An exploratory pilot study to assess drinking at bars or events located within grocery stores

Abstract

Objective

Grocery stores are creating opportunities, such as a separate bar area and including beer and wine tasting events, to create a unique experience that caters to particular groups of clientele to encourage drinking. The goals of the study were to determine whether assortative drinking (i.e. the process of drinking alcohol in places where individuals have similar characteristics) is occurring at grocery stores, assess drink pacing (e.g. drinks per hour), and observe whether grocery stores engaged in responsible beverage service practices during special events and at their bars.

Method

We conducted unobtrusive observations at four grocery stores in Central Ohio to understand who attended special events and/or drinks at the bars located within grocery stores. Demographic characteristics and drinking quantity of the 96 patrons were recorded. Data were analyzed with bivariate statistics.

Major findings

Patrons drank, on average, 3.8 drinks per hour, although standard drink size could not be determined. The locations showed evidence of assortative drinking at the individual level by age and consumption of food. At the establishment level, assortative drinking appears to have occurred by gender, age, and race/ethnicity. Drinks per person per hour differed by location, type of drink, and presence of food.

Conclusions

Our work suggests a need to better understand these emerging alcohol establishments, which may create more opportunities to drink while bringing in new or different clientele to drink alcohol. The effects of these locations on alcohol-related problems are an important next step in understanding the full impact of drinking in these locations.

Introduction

Grocery stores are one type of retail establishment that has more recently begun serving alcohol on premises. Whole Foods opened their first bar located within their store in 2009 (Sugar 2018). Although no listing appears to exist of all grocery stores that serve alcohol, articles in the press detail the following brands serve alcohol, at least at some locations, such as Kroger Marketplace, Whole Foods, Publix, and Albertsons (Sugar 2018; Rubak 2023). Some grocery store chains have created atmospheres that encourage drinking alcohol while shopping and/or provide opportunities for drinking and socializing (called ‘sip and shop’ or ‘sip and stroll’; Sugar 2018). In fact, events that pair drinking alcohol with shopping are encouraged for places looking to develop a wine culture (Taylor et al. 2024). Yet, the published research literature does not include any studies that specifically look at drinking behaviors in these locations, although they may be serving a specialty niche where alcohol-related problems may be less understood. Drinking in grocery stores may be a phenomenon similar to how some restaurants morph into bars later in the evening where it caters to a different clientele (Lee et al., 2018). Acting as a bar may present additional challenges as Halonen et al. (2013) found that individuals living less than one kilometer to a bar are more likely to experience an extreme drinking occasion and heavy alcohol use.

In addition, drinking in grocery stores in particular may appeal to the changing demographics of alcohol use. In the United States, women make up 68% of household grocery shoppers (FMI 2023). Drinking behaviors in these nontraditional venues is important as some of the events are held regularly and could lead to heavy alcohol use. For example, having a flight of wine paired with appetizers might encourage patrons to drink more than just one glass with limited food. If patrons attending and consuming alcohol at these events could drink a significant quantity (e.g. four glasses of wine) in a short amount of time, this could lead to alcohol-related problems such as driving after drinking. Further, those pouring and serving alcohol may not be trained in procedures that limit service to intoxicated patrons or what standard drink sizes are for these ad-hoc events.

Niche marketing refers to selective advertisements in an attempt to entice a particular type of clientele to buy a product. With regards to alcohol, niche marketing could lead to assortative drinking behaviors where specific types of clientele are attracted to specific venues (Gruenewald 2007). Manufacturers and establishments selling alcohol look to differentiate their offerings to target a distinct population. In areas with higher densities of alcohol establishments, they may have to work harder to differentiate themselves and thus focus on a particular type of consumer, developing a ‘niche’ in the alcohol market. This was tested in an agent-based model, finding that as more bars moved into an area, competition for clients increased and those individuals who drank in the same bar shared more characteristics than those who drank in different bars (e.g. assortative drinking; Fitzpatrick and Martinez 2012). At the establishment level for cannabis, Cooke et al. (2018) found patient populations at medical cannabis dispensaries showed evidence of market segmentation, such that the characteristics of patients at the dispensaries were more similar within dispensaries than across dispensaries, which included the medical reason for obtaining a medical cannabis card. Further, some dispensaries catered to clientele that differed demographically from the areas within which they were located, an indication of segmentation as would be seen for assortative use. This has not been assessed for drinking establishments.

At the individual-level, assortative drinking suggests that individuals go to those establishments that cater to their preferences. As an example, segmentation among individuals who drink wine appears to be a function of sex, level of wine-drinking experience (e.g. beginner) and the taste of the beverage (Pickering et al. 2014), while men and women drink wine for different reasons and at different places (Thach 2012). Women who are novice at drinking wine are more likely to prefer sweet wine and wine-based beverages than both men and women who have more wine expertise (Pickering et al. 2014). A higher percentage of women drink wine for social reasons and to relax at home; whereas men are more likely to drink wine because they like the taste and it enhances food (Thach 2012). In Italy, the profile of people categorized as ‘beer/winer lovers’ were older (ages 46–60 years) compared to ‘spirit lovers’ and ‘mild-drink lovers’ (Cravero et al. 2020). As alcohol establishments become aware of these differences among the population, they may cater to particular groups of clientele to encourage drinking. If assortative drinking is occurring, people with like characteristics will congregate at specific places (e.g. bars) that may or may not represent the demographics of the individuals who live in the area where the place is located (i.e. is the bar a separate destination for individuals looking to drink with people who have shared characteristics?)

Alcohol manufacturers and establishments engage in niche marketing to entice individuals to purchase more alcohol. For example, marketing alcohol to women has changed (Atkinson et al. 2022) resulting in new brands directed specifically to subsets of women (c.f. ‘Mom Water’). Messages around using alcohol to cope in the US and internationally intensified during the COVID-19 pandemic with #sendwine memes, #winemom on social media, and references to alcohol as ‘mother’s helper’ (Basch et al. 2021; Harding et al. 2021; Newman and Nelson 2021). Women’s drinking increased from 2002 to 2017 (Hasin et al. 2019) and during COVID-19 (Barbosa et al. 2021), along with higher alcohol-related mortality from 1999 to 2017 (White et al. 2020). These increases have occurred primarily among White and more highly educated women (Keating 2016). In addition to changes in drinking among women, rates of binge drinking in the United States increased for all age groups except for 18 to 29 year olds from 2000 to 2015 (Grucza et al. 2018). Older adults (50+ years old) have also seen increases in alcohol use, including binge drinking over the past two decades, largely fueled by the aging Baby Boomer generation (Kuerbis 2020; Kepner et al. 2023). These changes could be related to different types of alcohol being consumed, as there has been an increase in wine consumption during the late 2010’s (Castellini and Samoggia 2018; Fogarty & Voon, 2018), with wine volume at its highest in in the US in 2020 (Kerr et al. 2024). These changes in the drinking landscape could offer new opportunities for alcohol establishments that seek patronage from older adults, women, and mothers.

We conducted an observational exploratory pilot study to assess the drinking pace at grocery stores, including special drinking events, and to preliminarily assess whether assortative drinking is occurring at grocery stores. The goals of the study were to (i) provided demographic characteristics of those who we observed attending special events and/or drinking alcohol at bars in grocery stores; (ii) examine if demographic characteristics differ between stores and from neighborhood characteristics where the store is located, as an initial test of assortative drinking; (iii) assess how much patrons drank per hour; and (iv) observe whether establishments engaged in harm reduction/responsible beverage service (RBS) practices.

Materials and methods

Study design

This observational exploratory descriptive pilot study monitored the consumption of alcoholic beverages during special events with paired alcohol and food held by a convenience sample of a grocery store chain and in their internal bars located in Central Ohio. One member of the research team scouted eight locations of a regional grocery store chain in Central Ohio known to have bars located within the store or had special alcohol events to ascertain whether a specific location had either or both of these features (see Table S1). This brand was chosen because they were known to have bars and/or special drinking events. Out of the eight locations of the chain that had bars or special alcohol drinking events, observers went to four locations for observations. Three of these locations were visited two times, with Location 3 only being visited once for a total of seven visits. We only include Location 3 once as the special event we visited had limited attendance. These locations were chosen based on proximity to where the observers lived. Table S2 shows the Census tract characteristics of all eight locations and denotes which were included in the sample, which also allows for an assessment of possible assertive drinking based on the sociodemographics of individuals living in the vicinity of the store. Scouting notes from each location that provide more context on the bar or special events can be found in Table S3.

At some locations, the event was integrated into the bar area, whereas others had a separate area for the event. During these events, patrons were offered four alcoholic beverages paired with four appetizers. Beer and wine pairings were held on different nights. All observations were conducted on the second (beer) or third (wine) Friday of every month from 5 to 8 p.m. (wine) or 5:30 to 7:30 p.m. (beer) when a tasting event was occurring to maximize sample size. Observers were present for 90 minutes to conduct observations, with a continuous 60 minutes used for the observation time period, over the course of two months. However, individuals had the option to purchase additional alcoholic beverages and/or food from the rest of the store to be consumed in the event area. Some individuals only drank what was available on tap and did not partake in the tasting event. Live music was also featured at these events. Study procedures were approved by The Ohio State University Institutional Review Board 2023E1316.

This study was unobtrusive such that research assistants were non-drinking attendees at the event and observed drinking behaviors of other patrons. A training manual was developed. Observers had a one hour classroom training where they reviewed the training manual and discussed the procedures for the observations. This was followed by two 2-hour practice observation training sessions and debriefing. Given that this is a pilot study, resources did not permit more extensive training.

Each observer was to observe 1–3 of the same tables or bar stools as the other observer to assess interrater reliability and an additional 1–3 on their own. This was dependent on the line of sight of the observers, the number of patrons at the event/in the bar, and the size of the parties being observed. The line of sight could include just tables, bar stools, or other individuals in the bar or special event area. The observers attempted to identify a table that gave the largest unimpeded view of the entire area, but were limited based on available sitting areas. As there were two possible locations for individuals to drink (in the bar area or the separate sitting area) for two locations, the observers prioritized observations in the bar area. Four individuals (three white men, one white female) sat at the bar briefly to purchase a drink then took that drink elsewhere (grocery shopping or in the upstairs area). These individuals are not included in our analyses. The set up for the bar and special events differed at each location. For example, two locations had the special events located in a separate upstairs area with tables and chairs and community rooms available as private meeting space. However, tickets for the event were purchased at the bar. In one location, the food and wine pairings were brought to the table after tickets were purchased, two locations provided at the bar and the individuals could take them to the separate seating area, the final location had the pairings displayed on a table where the patron would go up and help themselves to the alcohol or food. Additionally, customers of the store could purchase food from the hot or cold bars and/or the Starbucks located within the store without participating in drinking. We did not complete observations on individuals who were not drinking alcohol, unless they were included in a party that had individuals who were drinking alcohol.

Observers were asked to group patrons by their table and record demographic data such as perceived race, gender, age, as well as the number of drinks or any food consumed by each patron. Distinct physical characteristics like clothing or accessories were also recorded to aid in the tracking of number of drinks per person. All observations needed at least two observers. When a third observer was present, that person recorded whether IDs were checked while purchasing an event ticket as it was open to the public and underage individuals were allowed to be present in the space.

These observations were recorded in Google Sheets files via smart phones during the event as they were taking place that were uploaded to a secure server. Observers completed a reflection after each observation, which asked them to record any tips for future observations, anything unusual (e.g. someone visibly intoxicated), and what they would do differently. These were used to provide context to the observations, including RBS practices. Each researcher completed a collaborative observation and observed one group of observed patrons together to ensure the data was reliable and to protect internal validity. Our interrater reliability was 87.25%.

Sample

The population of interest for this study was attendees at the grocery store bars or of the paired alcohol and food events. The original sample size of the study was 96 observed patrons at four different locations. Two observed patrons were excluded from the study due to their young age (e.g. <10). Four additional observed patrons were excluded after purchasing a drink and taking it with them outside of the bar area. As we could not observe their drinking behaviors while they were shopping in the store, they were dropped from subsequent analyses. The final sample size was 90 observed patrons.

Measures

The primary outcome measure for the study is drinks per person per hour. This variable was calculated for each patron by multiplying the total number of drinks consumed by 60, then dividing that total by the number of minutes the person was present, similar to Morrison et al. (2016). A drink was measured for each glass/single beverage that had alcohol in it that was consumed. For example, if a patron bought the pairing of alcohol and food and drank alcohol for all four pairings, that was counted as four drinks. If they only consumed the alcohol for three of the pairings, they had three drinks. We recognize this might not reflect standard drink sizes as they tended to vary across the different locations based on who was pouring the alcohol. Additional measures included types of alcohol consumed, whether food was consumed, characteristics of the patrons, and the length of time that patrons were observed. Demographics of patrons observed were age, gender, and race/ethnicity. These assessments were based on outward physical characteristics of the patrons and might be subject to bias.

Age originally had eight categories: ‘<20’, ‘20–29’, ‘30–39’, ‘40–49’, ‘50–59’, ‘60–69’, ‘70–79’, and ‘80+’. Age was recoded for analysis by excluding two groups: ‘<20’ and patrons in the ‘80+’ category. The remaining age range groups were recoded into ‘20–39’, ‘40–59’, and ‘60–79’ for analysis. Gender had three categories: ‘Male’, ‘Female’, and ‘Nonbinary/Other’. However, unused categories, such as the ‘Other’ option for gender and other locations that contained no data, were removed from the analysis. Race/Ethnicity categories included ‘Black’, ‘Asian American’, ‘Native Alaskan/Native American’, ‘White’, ‘Hispanic’, and ‘Other’. This variable was recoded into ‘White’ and ‘people of color’, due to the small sample size for individuals of color. Possible types of alcohol consumed included ‘Beer’ and ‘Wine’. For each alcoholic beverage in the pairing, we collected the alcohol by volume (ABV). Each event had a card that stated what each of the pairings were (e.g. Station 1: Roasted Butternut Squash Crostini paired with Imagery Pinot Noir). Observers would then obtain the ABV for the particular alcohol included. In this example, Imagery Pinot Noir had an ABV of 13.5%. Options for food included ‘Event Food Only’, ‘Event Food and Store Bought Food’, ‘Store Bought Food Only’, and ‘No Food’; however, due to the small sample size of some of the categories, this variable was recoded to represent ‘Food’ or ‘No Food’. Finally, to assess whether the observed patrons reflected the underlying demographic distribution of the Census tracts where the grocery stores were located, we used American Community Survey (ACS) 5-year estimates to obtain data on the percentage of the Census tract population that was female, White, aged 20–39 years, 40–59 years, and 60–79 years. Census tracts are administratively defined units designed to have populations that are fairly homogenous. In our study, the average number of Census tracts across all eight locations was ~4000 residents.

Data analysis

Data were imported and analyzed through IBM® SPSS® Statistics, version 29.2.0. Frequency tables for nominal and ordinal level variables such as the location of the event, the age range of each patron, gender, food purchased, type of drink consumed, and race were created and analyzed to determine the characteristics of individuals who attended the events. We assessed for differences and commonalities in patron characteristics stratified by location of event using chi-square tests for those variables that met the underlying assumptions for a chi-square analysis. These assumptions included mutually exclusive categories and the values for the expected count based on the underlying distribution being 5 or more individuals in at least 80% of the cells (McHugh 2013). For variables where >20% of cells had an expected count of <5 (gender and race/ethnicity), we did not report a test statistic. For the remaining variables we report Pearson’s chi-square test statistic. The differences between the demographics of the sample at each location were assessed to see whether they corresponded to those of the Census tract where the store was located using one-sample tests of proportions or chi-square goodness of fit test. More specifically, we used Census data to identify the expected distributions of the variables (e.g. age groups) and compared those to the percentages found in our data. This is a crude assessment of assortative drinking. We do not provide the results for the one-sample test of proportion when the sample size is <15 (Skaik 2015) or for chi-square goodness of fit test statistics for variables that have expected counts <5 (McHugh 2013). To determine the potential relationship between various characteristics and the outcome, or the number of drinks per hour, independent samples t-tests and analysis of variance (ANOVA) tests were conducted. We use Levene’s test to test the assumption of equality of variances.

Results

The sample consisted of 56.3% (n = 54) observed female patrons and 43.8% (n = 42) observed male patrons (see Table 1). No observed patrons were coded as having a gender other than male or female. The age distribution of the sample was most concentrated among the ages of 40–49 and 60–69, as each had 28.1% (n = 27 for each) of patrons. Almost 20% (n = 17) of observed patrons were from the 30–39 age group, 13.5% (n = 13) from the 50–59 age group, and 7.3% (n = 7) from the 20–29 age group. About 2% (n = 2) of observed patrons were under 20 years old and 3.1% (n = 3_over 70 years old. The majority of observed patrons that attended these events were White (97.9%, n = 94). Participation among locations varied, with Locations 2 and 4 having the highest number of attendees. About 35% (n = 34) of observed patrons were at the bar in Location 2, followed by 31.3% (n = 30) of patrons attending the paired alcohol and food tastings at Location 4. Location 3 had the lowest representation at 12.5% (n = 12). For Location 3, these patrons represented a census of the individuals attending the tasting event on that evening. The majority of the sample (54.2%, n = 52) consumed the food that came as part of the alcohol tasting event only, with 39.6% (n = 38) of observed patrons consuming no food. A smaller proportion of the sample ate store-bought food either with or without the food pairing.

Next, we compared demographic data across all four locations for gender age, race/ethnicity, and food consumption (Table 2). Some variation was found in the proportion of male and female patrons by location. Location 1 and Location 4 had a majority of female observed patrons while Location 2 and Location 3 had a majority of male observed patrons. A chi-square test showed that there was no statistically significant difference in the gender of observed patrons among the four locations X² (3, N = 90) 5.22, P = .156.

Age varied by location; the majority of observed patrons were between 60–79 years old at Location 1, 20–39 years old at Location 2, and 40–59 years old at Location 4. Location 3 had no clear majority. The only non-White individuals included in the study were observed at Location 2 with 6.3% (n = 2) of the sample were persons of color. We did not conduct statistical tests for significance on these measures.

Variation was seen across the locations when comparing the proportion of observed patrons who did and did not have food. At Locations 1 and 4, all observed patrons had food while at Location 3 no observed patrons had food. In Location 2, 18.8% of patrons had food, but the majority had no food (81.3%). A chi-square test showed there was a statistically significant difference in the proportion of observed patrons who had food or didn’t have food among the four locations X² (3, N = 90) = 69.487, P < .01.

We assessed whether the observed sample demographics were similar to those found in the Census tracts in which the grocery stores were found (Table 3). We found statistically significant differences in the observed proportion of females in our sample compared to the expected proportions for Location 4, which had higher proportions of females compared to the underlying demographics of the Census tract. These locations also both had a statistically significant higher observed proportion of White patrons than would be expected. Finally, Location 4 had a statistically significant distribution of age groups than would be expected, with higher proportions of observed patrons in the 40–59 age group. While Locations 2 trended with more observed patrons in the 40–59 age group and Locations 1 and 3 trended with a higher proportion of those aged 60–79, we did not conduct tests of statistical significance due to the violation of key assumptions for the statistical tests.

In Table 4, we provide information by characteristics of those who drank alcohol by the number of drinks per hour. Levene’s test for equality of variance was not statistically significant for any of our comparisons; thus, we report significance levels based on the assumption of equal variance. Observed patrons who ate food during the event (M = 3.89, SD = 1.67) drank statistically significantly more drinks than those who did not eat food (M = 3.01, SD = 2.01), t(90) = −2.23, P = .03 (Table 4). After standardizing to one hour, observed patrons who drank wine (M = 4.13, SD = 1.82) drank more than those who drank beer (M = 3.00, SD = 1.72), t(90) = 2.03, P = .003. Locations of the observations had statistically significantly different drinks per hour, F = 9.89, P = <.001. There was a statistically significant difference between location and drinks per hour; observed patrons at Location 3 had the most drinks per hour (M = 4.80, SD = 1.47), t(90) = 9.89, P = <.001. The correlation between party size and number of alcoholic drinks per hour was not statistically significant r(88) = .11, P = .31. Descriptively, the ABV for the beer options ranged from 4.10% to 9.70%, and the wine options ranged from 9.50% to 14.10%.

RBS practices (e.g. checking for identification) were highly dependent upon store location. All store locations allowed children to be present at the event. Location 1 and Location 4 did not check IDs for the observers, while Location 3 and Location 2 did check IDs for the observers. At Location 1, during the first visit, we had a third observer who was able to watch all individuals purchase tickets for the special event. Here, the observer saw 16 observed patrons purchased event tickets. Of the 16 observed patrons, 15 observed patrons did not have their IDs checked. All 16 observed patrons were older than 30, and the majority of them were men. That location, however, did not check IDs for the observers, who were younger than 30. One location had signs posted with information regarding purchasing limits, including postings with a two-drink limit per customer. Yet, each event promoted four different drinks for patrons to try at the event. It appeared that no store employees were monitoring number of drinks or drink sizes. Bartenders and other employees were observed promoting the event to all patrons. During the observation time period, two patrons were observed to possibly be intoxicated. The first was observed tripping over a bar stool when leaving that appeared to be due to intoxication. The second patron had purchased a bottle of wine at the grocery store that was half full when the observation period began and was finished during the observation. We further identified four individuals who did purchase a drink in the bar area, but did not stay there or in the special event space. As we did not follow these individuals throughout the store, we do not know if they purchased only one or more than one drink during their visit. Further, during one observation period, the driving conditions deteriorated due to heavy snow and ice, yet the event was still well attended with individuals who would need to drive after the drinking event. No attempt at limiting overservice appeared to have occurred. Overall, RBS was not standardized across store locations, with a wide range of practices at each store.

Discussion

We relied on niche marketing theory and assortative drinking to assess whether specific clientele were more likely to be drinking at grocery stores, whether it be events that serve alcohol or bars in the stores, in our exploratory pilot study. Consumption of food by patrons were found to be statistically significantly different when stratifying the data by location, which could suggest that assortative drinking is occurring. We saw trending differences with ages of patrons by location that could be explored in future studies as assortative drinking. With regards to the neighborhoods where the stores were located, patron age was also statistically different compared to the underlying Census tract for one location with the others showing possible differences, in addition to the locations being different from each other, which may be another indication of assortative drinking. Observed gender, age, and race/ethnicity of the patrons at location four differed most from the underlying neighborhood population characteristics. Specifically, many of the patrons were middle-aged or older, with 40–59 representing the highest average for age, which may indicate that advertised events were not directed toward younger individuals.

Regular drinking can be an agent of social interaction for older adults. Older adults have been noted to drink alone or in environments that are less noisy and crowded (Sacco et al. 2015) and may prefer only beer and wine options (Cravero et al. 2020), making these ideal events to attend for this age group. Grocery stores might be a suitable drinking spot for this generation who aim to drink in relaxed settings and/or do not feel comfortable frequenting the same venues in areas with a robust nightlife or a dense population of college-age individuals who drink. Grocery store bars have free parking and other commodities that may be more appealing and convenient for older populations. As a growing number of older adults are misusing alcohol at a higher rate than previous older generations (Barry and Blow 2016; Kuerbis 2020; Kepner et al. 2023), establishments may be designing activities or events that draw in this niche. As the misuse of alcohol among this age group can lead to negative health outcomes later in life, these findings are a potential cause for concern. Older adults are one of the largest groups among the living generations (Choi et al. 2016) and alcohol-related health concerns in this age group are frequently underrecognized and undertreated (Barry and Blow 2016).

Drinking quantity per hour also differed by location and event. Patrons who drank at locations where food and alcohol pairing events were also being held consumed one to two more drinks than patrons drinking in bar areas. Those who drank wine had one more drink per hour than those who drank beer. This could be signaling the overall increase in wine drinking in general or may speak to differences in who is drinking wine at these events (e.g. men vs. women; Castellini and Samoggia 2018; Cravero et al. 2020) and their reasons for drinking (Thach 2012). This is a concern as the range for ABV for wine is higher than ABV in beers (including IPAs).

The drinks per hour variable differed based on food consumption, with those eating food having about one more drink per hour than those not eating food. Food is often consumed alongside alcohol to mitigate the level of intoxication, as food can absorb alcohol in the stomach (Paton 2005). Additionally, drinking alcohol can increase someone’s appetite (Hindle and Orazio 2022).

Alcohol consumption varies depending on geographical location. Alcohol use patterns among those in the urban Midwest, like those studied here, exhibit more risky drinking behaviors, rates of heavier drinking, as well as signs of alcohol use dependence compared to other regions of the United States (Dixon and Chartier 2016). In this study, we also found that drinking quantity differed by location of the event. Market segmentation appears to be occurring based on the demographics at each location and differences in the neighborhood where the store was located. For example, Location 4 has more women in in the area where the store was located. These events may provide women greater opportunity to socialize in a place that feels safe (Thach 2012). Notably, drinks per hour did not differ by men and women, mirroring increases in drinking among women overall (Hasin et al. 2019), possibly identifying another ‘niche’ for alcohol manufacturers and establishments. Gender differences in drinking suggests that overall drinking and heavy drinking are more socially stigmatized for women in comparison to men (Greaves et al. 2022) and women have a lower threshold for binge drinking than men (Choi et al. 2016). Women may feel more comfortable drinking in a grocery store compared to bars or restaurants, especially if the majority of other individuals drinking are women.

Our findings suggest a need for increased RBS training for bartenders and servers in establishments that do not traditionally serve alcohol. Employees consistently demonstrated a lack of knowledge of IDing laws and did not appear to monitor drink size or number of drinks. In Ohio, the law around checking IDs of individuals purchasing alcohol state ‘It is the duty of the seller to question any person who, from their physical characteristics, appears to be underage’ (see Ohio Revised Code Section 4301.639). That ‘most’ individuals (excluding the observers) appeared to be >30 may have led those serving the alcohol to be lax when it came to checking IDs. While the food and alcohol pairing events consisted of four drinks and four appetizers, steps could be taken to stagger the service rate as opposed to serving all four at once. This method of delivery appeared to increase the rate of alcohol consumption. Additionally, the presence of clearly underage individuals in the space was commonly noted by observers, indicating a risk for underage drinking.

RBS training is offered for free in Ohio by the Ohio Investigative Unit which could easily be mandated for employees asked to work these special events. This online training is designed to provide education about and limit the negative impacts reinforced in this study. Policy changes could also be implemented to better separate the bar area from the general consumer area, as this could limit underage interaction with the space.

Limitations

We conducted unobtrusive observations in an exploratory study examining possible niche marketing at alcohol events and bars at grocery stores in Ohio. Drinks were measured based on consuming a single beverage, however drink sizes were likely not standard across the study given the number of different individuals who poured the drinks and were not trained in doing so. We also do not have information on the total number of drinks consumed by each patron or the total time spent at the bar/special event. We are limited in our small sample size which presents several drawbacks. Given that stores were chosen due to convenience and patrons were not randomly sampled, it could affect generalizability. Further, our study only includes bivariate statistics with no adjustments for multiple testing. Our findings should be interpreted cautiously. We recognize that observers’ assessments of the age, race/ethnicity, and gender of patrons were based solely on physical features and may be prone to bias (Thomas and Freisthler 2016). The study was conducted over the course of two months at multiple store locations in Central Ohio. These locations are not representative of the overall demographics of the state or nearby cities. Although multiple locations were used, the brand of grocery store was consistent. Consequently, the study’s findings are limited to the specific brand of grocery store that was used. This is further limited by the fact that no comprehensive listing of grocery stores with bars are available. Patrons were selected based on their position relative to the researchers’ location to promote the most accurate observations. Researchers were located at seating areas close to the food and alcohol pairing event which may limit our observations of the full scope of drinking at these locations.

Conclusion

Our work suggests a need to better understand these emerging alcohol establishments. These locations provide more opportunities to drink while possibly bringing in new or different clientele drinking alcohol. Understanding who is participating in these events, the practices bartenders or servers in reducing alcohol-related problems, and an assessment of the overall atmosphere and environment would provide much needed information on the context for drinking in grocery stores. The effects of these locations on alcohol-related problems are an important next step in understanding the full impact of drinking in these locations. Drinking in non-traditional alcohol establishments may provide additional risks for individuals in recovery from alcohol use disorder. Finally, we need to understand if RBS training in these locations will change alcohol service practices.

Acknowledgements

None.

Author contributions

Claudia Banke (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Ciera Feucht (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Allie Krile (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Orazia E. Loebsack (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Tristan L. Maynard (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Kethan N. Mokadam (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Abby Schneider (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), and Bridget Freisthler (Conceptualization [equal], Formal analysis-Supporting, Methodology [equal], Writing—original draft [equal], Writing—review & editing [equal])

Conflict of interest: The authors report no conflicts of interest.

Funding

None.

Data availability

Data are available upon request of the last author.

References

Author notes