Understanding the Motives, Contexts, and Consequences of Unplanned vs. Planned Drinking in Daily Life

Abstract

Objective:

Problematic alcohol use among college students remains a public health concern; thus, there is a need to understand distinct drinking events, such as unplanned and planned drinking. The present study examined motives and social and physical contexts as correlates of unplanned and planned drinking to help inform prevention and intervention

Method:

College student alcohol and cannabis users (N=341; 53% women) completed 56 days of data collection (two 28-day bursts) with five repeated daily surveys. Three-level generalized linear mixed-effects models were used to examine associations among motives, physical and social contexts, and unplanned vs. planned drinking. We also examined whether unplanned or planned drinking resulted in greater consumption and negative consequences.

Results:

Social and enhancement motives were related to planned drinking, whereas offered motives (i.e., offers of alcohol) and coping motives were linked to unplanned drinking. Drinking at home, with roommates, or alone was associated with unplanned drinking. Drinking at a bar/restaurant, a party, at a friend’s place, with friends, with strangers/acquaintances, with a significant other, or with intoxicated people was linked to planned drinking. Unplanned drinking was related to fewer drinks consumed and fewer negative consequences endorsed.

Conclusions:

Findings showed that planned drinking – irrespective of consumption – was related to more negative consequences than unplanned drinking. In addition to targeting intentions to drink, the present study provided specific motives and social and physical contexts that are associated with planned drinking that could be incorporated into ecological momentary interventions focused on harm reduction.

Introduction

Alcohol misuse among college students is linked to both short- and long-term negative consequences, including hangover, academic problems, injury, assault, poorer mental health functioning, and, in extreme cases, death (Gruenewald et al., 2010; Hingson et al., 2009; White & Hingson, 2013). Despite these consequences, alcohol misuse in this population remains high (Schulenberg et al., 2019; Substance Abuse and Mental Health Services Administration, 2017). There is a clear need to understand the nuances of drinking in daily life to reduce harms associated with alcohol misuse. One such factor is whether the drinking event is planned vs. unplanned, which has implications for both the content and delivery of harm-reduction approaches needed to address these distinct events.

Unplanned drinking may reflect impaired control over alcohol consumption, as impaired control is defined as the “breakdown of an intention to limit consumption in a particular situation” (Heather et al., 1993, p. 701). Impaired control is known to develop early in the course of problematic use (Langenbucher & Chung, 1995; Leeman et al., 2009) and is linked to poorer treatment outcomes (Heather et al., 1998; Heather & Dawe, 2005) – emphasizing the need to better understand impaired control in daily life. Consistently, the Model of Unplanned Drinking Behavior (MUDB) posits that unplanned drinking is riskier than planned drinking because minimizing consequences, such as by using protective behavioral strategies, inherently requires a degree of planning and control (Pearson & Henson, 2013). Though not directly tested in the original work, the MUDB identified drinking motives as potential antecedents of unplanned drinking, considering their robust relations to alcohol-related outcomes (Cooper et al., 2016, for a review). For example, Pearson and Henson (2013) speculated that unplanned drinking might underlie the relation between coping motives and negative consequences; however, very little work has tested these notions or has examined other antecedents of unplanned drinking aside from affect (Fairlie et al., 2019) and impulsivity (Pearson & Henson, 2013; Stevens et al., 2017). The present study seeks to examine the reasons and contexts under which students engage in unplanned vs. planned drinking, and the degree to which unplanned vs. planned drinking is linked to greater consumption and short-term negative consequences.

Antecedents of Unplanned vs. Planned Drinking

In the only examination of drinking motives and unplanned drinking to date, Stevens, Haikalis, et al. (2021) found that drinking to enhance positive affect (“to get high/buzzed/drunk,” “to make the night more fun”) was related to planned drinking on a given day in an ecological momentary assessment (EMA) study of college students. This is in line with a recent study by Fairlie et al. (2019), which found that celebratory events (e.g., birthdays) – that likely coincide with enhancement drinking motives – were more likely to be planned (heavy) drinking occasions. However, drinking to cope with depression or anxiety or for conformity reasons were not significantly related to planned or unplanned drinking (Stevens, Haikalis, et al., 2021)

Nevertheless, the theoretical rationale of coping motives being linked to unplanned drinking is supported by affect-regulation models of substance misuse. Indeed, there is a conceptual link between affective-oriented drinking motives (enhancement, coping) and affective-oriented impulsivity (positive and negative urgency), with the latter being particularly relevant to unplanned drinking – a potential marker of uncontrolled, or impulsive, behavior. Extant research has shown that negative urgency, or acting rashly under extreme negative mood, is related to drinking to cope, which, in turn, is associated with negative alcohol-related outcomes (Adams et al., 2012; Coskunpinar & Cyders, 2012). Similarly, Stojek and Fischer (2013) found that drinking to cope amplified the relation between negative urgency and subsequent alcohol dependence symptoms. Thus, it follows that drinking to cope may be linked to more impulsive, or unplanned, drinking.

Whether and how physical and social contexts relate to unplanned vs. planned drinking is also poorly understood but bears significant relevance to the burgeoning technological advances facilitating just-in-time interventions for alcohol misuse based on the physical location or social contexts of use (Carpenter et al., 2020; Carreiro et al., 2020). Fairlie et al. (2019) is the only study to date to examine social and physical contexts of unplanned drinking in an EMA sample of college students. Findings showed that weekends and special occasions were linked to more planned than unplanned heavy drinking, whereas drinking “somewhere else” (vs. home) was related to more unplanned heavy drinking (Fairlie et al., 2019). Contrary to the authors’ hypotheses, drinking with others (vs. alone) and drinking at a bar or party (vs. home) were unrelated to unplanned or planned heavy drinking. However, this study did not consider other nuanced physical locations (e.g., drinking at a friend’s place) or social contexts (e.g., drinking with friends) that may reveal differential relations to unplanned vs. planned drinking.

For example, when considering any unplanned drinking, on-premise drinking, such as at a bar/restaurant, likely involves some intentionality, whereas off-premise drinking, such as at a friend’s place or at home, might be more opportunistic and less intentional. The social influences of unplanned drinking may be more complex and situationally dependent, such that drinking with friends could be planned in some cases (i.e. drinking buddies; Lau-Barraco & Linden, 2014) and unplanned in others via peer influences (e.g., alcohol offers; Borsari & Carey, 2001, for a review). The number of people at a given drinking event, particularly the number of those intoxicated, could also be relevant to unplanned drinking, as larger groups tend to promote a heightened drinking culture that could result in more opportunistic drinking (Clapp et al., 2006; Clapp et al., 2003). Understanding contextual influences of distinct drinking events is in its nascent stages and is complicated by the myriad possible drinking contexts (Stanesby et al., 2019, for a review) and their atheoretical nature. The present study offers the first comprehensive evaluation of contexts as nuanced predictors of unplanned vs planned drinking.

Consequences of Unplanned vs. Planned drinking

Consistent with our understanding of impaired control, a test of the MUDB showed that unplanned drinking was cross-sectionally associated with alcohol-related problems over and above quantity and frequency of alcohol consumption (Pearson & Henson, 2013). Since the MUDB was developed, experience-sampling methods have become more widely used in alcohol research, which has permitted ecologically-valid assessments of unplanned drinking by examining a discrepancy between intentions for consumption and actual consumption on a given day (as opposed to retrospectively-assessing self-reported unplanned drinking). However, evidence from these more ecologically-valid studies is mixed regarding whether unplanned or planned drinking results in more negative consequences.

Supporting the MUDB, Fairlie et al. (2019) found that unplanned, compared to planned, heavy drinking was related to a greater number of negative consequences on that day. On the other hand, Lauher et al. (2020) showed that unplanned (vs. planned) drinking – irrespective of the intended and actual amount consumed – was linked to fewer negative consequences on that day. There is likely an important distinction between any vs. heavy unplanned drinking, though more research is needed to determine the replicability of Fairlie et al. (2019) and Lauher et al. (2020). Measurement differences may also explain this disparate relation, such that Fairlie et al. (2019) assessed intentions in the afternoon and more proximal to the drinking event, whereas Lauher et al. (2020) assessed intentions in the morning. Better understanding whether unplanned or planned drinking is riskier on a given day will inform which variables should be targeted in just-in-time interventions. For example, if planned drinking is more consequential, intentions for consumption would likely be a beneficial in-the-moment intervention target. By contrast, if unplanned drinking results in more consequences, other intervention factors (e.g., motives) would need to be targeted in the absence of drinking intentions.

Present Study

The purpose of the present study was to (1) examine the associations of drinking motives, physical contexts, and social contexts with unplanned vs. planned drinking, and (2) determine whether unplanned or planned drinking results in more alcohol consumption and a greater number of negative alcohol-related consequences. We analyzed data drawn from a larger multi-site parent study of college students who provided data spanning 56 days of assessment (two 28-day bursts) with five repeated daily surveys. We are the first study to assess intentions for drinking multiple times per day and examine relations at the occasion-level, which extends prior work in this area that has focused exclusively on day-level effects. For Aim 1, consistent with Stevens, Haikalis, et al. (2021), we a priori hypothesized that drinking for social reasons or to enhance positive affect would be related to planned drinking and that coping motives would be related to unplanned drinking, consistent with the MUDB. Considering that drinking because alcohol was offered may be opportunistic, we also a priori hypothesized that this reason would be related to unplanned drinking.

For physical contexts, we proposed a priori pre-registered hypotheses¹ that planned drinking would be related to drinking at a bar/restaurant, whereas unplanned drinking would be related to drinking at a party or friend’s house. However, since pre-registration, qualitative work conducted by our research group suggested that drinking at a party or friend’s house would be associated with planned drinking (Stevens, Boyle, et al., 2021). For social contexts, we a priori hypothesized that drinking with friends would relate to unplanned drinking; however, our qualitative themes indicated that drinking with friends was most often planned (Stevens, Boyle, et al., 2021). We also a priori hypothesized that unplanned drinking would be associated with the number of intoxicated people on a given drinking occasion, given the possibility for the social norms of the event to influence a student’s decision to drink. Lastly, we a priori hypothesized that drinking with strangers/acquaintances would be linked with unplanned drinking, whereas drinking with family would be associated with planned drinking. Informed by our qualitative themes (Stevens, Boyle, et al., 2021), we included three additional contexts, which should be considered exploratory: drinking alone, drinking at home, and drinking with roommates, with new hypotheses that these contexts would be linked to unplanned drinking.

For Aim 2, we a priori hypothesized that unplanned drinking would relate to greater consumption on that survey, given the presumed uncontrolled nature of unplanned drinking, and to a greater number of negative consequences on that day, consistent with Fairlie et al. (2019) and the MUDB (Pearson & Henson, 2013). On the other hand, findings from Lauher et al. (2020) – published after our initial pre-registration – would support the opposite direction of effects.

Materials and Methods

Design and Sample

Screening survey.

Full-time college students were recruited to participate in a larger parent study on simultaneous alcohol and marijuana (SAM) use from universities in three states with legalized medical cannabis and varying recreational cannabis laws (School A: recreational cannabis use illegal; School B: recreational cannabis use decriminalized; School C: recreational cannabis use legal for adults ages 21 and older). Eight thousand students were randomly selected from each university’s registrar database stratified by expected year of graduation (total N=24,000) and were emailed an invitation to participate in an online screener. Screening completers (N=7,000) included more women, more White students, fewer Black students, more Asian students, more Hispanic/Latinx students, and younger students (ages 18–21); effect sizes for these differences were small (see White et al., 2019, for details). Of those screened, 2,874 (41.1%) were considered eligible to participate (i.e., between ages 18–24, enrolled full-time in one of the three schools, endorsed past-year alcohol and cannabis use). Students who completed the screening survey were eligible for several lotteries to win $100.

Baseline survey.

Of those eligible for the parent study, a random sample of 2,501 students (stratified by university) was invited via email to participate in the parent study’s baseline survey, and 1,524 (60.9%) invitees completed the baseline survey. We retained the data for 1,390 (91.2%) of these students after excluding participants who provided responses inconsistent with baseline survey eligibility criteria (e.g., inconsistent reporting of past-year alcohol and cannabis use) and whose surveys had technological problems/did not complete the baseline survey. See White et al. (2019) and Stevens et al. (2020) for further details regarding the baseline survey. Three months later students completed a follow-up survey. Participants were compensated $25 for the baseline survey and $35 for the follow-up survey.

Repeated daily surveys.

Of those who completed the baseline survey, 693 used alcohol and cannabis at the same time ‘so that their effects overlapped’ (i.e., SAM use) within the past month, deeming them eligible for the repeated daily survey (RDS) phase. Invited participants (N=596) were stratified based on frequency of past-month SAM use and sex assigned at birth. A cap was placed on each category for sex and SAM-use frequency within each school (i.e., recruitment site), with invitations favoring more frequent SAM users (i.e., three or more times in the past month) and male participants to ensure sufficient base rates of SAM use in the RDS phase and achieve a balance of male and female participants. Enrollment for this phase was conducted on a rolling basis until quotas were filled; therefore, not all of those invited could be enrolled in this phase. Of the 379 students who were given access to the custom-designed smartphone application, 343 students (90.5%) agreed to participate in this study phase. See Supplemental Figure S1 for a recruitment and retention flowchart of the parent study.

Data collection directly followed the longer surveys (baseline and three-month follow-up) and comprised 28 days of RDS at each burst (56 total days). Surveys were prompted at 9:00 am, 2:00 pm, 5:00 pm, 8:00 pm, and 11:00 pm daily using a custom smartphone application (see Stevens et al., 2020 for details). For the first survey of the day prompted at 9:00 am, students were also asked additional questions assessing yesterday’s behavior. Participants were provided five hours to complete the 9:00am survey and two hours to complete all other surveys.² Reminders were provided to participants 15 minutes before surveys closed. Participants were compensated $1 for each completed daily survey, with weekly and overall bonuses (for a total possible compensation of $200 per burst) to encourage high response rates. See Stevens et al. (2020) and Sokolovsky et al. (2020) for additional details regarding the parent study’s RDS bursts. For each of the five surveys, compliance was as follows: 9:00 am (78%), 2:00 pm (59%), 5:00 pm (66%), 8:00 pm (65%), and 11:00 pm (56%). Survey-level compliance was significantly lower for the second burst compared to the first burst (χ² = 29.09, p < .05). An average of 3.75 surveys per day (SD=1.29) were completed by each participant across the study, with a range of 1 to 5. Across all study days, participants completed an average of 175 surveys (SD=69.44) out of 270 possible surveys. Notably, however, we achieved complete coverage (i.e., no missing data) on 75% of study days because the parent study design was such that if one survey was missed the next survey asked about the experiences that occurred anytime since the last survey submission (see Stevens et al., 2020, for details). All participants were trained on standardized drink equivalences set forth by the National Institute of Alcoholism and Alcohol Abuse (NIAAA; NIAAA, 2007). Study procedures were approved by the coordinating university’s Institutional Review Board. A Certificate of Confidentiality was obtained from the National Institute on Drug Abuse.

The first two study days (of the 56 total days) were excluded from analyses due to technical difficulties that occurred on these days; thus, 54 days were included in analyses. . Excluding two participants who only completed the first two study days, the final RDS sample comprised 341 students. Among these students (53% women; M age=19.79; 74% White, 11% Asian, 9% bi- or multi-racial, 3% Black, 0.7% American Indian, 0.2% Pacific Islander, 2% other race; 10% Hispanic/Latinx); 32.3% of students were from School A, 34.1% from School B, and 33.5% from School C. In the larger RDS parent study, 75% of surveys were non-use occasions (n=44,278), 7% were alcohol-only occasions (n=4,167), 14% were cannabis-only occasions (n=8,527), and 4% were alcohol and cannabis co-use (i.e., concurrent or simultaneous use) occasions³ (n=2,343). In the present study, our analytic sample was restricted to any occasions endorsing alcohol use (alcohol-only or co-use occasions; n=6,510) nested within the 54-day assessment window nested within 331 students who contributed at least one day of drinking data.

Measures

Demographics.

Participants self-reported demographic information at the baseline survey, including age (continuous) and sex assigned at birth (‘1’ for male, ‘0’ for female).

Unplanned vs. planned drinking.

At each RDS, regardless of substance use, participants were asked: “Had you planned to drink between [time X] and [time Y]?” Options included yes (‘0’) or no (‘1’). The timeframe for each question was from the time the previous survey was submitted to the time the current survey began. An unplanned drinking (‘1’) occasion was defined by a participant endorsing ‘no’ to planning to drink in that time window while also reporting any alcohol consumption in that same time window (see Alcohol consumption below). A planned drinking (‘0’) occasion was defined by endorsing ‘yes’ to planning to drink in that time window while also reporting alcohol consumption in that same time window. For most drinking occasions, drinking was planned (n=4,612 occasions; 73%) vs. unplanned (n=1,746; 27%).⁴ Days where participants reported a mix of both unplanned and planned drinking occasions were minimal (n = 271 days; 0.07%).

Drinking motives.

Following the endorsement of alcohol-only use, participants were asked at each RDS: “What motivated you to drink between [time X] and [time Y]?” Following the endorsement of co-use, participants were asked at each RDS: “What motivated you to drink and use marijuana between [time X] and [time Y]?” Participants were instructed to select all motives from the following: “to be social” (social; 40% of drinking occasions), “to cope” (coping; 10%), “it was offered” (offered; 27%), “to have fun” (enhancement; 75%), “to fit in” (conformity; 2%), “expand awareness” (expansion; 5%), “get higher from another drug” (effect enhancement; 0.36%), and “was too high from the other drug” (offset; 0.37%). As part of the parent study, motives were adapted from a psychometrically-valid measure of drinking motives (Drinking Motives Questionnaire Revised; Cooper, 1994) as well as a validated measure of cannabis use motives (Marijuana Motives Measures, Simons et al., 1998) and of alcohol and cannabis co-use motives (Patrick et al., 2018). Consistent with our pre-registration, we selected four motives a priori for the present study that are conceptually related to planned (social, enhancement) vs. unplanned drinking (coping, offered).

Drinking contexts.

Physical contexts.

Participants reported on their physical contexts at each RDS: “Where were you while you were using alcohol?” Options included (select all that apply): home (47% of drinking occasions), friend’s place (30%), party (11%), bar/restaurant (14%), outside (2%), study space (0.51%), athletic facility (0.22%), elsewhere (7%). In the present study, we only considered home, friend’s place, party, and bar/restaurant as indicators of physical context.

Social contexts.

Participants reported on their social context of drinking at each RDS: “Who were you with while you were using alcohol?” Options included (selected all that apply): alone (11% of drinking occasions), significant other (20%), roommate (25%), friend (57%), family (12%), strangers (8%), acquaintance (9%), and others (3%). Consistent with our pre-registration and recent qualitative work, we examined all social contexts except for drinking with others.⁵ We initially examined strangers and acquaintances in separate models; however, considering these models produced similar effects (both directionally and effect size), we collapsed drinking with strangers/acquaintances into a single social context. We also included the number of people reported being intoxicated with the participant at a given RDS, with options for none (39% of drinking occasions), some (18%), most (21%), and all (22%).

Alcohol consumption.

At each RDS, participants indicated the number of drinks since their last RDS using a graphical interface, tapping on the timeline at each specific time a drink was consumed (see Stevens et al., 2020, Supplemental Materials, for screenshots): “Tap your finger in the blue box each time you had a drink at the corresponding time.” The sum of drinks reported at each completed RDS on a given day was analyzed as an Aim 2 outcome (M=3.12 drinks per occasion, SD=2.36).⁶

Negative consequences.

On the morning survey following a drinking day, participants indicated whether the following consequences occurred “because of yesterday’s use of alcohol” and/or “because of yesterday’s use of alcohol and marijuana together”⁷: hangover (10% of daily observations), nausea/vomiting (4%), hurt self (0.59%), drove car drunk/high (5%), blackout (2%), rude/aggressive (0.80%), and unwanted sex (0.30%). As part of the parent RDS study, we considered consequences across several validated measures, including the Brief Young Adult Alcohol Consequence Questionnaire (Kahler et al., 2005), Brief Marijuana Consequences Questionnaire (Simons et al., 2012), Young Adult Alcohol Consequences Questionnaire (Read et al., 2006), Rutgers Alcohol Problem Index (White & Labouvie, 1989), and the Rutgers Marijuana Problem Index (White et al., 2005), selecting acute consequences that we expected to vary at the daily level. In the present study, we examined the total number (0–7) of consequences on a given day as an Aim 2 outcome (M=0.23 consequences per day, SD=0.54, range= 0–5).

Covariates.

We included age and sex (male vs. female [ref]) as covariates, given the potential for alcohol consumption to differ for these demographic variables (Erol & Karpyak, 2015). We also adjusted for school (i.e., recruitment site) because this is a multi-site study, and models should account for potential differences across sites (ref=School C). We adjusted for weekend (i.e., Friday and Saturday vs. weekday [Sunday-Thursday] = ref), consistent with prior research showing the alcohol consumption is heavier on weekends (Finlay et al., 2012; Maggs et al., 2011) and the possibility that planned drinking is more frequent on weekends (Lauher et al., 2020; Stevens et al., 2017). We included type of use occasion (alcohol-only vs. alcohol and cannabis co-use [ref]) as a covariate as well, given that alcohol occasions analyzed in the present study could also include cannabis use, which may be a potential confound (Risso et al., 2020). In the Aim 2 model predicting consequences, we also included number of drinks consumed on that day as a covariate, providing a more robust test between unplanned drinking and consequences.

Analytic Strategy

Data management was conducted in SAS 9.4^™ software (SAS Institute Inc., 2012). Analyses were restricted to drinking occasions (n=6,510), including a subset of those occasions where cannabis was also consumed (n=2,343; 36% of occasions)⁸. Occasions (n=6,510; Level-1) were nested within days (n=4,142 days; Level-2) within participants (N=331; Level-3).⁹ Multilevel modeling was used to account for the nested structure of the data. Generalized linear mixed-effects models (GLMMs), an extension of multilevel modeling, were used for Aim 1 to account for the binary outcome (unplanned vs. planned drinking); these models included all three levels, apart from the model examining associations between unplanned vs. planned drinking and negative consequences. Consistent with recommendations, occasion-level variables were day-mean centered, day-level variables were person-mean centered, and person-level variables were grand-mean centered (Curran & Bauer, 2011).¹⁰ Day- and person-level variables for binary indicators (motives, contexts) were defined as the proportion of endorsement of a given construct across surveys on a given day for a given person (day-level) and the proportion of endorsement across study days for a given person (person-level).

Aim 2 models included an occasion-level negative binomial outcome (alcohol consumption) and a day-level negative binomial outcome (negative consequences). For the former model, we included all three levels. For this latter model, unplanned vs. planned drinking was an indicator at the day-level, rather than occasion-level, such that this model included only two-levels (days within participants). In this two-level model, day-level unplanned vs. planned drinking was defined as the proportion of surveys across a given day in which unplanned drinking was endorsed, and person-level unplanned drinking was defined as the proportion of days across all study days in which unplanned drinking was endorsed for a given person.

Results

See Table 1 for descriptive statistics of alcohol consumption, motives, and contexts by unplanned and planned drinking occasions and days. See Supplemental Table S1 for descriptive information about the overlap of social and physical contexts (e.g., drinking at home alone) by unplanned and planned drinking occasions. The intraclass correlation coefficients (ICC) for the three-level Aim 1 outcome indicated 7% of the variance in unplanned vs. planned drinking was at the person-level (ICC=.07) and 18% was at the day-level (ICC=.18). The ICC for the three-level Aim 2 outcome also suggested 7% of the variance in alcohol consumption was at the person-level (ICC=.07) and 25% was at the day-level (ICC=.25). The ICC for the two-level Aim 2 outcome indicated 21% of the variance in negative consequences occurred at the person-level. Results for occasion-level effects are interpreted as deviations from a given day, accounting for typical endorsement of a construct (motives, contexts, unplanned drinking) for a given person across study days. Occasion- and day-level effects are of central interest in the present study; person-level effects are accounted for in each model but not interpreted for parsimony. See Supplemental Tables S2, S3, and S4 for person-level and covariate effects.¹¹

Table 1.

Descriptive statistics of alcohol consumption, motives, and contexts by unplanned and planned drinking occasions and days

Focal variable	Unplanned Drinking Occasion (n = 1,746)	Planned Drinking Occasion (n = 4,612)	Unplanned Drinking Day (n = 1,366)	Planned Drinking Day (n = 2,609)
Alcohol Consumption
Number of drinks	M= 2.38 (SD = 1.89)	M= 3.55 (SD = 2.40)	M= 2.44 (SD = 1.83)	M= 3.59 (SD = 2.28)
N (%) HED	274 (16.10%)	1,574 (34.89%)	301 (22.04%)	1,174 (45.00%)
Motives
Social	458 (26.28%)	2,109 (45.83%)	448 (31.93%)	1,352 (50.30%)
Coping	285 (16.35%)	373 (8.11%)	246 (17.53%)	251 (9.34%)
Offered	587 (33.68%)	1,106 (24.03%)	545 (38.85%)	784 (29.17%)
Enhancement	928 (53.24%)	3,836 (83.36%)	821 (58.52%)	2,301 (85.60%)
Conformity	38 (2.18%)	102 (2.22%)	38 (2.71%)	85 (3.16%)
Expansion	87 (4.99%)	221 (4.80%)	74 (5.27%)	159 (5.92%)
Effect-enhancement	5 (0.39%)	10 (0.36%)	5 (0.47%)	10 (0.57%)
Offset	8 (0.46%)	16 (0.35%)	7 (0.50%)	16 (0.60%)
Contexts
Home	998 (57.13%)	1,993 (43.22%)	837 (59.62%)	1,300 (48.31%)
Friend’s place	404 (23.13%)	1,489 (32.29%)	374 (26.64%)	971 (36.08%)
Party	66 (3.78%)	651 (14.12%)	84 (5.98%)	519 (19.29%)
Bar/restaurant	183 (10.48%)	693 (15.03%)	192 (13.68%)	1,212 (86.32%)
Outside	36 (2.06%)	88 (1.91%)	42 (2.99%)	70 (2.60%)
Study space	18 (1.03%)	15 (0.33%)	16 (1.14%)	15 (0.56%)
Athletic facility	1 (0.06%)	13 (0.28%)	1 (0.07%)	13 (0.48%)
Elsewhere	95 (5.44%)	341 (7.40%)	100 (7.12%)	255 (9.48%)
Alone	323 (18.49%)	341 (7.40%)	284 (20.23%)	266 (9.88%)
Significant other	314 (17.97%)	972 (21.08%)	282 (20.09%)	590 (21.92%)
Roommate	409 (23.41%)	1,194 (25.89%)	364 (25.93%)	782 (29.06%)
Friends	784 (44.88%)	3,089 (66.99%)	691 (49.22%)	1,904 (70.75%)
Family	210 (12.02%)	565 (12.25%)	189 (13.46%)	357 (13.27%)
Strangers	89 (5.09%)	463 (10.04%)	100 (7.12%)	358 (13.30%)
Acquaintance	102 (5.84%)	510 (11.06%)	114 (8.12%)	383 (14.23%)
Others	52 (2.98%)	121 (2.62%)	56 (3.99%)	106 (3.94%)
Number of intoxicated people
’None’	1,109 (63.52%)	1,354 (29.36%)	955 (69.91%)	985 (37.75%)
’Some’	242 (13.86%)	941 (20.41%)	264 (19.33%)	711 (27.25%)
’Most’	181 (10.37%)	1,152 (24.98%)	214 (15.67%)	866 (33.19%)
’All’	214 (12.26%)	1,164 (25.24%)	248 (18.16%)	824 (31.58%)

Note. Heavy episodic drinking (HED) was defined as 4+/5+ drinks for females/males. Unplanned drinking day was defined if a participant reported any unplanned drinking occasions on a given day. Planned drinking day was defined if a participant reported no unplanned drinking occasions on a given day. Day-level motives and contexts variables were constructed if the participant endorsed a given motive or context on at least one of five surveys on a given day.

Aim 1: Predictors of Unplanned vs. Planned Drinking

Motives.

At the occasion- and daily-levels, offered motives were related to greater odds of unplanned drinking, after adjusting for covariates, whereas social and enhancement motives were associated with greater odds of planned drinking. At the daily-level, but not at the occasion-level, coping motives were related to greater odds of unplanned drinking (see Table 2).

Table 2.

Associations between drinking motives and contexts and unplanned vs. planned drinking

Independent variable	Est	SE	OR	95% CI	p-value
Motives
Occasion-level (Level-1)
Offered	0.53	0.24	1.71	(1.06, 2.73)	.03
Coping	0.33	0.43	1.39	(0.60, 3.24)	.45
Social	−0.81	0.25	0.45	(0.27, 0.72)	<.01
Enhancement	−1.89	0.34	0.15	(0.08, 0.30)	<.01
Day-level (Level-2)
Offered	1.25	0.22	3.49	(2.29, 5.32)	<.01
Coping	2.14	0.33	8.53	(4.45, 16.36)	<.01
Social	−2.11	0.24	0.12	(0.08, 0.19)	<.01
Enhancement	−13.97	0.56	<.01	(<.01, <.01)	<.01
Contexts
Occasion-level (Level-1)
Physical Contexts
Home	0.54	0.25	1.71	(1.04, 2.80)	.03
Party	−0.83	0.36	0.43	(0.22, 0.88)	.02
Bar/restaurant	−0.95	0.31	0.39	(0.21, 0.71)	<.01
Friend’s place	−0.07	0.25	0.93	(0.58, 1.51)	.77
Social Contexts
Alone	1.22	0.36	3.39	(1.67, 6.89)	<.01
Friends	−1.63	0.34	0.20	(0.10, 0.39)	<.01
Roommate	0.01	0.33	1.01	(0.53, 1.92)	.98
Significant other	−1.27	0.41	0.28	(0.12, 0.63)	<.01
Family	−0.29	0.43	0.75	(0.32, 1.72)	.49
Stranger/acquaintance	−0.72	0.30	0.49	(0.27, 0.88)	.02
Number of intoxicated people	−0.97	0.13	0.39	(0.30, 0.49)	<.01
Day-level (Level-2)
Physical Contexts
Home	1.90	0.23	6.67	(4.29, 10.36)	<.01
Party	−2.56	0.31	0.08	(0.04, 0.14)	<.01
Bar/restaurant	−1.50	0.26	0.22	(0.13, 0.37)	<.01
Friend’s place	−1.00	0.19	0.37	(0.25, 0.53)	<.01
Social Contexts
Alone	2.94	0.39	18.90	(8.79, 40.63)	<.01
Friends	−3.34	0.61	0.04	(0.01, 0.12)	<.01
Roommate	0.53	0.20	1.70	(1.15, 2.51)	<.01
Significant other	−0.44	0.23	0.64	(0.41, 1.01)	.05
Family	0.06	0.22	1.06	(0.69, 1.62)	.80
Stranger/acquaintance	−1.50	0.26	0.22	(0.14, 0.37)	<.01
Number of intoxicated people	−2.48	0.30	0.08	(0.05, 0.15)	<.01

Note. N = 331; OR = odds ratio. Effects reflect odds of an unplanned drinking occasion, relative to a planned drinking occasion (reference group). The four motives were tested in separate models for interpretability. Significant findings and directions of effects were unchanged when tested in a multivariate model. Each social and physical context was tested in a separate model for interpretability. All models included age, sex (male vs. female = ref), school (School A, School B vs. School C = ref), weekend (vs. weekday = ref), and type of use survey (alcohol-only vs. co-use = ref) as covariates. See Supplemental Tables S1 and S2 for person-level and covariate effects

Contexts.

At both the occasion- and daily-levels, drinking at home or alone was related to unplanned drinking, whereas drinking at a party, bar/restaurant, with friends, with strangers/acquaintances, or with a greater number of intoxicated people was associated with planned drinking. Drinking with a significant other was related to planned drinking at the occasion-level; this relation was marginally significant at the daily-level (p = .05). On the other hand, drinking with a roommate was unrelated to planned or unplanned drinking at the occasion-level but was associated with greater odds of unplanned drinking at the daily-level. Drinking at a friend’s place was unrelated to planned or unplanned drinking at the occasion-level but was related to greater odds of planned drinking at the daily-level. Drinking with family was the only context that was unrelated to planned or unplanned drinking at both the occasion- and daily-level (see Table 2).

Aim 2: Outcomes of Unplanned vs. Planned Drinking

At the occasion- and daily-levels, unplanned drinking was associated with fewer drinks consumed after adjusting for covariates. Unplanned drinking on a given day was also related to fewer consequences after adjusting for covariates, including the amount of alcohol consumed on that day (see Table 3).

Table 3.

Outcomes of Unplanned vs. Planned Drinking

Independent variable	Est	SE	IRR	95% CI	p-value
DV: Number of Drinksa
Occasion-level (Level-1)
Unplanned occasion (ref = planned)	−0.26	0.05	0.77	(0.70, 0.85)	<.01
Day-level (Level-2)
Unplanned occasion (ref = planned)	−0.42	0.02	0.66	(0.63, 0.69)	<.01
DV: Number of Consequencesb
Day-level (Level-1)
Unplanned occasion (ref = planned)	−0.18	0.08	0.83	(0.72, 0.97)	.02

Note. IRR = incidence rate ratio

^aOccasion-level (Level-1) effects were day-mean centered, daily-level (Level-2) effect were person-mean centered.

^bConsequences were assessed once daily, and this model is two-levels: Day-level (Level-1) effects are person-mean centered and Level-2 (person-level) effects were grand-mean centered.

Discussion

In the present study, we examined how drinking motives and physical and social contexts were associated with unplanned vs. planned drinking in a large, multi-site study of college students using repeated daily surveys spanning 54 days of assessment. We also investigated whether unplanned or planned drinking yielded greater alcohol consumption and negative consequences. To our knowledge, this is the second study to date to examine drinking motives as indicators of any unplanned or planned drinking and is the first to also include physical and social contexts of drinking. We are also the first study to examine multiple unplanned and planned drinking occasions within a given day, which extends prior literature that has focused exclusively on unplanned and planned drinking at the day level.

Motives and Unplanned vs. Planned Drinking

Our findings for relations between specific motives and contexts and unplanned vs. planned drinking largely supported our a priori hypotheses as well as emerging work in this area. Specifically, at the occasion- and daily-levels, drinking for social reasons and for enhancement was related to greater odds of planned drinking, which corroborates recent findings by Stevens, Haikalis, et al. (2021) showing that social drinking events and those for the purpose of having fun are often planned, rather than unplanned. Though Stevens, Haikalis, et al. (2021) examined drinking motives and unplanned and planned drinking, the present quantitative study extends this work by isolating the occasion-level associations between motives and unplanned and planned drinking, which provides further nuance about these relations that are more proximal to the drinking event than daily-level effects.

The present study is the first, to our knowledge, to examine drinking because alcohol was offered as a reason for alcohol use and also to examine its relation to unplanned vs. planned drinking. Because of its inherent opportunistic nature, we hypothesized that using alcohol for this reason would be more strongly related to unplanned drinking. Our findings robustly supported this direction of effect at the occasion and daily levels. Consistent with the MUDB (Pearson & Henson, 2013), we also hypothesized that coping motives would be related to unplanned, rather than planned, drinking, and our findings supported this hypothesis at the daily-level but not at the occasion-level. By contrast, when examining coping motives disaggregated by depression and anxiety, Stevens, Haikalis, et al. (2021) did not find a daily-level relation between drinking to cope and unplanned or planned drinking. Thus, though the present study largely supported the MUDB, more research is needed to determine the replicability of the relation between coping motives and unplanned drinking, given the disparate relation found between Stevens, Haikalis, et al. (2021) and the present work.

Physical Contexts and Unplanned vs. Planned Drinking

Building upon initial work by Fairlie et al. (2019), we examined associations between four physical contexts and unplanned vs. planned drinking. As hypothesized, we found that drinking at a bar or restaurant was related to planned drinking at the occasion- and daily-levels. We provided competing hypotheses for the relation between drinking at a party; findings from the present study largely supported our recent qualitative work in this area (Stevens, Boyle, et al., 2021), but not our a priori hypothesis, by robustly showing that drinking at a party was related to planned drinking at the occasion- and daily-levels. However, we did not distinguish between the size and type of party, which might account for differences in unplanned vs. planned drinking at parties. For example, informal gatherings in one’s own home or a nearby location, such as a residence hall, might be more opportunistic and unplanned, whereas a large party, such as a birthday celebration, is more likely to be planned. Our findings also are somewhat in contrast to Fairlie et al. (2019) who did not find significant relations between drinking at a bar or party, relative to drinking at home, and unplanned or planned heavy drinking. Thus, there appears to be an important distinction between drinking without any intention, as examined here, vs. the definition used by Fairlie et al. (2019) that included drinking with no intention for heavy drinking and drinking more than intended.

We provided similar competing hypotheses for drinking at a friend’s place and ultimately supported our qualitative themes (Stevens, Boyle, et al., 2021) by showing that drinking in this location on a given day, but not on a given occasion, was related to planned drinking. To our knowledge, we are the first to examine this location for these two types of drinking events; thus, more research is needed to determine the replicability of this effect, including whether these findings hold when examining unplanned heavy drinking or drinking more than intended. Though speculative, it is possible that drinking at a friend’s place may require additional planning (e.g., alcohol availability) that is not required by other physical contexts (e.g., bar/restaurant), which may explain why we did not find an occasion-level effect (more impromptu drinking) but did find a daily-level effect for this context. Indeed, these competing effects at the occasion- vs. daily-level also highlight the importance of disaggregating effects when using nested data, as it cannot be assumed that directions of effects or statistically significant effects will be similar across levels – an ecological fallacy (Curran & Bauer, 2011). Informed by our recent qualitative work, we also examined whether drinking at home relates to unplanned drinking; we supported this effect at both occasion- and daily-levels. Again, it would be important for future research to determine whether drinking at home is significantly related to drinking more alcohol than intended, as opposed to any unplanned drinking, which would better inform our understanding of situations marked by impaired control over alcohol consumption.

Social Contexts and Unplanned vs. Planned Drinking

We also considered associations between social contexts and unplanned vs. planned drinking. At both occasion- and daily-levels, drinking with friends, with strangers/acquaintances, or with a greater number of intoxicated people was related to planned drinking, rather than unplanned drinking, which corroborated findings from our recent qualitative research but not necessarily our a priori hypotheses. Indeed, we originally expected that drinking with friends, strangers/acquaintances, or with larger group of intoxicated people might influence an individual’s decision to drink via the social norms of the drinking event. That said, when young adults described intending to drink – at all and regardless of level of consumption – they were most often with friends or around other people, including those who were intoxicated (Stevens, Boyle, et al., 2021). Thus, future research that examines the influence of drinking with friends or intoxicated people on other types of unplanned drinking (e.g., drinking more than intended) may find disparate relations. Though conjecture, this would support the role of social influence on impaired control, but this notion has yet to be tested.

We also tested whether drinking with a significant other or with family would be related to planned drinking. Supporting our original hypotheses and our recent qualitative work, drinking with a significant other at the occasion-level was linked to planned drinking; this same effect approached significance at the daily-level (p = .05). Contrary to our hypotheses and our qualitative findings, drinking with family was not related to unplanned or planned drinking. To our knowledge, these two social contexts have not been assessed quantitatively in relation unplanned vs. planned drinking; thus, these findings are preliminary and require replication before firm conclusions can be drawn.

As an exploratory analysis, we examined the associations between drinking alone and unplanned drinking and supported this relation at the occasion and daily levels. Again, this was not found by Fairlie et al. (2019) for unplanned heavy drinking, so it is important that future research examines the replicability of this effect when investigating other types of unplanned drinking aside from drinking with no intentions. Building upon our qualitative work, we examined whether drinking with roommates was related to unplanned drinking, and we supported this direction of effect at the daily-level but not at the occasion-level. This effect is largely consistent with our findings for drinking at home, with both contexts likely providing more opportunity for fortuitous drinking to occur with little planning required.

Though our modeling approach examined social and physical contexts as focal variables and unplanned vs. planned as the dependent variable, it is likely that individuals who plan to drink then go to a bar/restaurant, a party, or to a friend’s place – all physical contexts in which individuals are likely to be with friends and/or with individuals who are also intoxicated. Thus, it follows that these social situations– marked by intentions to drink – then lead to greater consumption and consequences (Clapp et al., 2007; Wechsler & Nelson, 2008). On the other hand, college students who are alone or with roommates and/or at home may end up drinking, likely because planning it is not required for these situations (e.g., no travel required), but their alcohol consumption is lower in these situations as is the number of negative consequences endorsed. This is likely because these situations are less social and thus less conducive to heavy drinking. That said, some research has also shown that some college students engage in heavy drinking while they are alone, which has been shown to result in a greater number of negative consequences than students who drink heavily in social contexts (Christiansen et al., 2002). Thus, there remains a need to better understand the situations and motives under which students engage in heavy drinking – including whether this event was planned or unplanned – to best inform intervention targets for ecological momentary interventions.

Alcohol Consumption and Negative Consequences

Informed by Fairlie et al. (2019) and the MUDB (Pearson & Henson, 2013), we originally expected that unplanned drinking would yield greater consumption and a greater number of negative consequences, as the MUDB postulates that using harm-reduction strategies inherently requires a degree of planning. However, our findings for any unplanned drinking did not support this notion. Unplanned drinking in the present study was consistently linked to lower levels of consumption at the occasion- and daily-levels. This corroborates another recent study examining any unplanned drinking among college students (Lauher et al., 2020). As indicated above, the contexts under which students reported planned drinking were more social (e.g., at a party, with friends); thus, greater alcohol consumption is expected in these situations. For example, though not assessed in the present study, these contexts that were associated with planned drinking may have involved pre-gaming or drinking games, which is linked to heavier consumption (e.g., Borsari, 2004; Pedersen & LaBrie, 2007). Planned drinking may also have ensured there was sufficient alcohol available, which may have resulted in heavier drinking than occasions where alcohol is more limited. Consistent with Lauher et al. (2020), we also found that unplanned drinking was associated with experiencing fewer consequences, even after accounting for the amount of alcohol consumed on that day. As reiterated above, and coupled with Lauher et al. (2020), our findings support a consistent and important distinction between any unplanned drinking and other definitions of unplanned drinking, such as unplanned heavy drinking/drinking more than intended. In sum, planning to drink at all, irrespective of planned and actual consumption, appears to be more consequential than any unplanned drinking.

Strengths and Implications

To our knowledge, this is the first study to examine associations of drinking motives and social and physical contexts with any unplanned or planned drinking while disaggregating the occasion-level effects (spanning approximately three hours of time) from the daily-level effects. Both occasion-level and daily-level findings gleaned using this rigorous analytic approach have significant implications for informing ecological momentary interventions (EMI). For example, we showed that drinking for social reasons or to enhance positive affect is robustly related to planned, rather than unplanned, drinking. We also showed that drinking at a party, at a bar restaurant, at a friend’s place, with friends, with strangers/acquaintances, or with a greater number of intoxicated people is also linked to planned drinking. Considering we also found that planned drinking was associated with greater consumption and experiencing a greater number of negative consequences, these specific motives and contexts could be used as intervention targets in harm reduction-focused EMIs, in addition to targeting drinking intentions. For example, individuals who endorse these reasons for drinking or drinking in these contexts could receive a follow-up text message that provides them with suggestions for protective behavioral strategies to both reduce consumption (e.g., alternating alcoholic and nonalcoholic drinks) and negative consequences (e.g., eat before or during drinking).

We also found that drinking to cope or because alcohol was offered and drinking at home, alone, or with roommates was associated with unplanned, rather than planned, drinking. Though we did not support that any unplanned, compared to planned, drinking was related to more consumption and consequences, as originally expected, future research is needed to determine whether these specific drinking motives and contexts are linked to unplanned heavy drinking or drinking more than intended, which has been shown to be more problematic than planned heavy drinking (Fairlie et al., 2019). Indeed, our recent qualitative research (Stevens, Boyle, et al., 2021) showed that intentions and willingness for drinking shift across the drinking event. Thus, though these contexts may be initially associated with unplanned any drinking (and thus less problematic), intentions and willingness to continue drinking may shift, thereby creating additional risks that warrant harm-reduction intervention. Thus, it may be beneficial for EMIs to consider incorporating these proximal motives/contexts as well, as much work has shown the potential long-term consequences of drinking to cope (Cooper et al., 2016) and drinking alone (Corbin et al., 2020; Gonzalez et al., 2009; Keough et al., 2015), in particular.

For example, research examining solitary drinking has shown that this behavior often co-occurs with drinking to cope (Corbin et al., 2020; Gonzalez et al., 2009), suggesting that this motive-context “match” is a potentially risky situation that could be targeted in tandem using EMI, if future research supports that drinking alone and to cope are also associated with unplanned heavy drinking – a more problematic outcome. If so, individuals endorsing this motive and/or context could be provided with alternative coping strategies (e.g., distress tolerance skills; Linehan, 1993) and/or could be encouraged to engage with their support network, if reporting solitary drinking. Reducing the likelihood of unplanned drinking among those endorsing drinking because alcohol was offered poses unique challenges considering the inherent opportunistic nature of this reason for drinking. Considering that we found a daily-level relation between offered motives and unplanned drinking, in addition to an occasion-level effect, individuals who endorse this reason for drinking on a given day could be provided with strategies to enhance their drink refusal self-efficacy skills to reduce the likelihood of drinking for this reason on a subsequent day (Choi et al., 2013).

Limitations and Future Directions

Despite the strengths and potential implications of the present study, findings should be considered alongside their limitations. First, data were drawn from a large sample of college students who also endorsed past-month cannabis use and who were majority White; thus, our findings may not generalize to college students who do not use cannabis or to other demographic groups. Future research should replicate our study aims including both college and non-college individuals to improve the generalizability of our conclusions. In particular, the social and physical contexts of drinking likely differ for college and non-college young adults, which may reveal differential effects with unplanned and planned drinking for these two groups. Further, it is possible that motives and contexts for planned and unplanned drinking may vary in a younger population of adolescent drinkers whose drinking is inherently more opportunistic. Second, the repeated daily surveys did not capture data in real-time and thus required a small degree of retrospection. We recommend that future studies consider examining the effects of motives and contexts on unplanned and planned drinking when collected in real-time. Third, the morning survey completion rate was similar to a recent meta-analysis of pooled completion rates of EMA studies focusing on substance use (Jones et al., 2019); however, afternoon and evening surveys had lower completions rates (<66%), which may have affected study findings given plans for drinking and drinking events may be more likely to occur later in the day.

Fourth, intentions for drinking were retrospectively assessed (i.e., “Had you planned to drink between [time X] and [time Y]?”), and recall bias may have precluded students from accurately remembering if the drinking event was unplanned. Future work in this area should prioritize assessing intentions for drinking prior to the drinking event. Assessing intentions prospectively, prior to drinking, would also allow researchers to ask participants about whether they anticipate having alcohol available that day to determine the interplay between intentions for drinking and availability. Intentions for a specific amount of consumption were not assessed, as has been done in prior work (Fairlie et al., 2019), which affords an examination of unplanned vs. planned heavy drinking. Future work should prioritize multi-part questions to assess intentions, including any plans for drinking, intentions for amount of consumption, and intentions for level of intoxication (Stevens, Boyle, et al., 2021). To our knowledge, changes in alcohol availability (e.g., not anticipating access to alcohol in the morning but gaining access later that night) have not been examined as a potential factor that may influence unplanned drinking. Future research could test this notion. Fifth, we included drinking at a party as a physical context in the present study, given it was endorsed on 11% of drinking occasions. However, we acknowledge that drinking at a party could be both a physical and social context. Future research could consider assessing the physical location and nature of the party (e.g., informal vs. formal gathering; at a friend’s home vs. at a fraternity/sorority house).

There is a similar limitation for drinking at a friend’s place. Future research could collect additional details about drinking at specific locations to better understand their relations to unplanned and planned drinking. Further, given the inevitable overlap between social and physical contexts, future research could extend findings from the present study by examining interactions between contexts (e.g., drinking at home while alone). The present study was limited to examining negative consequences and future research could consider including positive consequences of planned and unplanned drinking. We also showed descriptively that students endorsed planning to drink but ultimately did not do so on approximately 13% of study days; future research is needed to determine factors that influence individuals’ decisions to abstain from drinking after initially planning to do so (e.g., reasons for not drinking, mood).

Conclusion

The present study extended prior daily-level work examining motives (Stevens, Haikalis, et al., 2021) and contexts (Fairlie et al., 2019) of unplanned vs. planned drinking by assessing effects at both the occasion-level and daily-level, providing further nuance regarding these effects. We supported prior work by showing that drinking for social reasons or for enhancement was related to planned drinking. We contributed novel findings for coping motives and unplanned drinking, first postulated by the MUDB, and for offered motives and unplanned drinking. Unplanned drinking occasions were more likely to occur at home or alone, whereas planned drinking occasions were more likely to occur at a bar/restaurant, a party, with friends, with strangers/acquaintances, or with intoxicated people. Contrary to the MUDB and Fairlie et al. (2019), but corroborating Lauher et al. (2020), we found that planned drinking was linked with more consequences and a greater number of negative consequences. Findings suggest an important distinction between any unplanned drinking vs. unplanned heavy drinking. In future studies, we recommend that researchers consider the varying definitions of unplanned drinking (e.g., any unplanned, heavy unplanned, drinking more than intended, drinking faster than intended) to better improve our understanding of impaired control in daily life and inform in-the-moment harm-reduction interventions.

Public Health Significance:

This study identified that drinking for social reasons and to have fun is linked with planned drinking, whereas drinking to cope or because alcohol was offered is related to unplanned drinking. This study also showed several social and physical contexts that are associated with planned drinking (bar/restaurant, friend’s place, party, with friends, with intoxicated people, with strangers/acquaintances) and unplanned drinking (at home, alone, with roommates). These findings also highlighted that planned drinking events were related to more alcohol consumption and negative consequences, and the motives and contexts associated with planned drinking could be incorporated into in-the-moment harm-reduction interventions, in addition to targeting intentions to drink.