages: (1) kombucha; (2) seltzer (non-alcoholic); (3) soda/pop; (4) non-alcoholic beer, non-alcoholic wine or non-alcoholic liquor/’mocktail’; and (5) alcohol. Response options were: no, never; yes, more than a year ago; or yes, within the past year. Beverages other than NABs were assessed to reduce the likelihood of respondents discerning the screener purpose and providing biased responses.

Survey 2

Respondents reported demographic characteristics, NAB use behaviors² and alcohol use history (see Tables 1 and 2). A nascent area of research, novel items were created to assess NAB use. NAB use frequency/quantity were modeled after AUDIT-C items [25]. Reasons for drinking NABs were developed through co-author discussion, based on relevant items from the Reasons for Not Drinking Scale [26], knowledge from clinical practice and recommendations from people who consume NABs. We piloted the survey with six colleagues for feedback, which yielded 19 response options. An ‘Other’ textbox captured 14 free-responses that were either already covered by the multiple-choice reasons (e.g. taste) or endorsed by no more than two respondents (e.g. receiving an NAB for free)—we did not analyze these data.

TABLE 1.

Survey 2 non-alcoholic beverage use measures.

Item	Item wording	Response options
Frequency	How often do you drink the following non-alcoholic (NA) drinks? (asked separately for NA beer, wine and liquor/ ‘mocktail’)	Never; monthly or less; 2–4 times a month; 2–3 times a week; 4 or more times a week
Quantity	How many NA drinks do you have on a typical day when you are drinking NA beverages?	Range from ‘1’ to ‘10 or more’.
Age of first use	How old were you when you first drank an NA drink?	Free text numerical entry
Reasonsa,b	Why do you drink NA drinks?	See Table 5 for response options
Timea,b	What time of day do you drink NA drinks?	Morning, afternoon, evening
Social settinga,b	In what social settings do you drink NA drinks?	Alone; around other people drinking alcohol; around other people drinking NA beverages
Locationa,b	Where do you drink NA drinks?	My home; cannabis lounge/club; work; bar or brewery; friend’s or family member’s home; restaurant; other
Alcohol desireb	When I drink NA drinks, I experience…	Less desire to drink alcohol; more desire to drink alcohol; no change in my desire to drink alcohol
Alcohol useb	Because I drink NA drinks, I drink…	more alcohol than I otherwise would; the same amount of alcohol that I otherwise would; less alcohol than I otherwise would

Abbreviations: AUD = alcohol use disorder; NA = non-alcoholic; NAB = non-alcoholic beverages.

^aCheck all that apply;

^bOrder of options was randomized.

TABLE 2.

Participant demographic characteristics (survey 2; n = 466).

Item	n (%)
Race or ethnicity
Hispanic/Latinx White	25 (5.36)
Multi-racial/other	29 (6.22)
Non-Hispanic Asian/Native Hawaiian/Pacific Islander	74 (15.88)
Non-Hispanic Black	51 (10.94)
Non-Hispanic White	287 (61.59)
Gender
Man	240 (51.50)
Non-binary	7 (1.50)
Prefer to self-describe	1 (0.21)
Woman	218 (46.78)
Sexual orientation
Asexual	7 (1.50)
Bisexual	57 (12.23)
Gay or lesbian	25 (5.36)
Heterosexual	360 (77.25)
Pansexual	5 (1.07)
Queer	8 (1.72)
Questioning	3 (0.64)
Prefer to self-describe	1 (0.21)
Education
Less than high school	1 (0.21)
High school graduate or GED	20 (4.29)
Some college, no degree	71 (15.24)
Associate degree or vocational program	39 (8.37)
Bachelor’s degree	224 (48.07)
Master’s degree	85 (18.24)
Professional doctorate degree	26 (5.58)
Age	Mean = 36.62
	SD = 11.76
	Range: 18–85

Abbreviations: GED = general educational development; SD = standard deviation.

The AUDIT-C measured alcohol consumption frequency, quantity and severity. The four-item CAGE questionnaire screened for AUD [27]. Responses of ‘yes’ are summed, with higher scores indicating alcohol problems; a score of two or more is considered clinically significant. Scale reliability in this study was acceptable (α = 0.72) [28]. Participants also reported age of first alcohol use.

Analyses

Logistic and linear regression analyses for binary and continuous outcomes tested associations between demographic variables and alcohol use characteristics with NAB consumption. All demographic and alcohol use variables were entered simultaneously into the models. Regarding NAB use reasons, we examined associations with the top five reasons endorsed.

We combined responses for some racial–ethnic subgroups (Table 2). Gender and sexual orientation were combined to create three subgroups: (1) sexual or gender minority (SGM; 22.75%), (2) heterosexual woman (34.76%) and (3) heterosexual man (42.49%). Highest education was dichotomized to compare those with and without a college degree.

To enhance the stringency of these exploratory analyses, significance cut-offs were set at P < 0.01. Given relevant theories predicting substitution mechanisms and differences in NAB consumption by AUD status, we conducted a post-hoc interaction test of AUD status by NAB quantity and frequency (respectively) predicting alcohol use severity (three-item total score of the AUDIT-C) [25].

RESULTS

Survey 1: past-year NAB use among adults who drink alcohol

Of those invited to survey 1, 1906 provided eligible responses (Figure 1). Mean age was 39.38 years [standard deviation (SD) = 13.34; range = 18–89]. Most respondents endorsed having ever consumed NABs (61.70%), with 28.44% (n = 542) reporting past-year use. Past-year NAB consumers were significantly younger (mean = 35.95, SD = 11.81) than those who had not consumed NABs within the past year (mean = 40.19, SD = 13.72; t₍₁₉₀₄₎ = −6.32, P < 0.001), although the effect size was small (d = 0.32; 95% CI = 0.22,0.42) [29].

FIGURE 1.

Participant recruitment and inclusion for surveys 1 and 2. NABs, non-alcoholic beverages.

Survey 2: descriptive characteristics of adults who drink alcohol and report past-year NAB consumption

Figure 1 shows recruitment flow to the final sample of 466 participants. Table 2 summarizes sample socio-demographic characteristics and Table 3 presents alcohol and NAB use characteristics. Sample representation of White and Black participants was comparable to US demographics; there was a larger proportion of Asian and Pacific Islander (API) participants and a smaller proportion of Hispanic participants [30]. Educational level and proportion of SGM respondents was higher than national rates, which is expected, as alcohol use is more common among these groups [30–33].

TABLE 3.

Participant beverage use characteristics (survey 2; n = 466).

Item	n (%)	n (%)
Frequency of use	Alcohol	NABs
Monthly or less	145 (31.12)	239 (51.29)
2–4 times per month	160 (34.33)	141 (30.26)
2–3 times per week	118 (25.32)	74 (15.88)
4 or more times per week	43 (9.23)	12 (2.58)
Quantity per occasion
One drink	133 (28.54)	215 (46.14)
Two drinks	155 (33.26)	177 (37.98)
Three drinks	84 (18.03)	48 (10.30)
Four drinks	51 (10.94)	21 (4.51)
Five drinks	18 (3.86)	1 (0.21)
Six drinks	12 (2.58)	3 (0.64)
Seven drinks	2 (0.43)	–
Eight drinks	3 (0.64)	–
Nine drinks	1 (0.21)	–
Ten or more drinks	7 (1.50)	–
Age of first use	Mean = 17.93 (SD = 3.92)	Mean = 23.38 (SD = 9.18)
Frequency of heavy alcohol use
Never	214 (45.92)	–
Less than monthly	140 (30.04)	–
Monthly	67 (14.38)	–
Weekly	36 (7.73)	–
Daily or almost daily	9 (1.93)	–
Alcohol use disorder
Diagnosed or suspected
Yes, currently	12 (2.58)	–
Yes, in the past but not currently	33 (7.08)	–
No	421 (90.34)	–
CAGE screener
Clinically significant (total score ≥ 2)	153 (32.83)	–
Not clinically significant (total score < 2)	313 (67.17)	–

n = 18 participants endorsed being younger than 10 years of age when they first consumed NABs. We treated these responses as missing data due to suspicion of misreporting; other data from these respondents were retained.

Abbreviations: NABs = non-alcoholic beverages; SD = standard deviation.

NAB consumption among adults who consume alcohol and NABs

Nearly half (48.71%) of the sample consumed NABs more than monthly, with non-alcoholic liquor or ‘mocktails’ most commonly reported (83.69%), followed by non-alcoholic beer (65.67%) and non-alcoholic wine (48.50%) (Figure 2). All significant associations with NAB use outcomes are reported in Table 4 (see Supporting Information Tables S1–S21 for full model results). All tested variables had some significant associations, with eight for racial–ethnic identity (four Black, two Hispanic, two API), five for AUD positive screening, four for age, three for education, three for alcohol quantity, two for alcohol frequency and one for SGM.

FIGURE 2.

Frequency of non-alcoholic beverage consumption by type (survey 2).

TABLE 4.

Significant associations with non-alcoholic beverage (NAB)-related measures at P < 0.01.

Outcome	Predictor	B	SE	Std Coef.	95% CI
Greater than monthly NAB use	Frequency of alcohol use	0.38	0.11	AOR = 1.46	1.17, 1.83
Quantity of NAB use	Hispanic (ref: White)	0.53	0.18	β = 0.13	0.04, 0.22
	Quantity of alcohol use	0.14	0.03	β = 0.25	0.15, 0.35
Age of first NAB use	Age	0.33	0.04	β = 0.43	0.34, 0.51
	Black (ref: White)	−5.48	1.28	β = −0.18	−0.27, −0.10
	Education: college degree or higher (ref: no college degree)	−2.37	0.86	β = −0.12	−0.20, −0.03
Number of reasons for NAB use	Black (ref: White)	−0.89	0.32	β = −0.13	−0.22, −0.04
	AUD screen: positive (ref: negative)	0.65	0.22	β = 0.14	0.05, 0.24
NAB use reason: avoid the negative effects of alcohol	API (ref: White)	−0.90	0.30	AOR = 0.41	0.23, 0.73
	Black (ref: White)	−1.04	0.35	AOR = 0.35	0.18, 0.70
	Quantity of alcohol use	−0.23	0.07	AOR = 0.80	0.69, 0.92
NAB use reason: health	Age	0.03	0.01	AOR = 1.03	1.01, 1.04
	AUD screen: positive (ref: negative)	0.76	0.23	AOR = 2.13	1.36, 3.33
NAB use reason: curiosity	Hispanic (ref: White)	1.41	0.46	AOR = 4.09	1.67, 10.00
	Frequency of alcohol use	0.35	0.12	AOR = 1.42	1.13, 1.78
NAB use reason: taste	Quantity of alcohol use	−0.22	0.08	AOR = 0.81	0.69, 0.94
NAB use reason: decrease alcohol use	AUD screen: positive (ref: negative)	1.26	0.23	AOR = 3.54	2.24, 5.58
NAB use time: afternoon	API (ref: White)	−0.80	0.30	AOR = 0.45	0.25, 0.80
NAB use social context: others drinking alcohol	Black (ref: White)	−0.97	0.32	AOR = 0.38	0.20, 0.72
NAB use social context: others drinking NABs	Age	−0.02	0.01	AOR = 0.98	0.96, 0.99
NAB use location: home	Age	0.03	0.01	AOR = 1.03	1.01, 1.05
NAB use location: restaurant	Heterosexual man (ref: SGM)	−0.79	0.26	AOR = 0.46	0.27, 0.76
NAB use location: bar or brewery	Education: college degree or higher (ref: no college degree)	0.62	0.23	AOR = 1.86	1.17, 2.94
NAB use and desire for alcohol	Education: college degree or higher (ref: no college degree)	−0.16	0.06	β = −0.14	−0.23, −0.04
	AUD screen: positive (ref: negative)	−0.19	0.06	β = −0.17	−0.27, −0.07
NAB use and alcohol use	AUD screen: positive (ref: negative)	−0.27	0.06	β = −0.23	−0.33, −0.14

All models included the following covariates: age, race/ethnicity, gender/sexual orientation, education level, alcohol use frequency, alcohol use quantity, screen for AUD.

Abbreviations: AOR = adjusted odds ratio; API = Asian/Native Hawaiian/Pacific Islander; AUD = alcohol use disorder; B = unstandardized coefficient; NAB = non-alcoholic beverages; SE = standard error; SGM, = sexual or gender minority; Std coef. = standardized coefficient; β, standardized coefficient.

Participants reported consuming an average of 1.76 (SD = 0.91) NABs on a typical day that they drink NABs. NAB consumption frequency and quantity were each significantly associated with alcohol use frequency and quantity. Upon inspection, people who consumed alcohol monthly or less were less likely to report monthly NAB use (35.17%) compared to people who consumed alcohol multiple times per month (54.83%).

Reasons for consuming NABs

On average, participants endorsed three to four reasons for consuming NABs (SD = 2.13) (Table 5). Participants who screened positive for AUD endorsed significantly more reasons for consuming NABs [positive screen: mean = 3.99 (SD = 2.06); negative screen: mean = 3.47 (SD = 2.15)], and Black participants (mean = 2.90, SD = 2.18) endorsed significantly fewer reasons than White participants (mean = 3.78, SD = 2.18).

TABLE 5.

Reasons for consuming non-alcoholic beverages (survey 2, n = 466).

Response	n (% yes)
I do not want the negative effects of alcohol (e.g. hangover)	213 (45.71)
Better for my health	180 (38.63)
Curiosity	174 (37.34)
I like the taste of NA beverages	173 (37.12)
I am trying to decrease or abstain from drinking alcohol	154 (33.05)
Driving	143 (30.69)
Time of day—for example, it is too early in the day to drink alcohol	123 (26.39)
It helps me blend in socially—it makes it look like I’m drinking alcohol even though I’m not	123 (26.39)
Cost—it is cheaper than beverages containing alcohol	70 (15.02)
Caretaking responsibilities	63 (13.52)
People I know drink NA beverages	58 (12.45)
I’ve seen them promoted/advertised	58 (12.45)
I’m working or on call for work	54 (11.59)
I do not want to mix substances (e.g. cannabis and alcohol)	27 (5.79)
I do not have access to beverages containing alcohol	24 (5.15)
I cannot drink alcohol for medical reasons	21 (4.51)
Pregnant, trying to get pregnant or breastfeeding	18 (3.86)
Other	14 (3.00)
Religious reasons	8 (1.72)
Mean number of reasons endorsed	3.64 (SD = 2.13)

Abbreviations: NA = non-alcoholic; SD = standard deviation.

Contexts of consuming NABs

Most respondents consumed NABs in the evening (88.19%), 51.72% in the afternoon and 9.01% in the morning (due to the low endorsement rate for morning consumption, tests of associations were not conducted). NABs were most commonly consumed in social settings where others were drinking alcohol (73.61%) and where others were drinking NABs (66.31%); 39.91% consumed NABs when alone. Regarding location of consuming NABs, home was most common (66.95%), followed by at a friend’s or family member’s home (59.44%), restaurant (50.43%) and bar or brewery (36.05%). Fewer reported drinking NABs at work (8.37%), at a cannabis lounge or club (2.58%) or ‘other’ (2.58%) (due to the low endorsement rates for these items, tests of associations were not conducted).

NAB consumption association with desire for and consumption of alcohol

Most reported (69.10%) that when they consume NABs they experience no change in desire to drink alcohol, 24.46% reported experiencing less desire to drink alcohol and 6.44% reported experiencing more desire to drink alcohol. More than half (51.29%) reported that because they consume NABs they drink less alcohol than they otherwise would, 45.92% reported consuming the same amount of alcohol that they otherwise would, and 2.79% reported consuming more alcohol. Screening positive for AUD was negatively associated with reporting changes in both desire for and perceived consumption of alcohol as a function of NAB use, with 36.60% of this group endorsing less desire for alcohol (9.8% endorsed more desire) and 67.97% endorsing less alcohol consumption (3.23% endorsed more alcohol consumption). Participants with a college degree, compared to those without a college degree, were more likely to endorse less desire for alcohol (26.87 versus 18.32%) and less likely to endorse greater desire for alcohol (4.78 versus 10.79%) when drinking NABs.

Post-hoc analysis of AUD status and NAB frequency and quantity predicting AUDIT-C

That greater frequency and quantity of NAB use was associated with greater frequency and quantity of alcohol use seemed incongruent with the finding that people with AUD reported believing they drink less alcohol because of their NAB use. Thus, we ran an exploratory analysis testing whether screening positive for AUD interacted with NAB use frequency and quantity, respectively, in predicting alcohol use severity, controlling for the same demographic variables as in the primary analyses. Screening positive for AUD [β = 2.52, P < 0.001, B = 0.34 (95% CI = 0.26,0.42)] and NAB quantity [β = 0.61, P < 0.001, β = 0.16 (95% CI = 0.08,0.25)] were positively associated with alcohol use severity. The CAGE × NAB quantity interaction approached significance [β = −0.59, P = 0.01, β = −0.11 (95% CI = −0.20,−0.02)], suggesting that alcohol use severity may be stable across NAB quantity among people with AUD, whereas alcohol use severity may increase across NAB quantity among social drinkers (Supporting information, Figure S1). NAB frequency was not a significant predictor.

DISCUSSION

Among surveyed adult US residents who currently drink alcohol, more than a quarter consumed NABs in the past year and most did so at least monthly. Consistent with market research [4], non-alcoholic liquor/’mocktails’ were the most commonly reported NAB type consumed. A key goal was to explore how NAB consumption practices differ across subgroups defined by demographic and alcohol use characteristics—differences are outlined below.

Associations between NAB consumption practices and demographic and alcohol use characteristics

Educational level and age

Educational level, which can be a socio-economic status proxy, was associated with measures of NAB use. Compared to those without a college degree, individuals with a college degree started drinking NABs at an earlier age, and were more likely to consume NABs in a bar or brewery and to report NABs reduce their desire to drink alcohol. Future research with store observations of NAB marketing and sales split by geographic socio-economic indicators would offer relevant insights into whether differences in exposure contribute to socio-economic differences in age of first use.

We observed age differences in reasons for consuming NABs, settings in which NABs are typically consumed, age of first NAB use and NAB use frequency. Younger participants were more likely to consume NABs in social settings where others are drinking NABs, suggesting that NAB consumption may be more socially accepted among younger generations. However, it is also possible that older participants engage in less social drinking contexts [34]. While age of first NAB use was not associated with alcohol consumption, longitudinal studies of adolescents would afford greater sensitivity to elucidate the effect of NAB consumption—and NAB marketing exposure—on alcohol use outcomes and would help to guide marketing and sales regulations.

Racial–ethnic and SGM identities

There were multiple differences in NAB consumption practices based on racial–ethnic identity in terms of reason for using, including White participants having greater odds of using NABs to avoid the negative effects of alcohol (compared to Black and API participants) and to try to decrease or abstain from alcohol (compared to API participants). Prior work has demonstrated differences in alcohol drinking motives, use of protective behavioral strategies and drinking practices based on race [35]. Exploration of how people are socialized to NABs would help to elucidate why NABs may be used as a protective behavioral strategy by some, but not all.

In addition to differences in reasons for consuming NABs, compared to White participants, Black participants were more likely to start consuming NABs at a younger age and Hispanic participants reported greater NAB consumption quantity per occasion. Given that exposure to alcohol advertising is higher among Black and Hispanic (compared to White) youth [36], future research on marketing exposure and access to NABs will help to elucidate possible mechanisms of racial–ethnic differences in NAB consumption. A limitation is that heterogeneity within the multi-racial group may have precluded observations of differences based on this grouping.

Similarly, heterogeneity within our SGM grouping may be why we only observed one difference based on SGM identity, which related to NAB consumption location. Because SGM individuals are at greater risk for alcohol misuse and related harms [37, 38], and prior work has observed differences based on intersections of minoritized sexual and gender identities [33], future work that is powered to examine differences in the relation between NAB and alcohol use among SGM individuals (rather than solely between SGM and heterosexual cisgender men and women) is needed to clarify NAB’s protective versus harmful effects within this population.

Alcohol use history

Quantity and frequency of alcohol use were positively associated with quantity and frequency of NAB use. Research is needed to clarify whether heavy drinkers alternate beverage type within or across drinking occasions, as the former may be particularly useful in mitigating heavy drinking episodes. Frequency of alcohol use was positively associated with endorsing curiosity as a reason, perhaps indicating that individuals who drink alcohol regularly enjoy trying new alcohol-related products, such as NABs. Quantity of alcohol use was negatively associated with endorsing taste and not wanting the negative effects of alcohol as reasons for consuming NABs. The latter may be due to participants who consume greater quantities of alcohol having less sensitivity to negative alcohol effects [39].

Participants who screened positive for having AUD endorsed more reasons for consuming NABs, and specifically were more likely to endorse health and trying to decrease or abstain from alcohol. This underlines the notion that people at risk of AUD leverage NABs as a harm reduction tool. We observed that more than a third of people who screened positive for AUD endorsed less desire for alcohol because of their NAB use and, consistent with predictions from habit theory and the contextualized reinforcer pathology model, more than two-thirds reported that their NAB use leads them to consume less alcohol. Of note, there was minimal support for the incentive salience theory of addiction, with a small proportion of participants endorsing increased desire for and consumption of alcohol due to their NAB use.

Although people screening positive for AUD typically endorsed using NABs in an effort to drink less alcohol and perceived that they drink less alcohol because of their NAB use, in a model predicting risky drinking (AUDIT-C), a significant main effect was found for NAB quantity. NAB quantity is associated with greater levels of risky drinking even when accounting for AUD. The non-significant trend of an interaction between NAB quantity and AUD indicated that, if anything, there may be a positive relation between NAB quantity and alcohol use severity among social drinkers. These findings are based upon self-reported perceptions of NAB and alcohol use, and further evaluation with real-time, objective observations are needed.

We focused the present study upon current alcohol consumers, due to prior work indicating the majority of people who consume NABs also consume alcohol [40]. However, classical conditioning theory would predict the utility of NAB use among people initiating alcohol abstinence. Repeated pairing of the unconditioned stimulus (US) of alcohol ingestion with alcohol-related cues leads alcohol cues to become conditioned stimuli (CS) that elicit the conditioned response (CR) of alcohol craving and subsequent alcohol-seeking [41]. When a CS is repeatedly encountered without the US, the CR is expected to extinguish [42]. Because NABs necessarily contain alcohol cues, people who consume NABs without alternating with alcohol will encounter the CS without the US. Thus, classical conditioning theory would predict that NAB consumption in the context of abstinence from alcohol should facilitate the extinction of cue-elicited craving and support sustained abstinence.

Limitations

While Prolific permitted nation-wide recruitment, it may be subject to typical limitations of convenience sampling (e.g. selection bias, skewed towards women and individuals with higher educational level) [43]. This study probably underestimates NAB use, as we excluded individuals who do not consume alcohol. Our measure of alcohol use quantity, frequency and severity may have limited precision as we did not specify a past-year time-frame, which the AUDIT-C was originally designed to include [25]. Future studies should assess current health conditions, as health was one of the top NAB use reasons endorsed.

CONCLUSION

The present study contributes to a small, but growing, literature on NABs. Findings offer detailed exploration of NAB consumption practices and reasons for consuming, as well as differences across consumer subgroups. The finding that most people who screened positive for AUD believed they consume less alcohol because of their NAB use is encouraging, and suggests that NABs may prove to be a harm reduction tool for some. However, there appeared to be no difference in reported alcohol use severity in relation to NAB quantity among people screening positive for AUD. In order to inform clinical guidelines for working with people with AUD and precision health approaches [44] for intervening based on demographic differences in NAB consumption, additional research with more objective, real-time assessment is needed to confidently establish the effect of NAB on alcohol use.

DATA AVAILABILITY STATEMENT

Supporting information, including data and syntax, are openly available at: https://osf.io/se279/.