Pregame Heavy Episodic Drinking and Its Association With Negative Consequences and Other Risky Substance Use Behaviors

Abstract

Objective:

Pregaming, or drinking before going out, is common among college students and has been linked with greater alcohol use and experiencing more negative consequences. This study tested within- and between-person associations between pregame heavy episodic drinking (Pregame HED; 4+/5+ drinks for women/men while pregaming) and high-intensity drinking (8+/10+ drinks), negative consequences, and three risky behaviors.

Method:

College students at a large, public university in the Northeast United States who participated in a longitudinal measurement-burst design study completed a longer survey and up to 14 daily surveys in up to four consecutive semesters (n days = 4,706; n persons = 547). Hypotheses were primarily tested using logistic and Poisson multilevel models.

Results:

Pregame HED was reported by 41% of drinkers and on 15% of drinking days and 38% of pregaming days. Students were more likely to engage in high-intensity drinking on Pregame HED days than on moderate pregaming (1–3 and 1–4 pregaming drinks for women and men, respectively) or no pregaming drinking days. Students experienced more negative consequences on Pregame HED days than moderate or no pregaming drinking days, but there was no unique daily-level association between Pregame HED and negative consequences after alcohol intake was controlled. Students were more likely to use marijuana on Pregame HED days than on moderate and no pregaming drinking days.

Conclusions:

Pregame HED appears to be a characteristic of extremely heavy drinking days and fundamentally different from moderate pregaming and no pregaming drinking days. Findings highlight the importance of accounting for amounts of alcohol consumed while pregaming and the notion that drinking episodes can be dynamic.

Pregaming (also called predrinking, preloading, or prepartying), or drinking before going out, is normative within college drinking culture. Approximately 60%–70% of drinkers report past-month pregaming, and pregaming occurs on approximately one third of drinking days (Pedersen, 2016; Zamboanga & Olthuis, 2016). Numerous studies report between- and within-person associations between pregaming and greater alcohol intake and negative consequences (Pedersen, 2016; Zamboanga & Olthuis, 2016). For instance, Barnett et al. (2013) found that students consumed approximately two more drinks on pregaming versus nonpregaming drinking days, and earlier analyses from the present study found that students were more likely to reach high-intensity drinking thresholds (HID; 8+/10+ drinks for women/men) on pregaming days (Fairlie et al., 2015). Radomski et al. (2016) found that participants experienced nearly 50% more negative consequences on pregaming days. Merrill et al. (2013) reported similar findings when also controlling for same-day alcohol intake.

Limitations of past work on pregaming

Inferences drawn from many past studies on pregaming are mitigated by similar limitations. First, many relied on data collected using Timeline Followback (TLFB) assessments in which participants reported on drinking episodes days, weeks, or months later (e.g., Merrill et al., 2013; Radomski et al., 2016). Although TLFB is reliable and valid, reports of drinking collected daily may differ from those obtained via TLFB, and daily reports appear to minimize recall and social desirability bias and offer improved measurement validity (Dulin et al., 2017; Merrill et al., 2020; Patterson et al., 2019). Second, many day-level pregaming studies relied on small numbers of days per person and on single “bursts” of data collection (e.g., Fairlie et al., 2015). Third, much past work relied on majority White samples (e.g., Merrill et al., 2013, Radomski et al., 2016), leading to questions about generalizability to other groups. Many also used non-probability convenience samples, which may introduce selection effects.

Extant research has predominantly assessed correlates and consequences of any pregaming, that is, whether students ever pregamed or did so on specific days. Although these studies concluded that pregaming is risky, there is likely substantial variability in the amount of risk resulting from pregaming. Yet, reaching heavy episodic drinking (HED) thresholds (4+/5+ drinks for women/men) while pregaming, a behavior we refer to as Pregame HED, appears to be common based on descriptive statistics (Labhart et al., 2013; Read et al., 2010). Given that HED thresholds are intended to indicate an elevated risk of acute negative consequences if consumed across an entire drinking episode (Wechsler et al., 1995), drinking to this level before going out, then presumably more later, is likely associated with even greater risk. Although this may seem obvious, no studies to our knowledge have specifically tested such associations.

The need for considering how much students drink while pregaming

Since some work suggests that pregaming may be a substitute for later drinking (LaBrie et al., 2012; Read et al., 2010), it is important to determine whether drinking typically continues after pregaming, especially on Pregame HED days on which continued drinking may approach HID levels. HID has gained increased attention in the past decade as researchers have realized that many college students drink well above the HED threshold and the traditional dichotomous measure of HED does not capture this variability (Patrick, 2016; White et al., 2006). HID is associated with a greater likelihood of acute consequences than HED (Linden-Carmichael et al., 2018), and understanding more about the day-level factors that lead to such extreme drinking is important.

We posit that Pregame HED days may be meaningfully different from other types of pregaming and/or drinking days such that they may resemble other types of extremely risky drinking days, such as event-specific drinking days (e.g., 21st birthdays, holidays; Neighbors et al., 2006; Tremblay et al., 2010) in which drinking tends to be particularly heavy and may have somewhat distinct motives. If this is the case, it is important to know whether other risky behaviors are also more likely to occur. Drinking games are a common risky behavior associated with pregaming and heavy drinking (Zamboanga et al., 2014). Mixing alcohol with energy drinks has also been linked with pregaming as well as with greater harms (Linden-Carmichael & Lau-Barraco, 2017). Researchers have increasingly been interested in simultaneous and same-day co-use of alcohol and marijuana, as both are associated with greater drinking and negative consequences (e.g., Lee et al., 2020). To the extent that Pregame HED days are meaningfully different from other types of pregaming and/or drinking days, it would be important to know whether playing drinking games, mixing alcohol with energy drinks, and using marijuana are more likely on Pregame HED days than on more moderate pregaming days and non-pregaming drinking days.

Similarly, it would be important to determine whether there are certain types of students who are more likely to engage in Pregame HED or do so more often. Considerable research has shown that men, fraternity/sorority affiliates, and student athletes tend to engage in heavier, riskier drinking than women, those not affiliated with a fraternity or sorority, and non-athletes, respectively (Schulenberg et al., 2020; Turrisi et al., 2006). Men appear to drink greater quantities of alcohol than women on pregaming days (La-Brie & Pederson, 2008; Read et al., 2010) but do not seem to have higher pregaming prevalence (Rutledge et al., 2016). Students affiliated with fraternities or sororities (Haas et al., 2018) and student athletes (Mastroleo et al., 2019) tend to pregame more frequently than those not affiliated with fraternities and sororities and non-athletes, respectively. Students in honors colleges generally report less frequent and/or risky drinking than non–honors students (Lanza et al., 2010; Rhoades & Maggs, 2006), although we found no studies comparing pregaming between these groups.

Hypotheses

This study introduces the concept of Pregame HED, or women/men consuming 4+/5+ drinks while pregaming, as an indicator of more extreme pregame drinking and assesses associations of Pregame HED with same-day HID, acute negative consequences, and three risky substance use behaviors. This article avoids some of the methodological limitations of past pregaming work by using a large, diverse, probability-based sample collected using a longitudinal measurement-burst design with each participant completing up to 56 daily reports across four consecutive college semesters.

There were three daily-level, within-person hypotheses. On Pregame HED days compared with Moderate Pregaming days (1–3 pregaming drinks for women/1–4 pregaming drinks for men), we hypothesized that students would (a) be more likely to drink at the HID level, (b) experience more negative consequences, and (c) be more likely to play drinking games, mix alcohol with energy drinks, and use marijuana. Between persons, we hypothesized that Pregame HED counts would be higher for men compared with women, fraternity/sorority participants compared with non-participants, student athletes compared with non-athletes, and non–honors students compared with honors students.

Method

Participants and procedure

Data came from a longitudinal measurement-burst design study examining developmental change and fluctuations in risk behaviors and daily activities in students at a large, public university in the Northeast United States (Howard et al., 2015). Participants completed longer web-based surveys followed by 14 consecutive daily surveys in each of seven sequential semesters. The study was approved by the university’s institutional review board and protected by a federal National Institutes of Health Certificate of Confidentiality.

Participants were sampled using a stratified random sampling procedure. Eligible participants were (a) first-year, first-time, full-time students living within 25 miles of campus, (b) 21 years of age or younger, and (c) U.S. citizens or permanent residents. Of selected students, 66% (n = 744) provided informed consent and completed the semester survey and set of daily surveys in Semester 1. Retention was strong, with 79.6% (n = 592) of the initial sample completing at least one daily survey in the final (seventh) semester. Participants completed an average of 12.8 of 14 possible daily surveys per semester.

Pregaming was assessed beginning in Semester 4. Analyses included data from Semesters 4–7, or Spring of second year through Fall of fourth year, from a possible 56 days per student. In Semesters 4–7, the sample included 689 students and 34,384 person-days. This sample was 51.7% female; 179 (26.0%) were Hispanic/Latinx, 104 (15.1%) African American/Black Non-Hispanic/Latinx (NHL), 160 (23.2%) Asian American/Pacific Islander NHL, 185 (26.9%) European American NHL, and 61 (8.9%) multiracial NHL students. In Semester 4, the average age of participants was 19.94 years (SD = 0.42).

Measures

Drinks. On daily surveys, students reported the number of drinks consumed the previous day (Dimeff et al., 1999; Hustad & Carey, 2005). Students were instructed as follows: “By one drink we mean half an ounce of absolute alcohol, for example, [a] 12 ounce can or bottle of beer or cooler, [a] 5 ounce glass of wine, [or] a drink containing one shot of liquor or spirits.” Then, students were asked, “How many drinks of alcohol did you drink?” in reference to the previous day, using a pull-down menu (0 to 25+ drinks).

Pregaming variables. Each day students reported drinking, they were asked, “You said you had [xx] drinks on [previous day, e.g., Saturday]. How many of those were pre-gaming, that is, consumed before going out?” with pull-down menu options of 0 to the number of drinks previously reported. Three dichotomous pregaming variables were computed (Supplemental Table A). (Supplemental material appears as an online-only addendum to this article on the journal’s website.) Pregame Heavy Episodic Drinking (Pregame HED) indicated days students reached the HED threshold while pregaming. Moderate Pregaming indicated days students pregamed but did not reach the HED threshold while pregaming. No Pregaming indicated days students drank but did not pregame.

High-intensity drinking. A dichotomous HID variable was computed. On days women/men reported consuming 8+/10+ drinks, HID was coded as 1, versus 0 for all other days (Linden-Carmichael et al., 2018; Patrick, 2016).

Negative alcohol-related consequences. Each day students reported drinking, they were asked, “As a result of drinking alcohol on [previous day], did you …” followed by 11 negative consequences (Lee et al., 2011; Patrick & Maggs, 2011) (Supplemental Table B). These were summed to produce a count variable indicating the number of different negative consequences experienced each day.

Drinking games. Each day students reported drinking, they were asked, “Did you participate in any drinking games?” with response options yes (1) and no (0).

Mixing alcohol with energy drinks. Each day students reported drinking, they were asked, “On [previous day], how many (1) high energy (caffeinated) drinks like Red Bull, not containing alcohol, did you drink? (2) high energy drinks with alcohol (e.g., Red Bull + vodka, or a premixed drink) did you drink?” Responses were summed to create a dichotomous variable indicating any mixing of alcohol with energy drinks that day (1).

Marijuana use. Each day students were asked, “Did you use any illegal drugs on [previous day]?” On days students reported using illegal drugs, they endorsed the specific substances they used (check all that apply). Days of marijuana use (1) were contrasted with all sampled days marijuana was not used (0).

Fraternity/sorority participation. Each semester, students reported their extracurricular activities (e.g., fraternity/sorority [social], intercollegiate athletics). Fraternity/sorority participation was coded as 1 for students who ever participated in a fraternity or sorority and 0 for all others.

Student athlete status. Similarly, students who reported ever participating in intercollegiate athletics were coded as 1, and students who never participated were coded as 0.

Honors college enrollment. Each semester, students were asked, “Are you in the [name/redacted] academic honors program at [name of university]?” This program is highly selective and comprises approximately 2% of the student body. Students who were ever in the academic honors program were coded as 1, and students who were never in the program were coded as 0.

Statistical analyses

Three-level logistic and Poisson multilevel models nesting days within semesters within persons were used to test the three daily-level, within-person hypotheses concerning associations between Pregame HED and HID, negative consequences, and risky substance use behaviors. Models were estimated using maximum likelihood estimation based on the Laplace Approximation in the lme4 package (Bates et al., 2015) of R 4.1.2 (R Core Team, 2021). A daily-level random variable captured overdispersion in Poisson models. Daily-level variables were semester-mean-centered, semester-level variables were person-mean-centered, and person-level variables were grand-mean-centered. A semester number variable was included to account for any change in pregaming or the outcome variables across semesters (Wang & Maxwell, 2015), and a daily-level social weekend variable (0 = Sunday–Wednesday, 1 = Thursday–Saturday) accounted for most heavy drinking occurring on the “social weekend” (i.e., Thursdays–Saturdays; Del Boca et al., 2004; Finlay et al., 2012; Maggs et al., 2011). Between-person hypotheses concerning group differences in counts of Pregame HED were tested using a negative binomial regression in the MASS (Venables & Ripley, 2002) R package. All models controlled for race/ethnicity.

Results

Descriptive statistics

Drinking was reported on 4,706 (13.7%) of the 34,384 days sampled. In total, 547 (79.4%) of the 689 students reported drinking on at least one of the up to 56 sampled days. The analytic sample consisted of 4,706 drinking days nested within 1,529 person-semesters nested within 547 drinkers. Pregaming was reported at least once by 387 students (70.7% of drinkers) and on 1,934 (41.1%) drinking days. Pregame HED was reported at least once by 226 students (41.3% of drinkers) and on 727 days (37.6% of pregaming days; 15.4% of drinking days) (Table 1). Students consumed an average of 10.10 (SD = 4.17) total drinks on Pregame HED days, 6.21 (SD = 3.33) on Moderate Pregaming days, and 4.68 (SD = 3.76) on No Pregaming drinking days. The mean number of drinks consumed after pregaming was 3.63 (SD = 3.24) on Pregame HED days and 4.08 (SD = 3.10) on Moderate Pregaming days. Thus, on Pregame HED days, roughly two thirds of total drinks were consumed while pre-gaming; on Moderate Pregaming days, roughly one third of total drinks were consumed while pregaming.

Table 1.

Descriptive statistics

Variable	n (%)	M (SD)	Min.	Max.
Daily level (n = 4,706 drinking days)
Pregame HED	727 (15.4)	-	0	1
Drinks	-	5.96 (4.21)	1	26
Negative consequences	-	1.02 (1.70)	0	11
Played drinking games	1,512 (32.1)	-	0	1
Mixed alcohol with energy drinks	452 (9.6)	-	0	1
Used marijuana	424 (9.0)	-	0	1
Person level (n = 547 students who reported drinking)
Male	253 (46.3)	-	0	1
Fraternity/sorority participant	114(20.8)	-	0	1
Student athlete	21 (3.8)	-	0	1
Honors student	46 (8.4)	-	0	1
Race/ethnicity
Asian NHL	110 (20.1)	-	0	1
Black NHL	75 (13.7)	-	0	1
Multiracial NHL	44 (8.0)	-	0	1
White NHL	162 (29.6)	-	0	1
Hispanic/Latinx	156 (28.5)	-	0	1

Notes: Min. = minimum; max. = maximum; HED = heavy episodic drinking; NHL = non-Hispanic/Latinx.

Within-person Hypothesis 1: Was pregame HED associated with same-day HID?

Compared with Moderate Pregaming days, students had 5.06 greater odds of engaging in HID on Pregame HED days, γ₁₀₀, and .55 lower odds of HID on No Pregaming drinking days, γ₂₀₀ (Table 2, Model 1). Thus, the likelihood of HID on a given drinking day was higher at greater levels of pregame drinking. Semesters in which students reported more Pregame HED days tended to have more HID days, γ₀₁₀, and students reporting more Pregame HED days across all four semesters tended to also report more HID days, γ₀₀₁.

Table 2.

Logistic multilevel model testing whether high-intensity drinking was more likely on pregame heavy episodic versus moderate pregaming days

Fixed effects	Model 1 Outcome: High-intensity drinking HID) OR [95% CI]
Daily level
Intercept, γ000	0.08 [0.06, 0.11]***
Pregame HED day, a γ100	5.06 [3.57, 7.18]***
No pregaming day, a γ200	0.45 [0.34, 0.60]***
Social weekend day, γ300	3.15 [2.36, 4.21]***
Semester level
Semester-mean pregame HED, γ010	7.52 [3.69, 15.31]***
Semester-mean no pregaming, γ020	0.65 [0.37, 1.14]
Semester number, γ030	0.93 [0.85, 1.03]
Person level
Person-mean pregame HED, γ001	196.66 [62.05, 623.26]***
Person-mean no pregaming, γ002	0.38 [0.16, 0.88]*
Male γ003	1.92 [1.30, 2.82]***
Fraternity/sorority participant, γ004	1.77 [1.13, 2.76]*
Student athlete, γ005	3.68 [1.39, 9.75]**
Honors student, γ006	0.31 [0.14, 0.68]**
Race/ethnicity (ref.: White NHL)
Asian NHL, γ007	0.60 [0.37, 0.95]*
Black NHL, γ008	0.16 [0.08, 0.32]***
Multiracial NHL, γ009	0.35 [0.19, 0.62]***
Hispanic/Latinx, γ0010	0.59 [0.29, 1.22]

Notes: n = 4,454 drinking days within 521 students. OR = odds ratio; CI = confidence interval; social weekend day = Thursday, Friday, or Saturday (in comparison to Sunday through Wednesday); HED = heavy episodic drinking; pregame HED = women/men consuming 4+/5+ drinks while pregaming (i.e., before “going out”); HID = women/men consuming 8+/10+ drinks; NHL = non-Hispanic/Latinx; ref. = reference.

^aReference group is moderate pregaming days, or women/men consuming 1-3/1-4 drinks before going out.

^*p < .05;

^**p < .01;

^***p < .001.

Within-person Hypothesis 2: Was pregame HED associated with greater same-day negative consequences?

In a model not controlling for alcohol intake (Table 3, Model 2), students experienced 68% more negative consequences on Pregame HED days, γ₁₀₀, and 35% fewer consequences on No Pregaming drinking days, γ₂₀₀, compared with Moderate Pregaming days. The average student was expected to experience 0.71 consequences on the average Pregame HED day, 0.42 on the average Moderate Pregaming day, and 0.27 on the average No Pregaming day when holding all other variables at their mean. Thus, the expected number of negative consequences experienced on a given drinking day was greater at higher levels of pregame drinking. There was no significant within-person association between Pregame HED and negative consequences at the semester level, γ₀₁₀, but students who engaged in Pregame HED more often reported more negative consequences throughout the study, γ₀₀₁.

Table 3.

Poisson multilevel models testing whether students experienced more negative alcohol- related consequences on pregame heavy episodic versus moderate pregaming days

Fixed effects	Outcome: Total negative consequences
	Model 2: Without controlling for alcohol intake IRR [95% CI]	Model 3: Controlling for alcohol intake IRR [95% CI]
Daily level
Intercept, γ000	0.41 [0.36, 0.46]***	0.39 [0.35, 0.44]***
Pregame HED day, a γ100	1.68 [1.44, 1.97]***	1.12 [0.97, 1.30]
No pregaming day, a γ200	0.65 [0.57, 0.74]***	0.82 [0.73, 0.93]**
Total drinks, γ300	-	1.21 [1.19, 1.23]***
Social weekend day, γ400	1.26 [1.10, 1.43]***	0.96 [0.85, 1.09]
Semester level
Semester-mean pregame HED, γ010	1.22 [0.87, 1.71]	0.78 [0.57, 1.06]
Semester-mean no pregaming, γ020	0.72 [0.56, 0.93]*	0.83 [0.66, 1.05]
Semester-mean total drinks, γ030	-	1.24 [1.20, 1.28]***
Semester number, γ040	0.92 [0.87, 0.96]***	0.92 [0.88, 0.96]***
Person level
Person-mean pregame HED, γ001	2.62 [1.43, 4.79]**	0.65 [0.33, 1.27]
Person-mean no pregaming, γ002	0.48 [0.31, 0.75]**	0.59 [0.39, 0.91]*
Person-mean total drinks, γ003	-	1.19 [1.14, 1.24]***
Male, γ004	1.28 [1.04, 1.59]*	0.91 [0.73, 1.13]
Fraternity/sorority participant, γ005	1.25 [0.98, 1.60]	1.04 [0.81, 1.32]
Student athlete, γ006	1.12 [0.64, 1.95]	0.82 [0.47, 1.41]
Honors student, γ007	0.63 [0.42, 0.93]*	0.79 [0.54, 1.17]
Race/ethnicity (ref.: White NHL)
Asian NHL, γ008	0.88 [0.68, 1.15]	0.99 [0.76, 1.29]
Black NHL, γ009	0.85 [0.60, 1.20]	1.28 [0.90, 1.82]
Multiracial NHL, γ0010	0.91 [0.67, 1.23]	1.12 [0.83, 1.50]
Hispanic/Latinx, γ0011	0.90 [0.53, 1.20]	0.88 [0.59, 1.31]

Notes: n= 4,420 days within 517 persons. IRR = incidence rate ratio; CI = confidence interval; HED = heavy episodic drinking; social weekend day = Thursday, Friday, or Saturday (in comparison to Sunday through Wednesday); pregame HED = women/men consuming 4+/5+ drinks while pregaming (i.e., before “going out”); NHL = non-Hispanic/Latinx.

^aReference group is moderate pregaming days, or women/men consuming 1-3/1-4 drinks before going out.

^*p < .05;

^**p < .01;

^***p < .001.

After we controlled for alcohol intake (Table 3, Model 3), there was no significant within-person association between Pregame HED and negative consequences at the daily level, γ₁₀₀, but students experienced 18% fewer consequences on No Pregaming drinking days versus Moderate Pregaming days, γ₂₀₀. Thus, the association between Pregame HED and greater negative consequences appeared to be largely attributable to heavier drinking on these days rather than to Pregame HED itself. Alcohol intake was positively associated with the average number of consequences experienced on drinking days at all three levels (γ₃₀₀, γ₀₃₀, γ₀₀₃).

Within-person Hypothesis 3: Were other risky substance use behaviors more likely on pregame HED days?

Students were no more likely to play drinking games on Pregame HED days versus Moderate Pregaming days, γ₁₀₀, but they were less likely to play drinking games on No Pregaming drinking days than on Moderate Pregaming days, γ₂₀₀ (Table 4, Model 4). Similarly, students were no more likely to mix alcohol with energy drinks on Pregame HED days versus Moderate Pregaming days, γ₁₀₀, but they were less likely to mix alcohol with energy drinks on No Pregaming drinking days than on Moderate Pregaming days, γ₂₀₀ (Table 4, Model 5). Students were more likely to use marijuana on Pregame HED days versus Moderate Pregaming days, γ₁₀₀, but they were no more likely to use marijuana on No Pregaming drinking days than on Moderate Pregaming days, γ₂₀₀ (Table 4, Model 6). There was no evidence of within-person, semester-level associations, γ₀₁₀, or between-person associations, γ₀₀₁, between number of Pregame HED days and the likelihood of engaging in the three risky substance use behaviors (Models 4–6).

Table 4.

Logistic multilevel models testing whether students were more likely to engage in risky substance use behaviors on pregame heavy episodic versus moderate pregaming days

Fixed effects	Model 4 Outcome: Any drinking games (n = 4,439 days) OR [95% CI]	Model 5 Outcome: Any alcohol mixed with energy drinks (n = 4,451 days) OR [95% CI]	Model 6 Outcome: Any marijuana use (n = 4,448 days) OR [95% CI]
Daily level
Intercept,γ000	0.32 [0.28, 0.37]***	0.04 [0.03, 0.05]***	<0.01 [<0.01, <0.01]***
Pregame HED day,a γ100	0.99 [0.74, 1.32]	0.93 [0.60, 1.44]	1.88 [1.07, 3.29]*
No pregaming day,a γ200	0.78 [0.62, 0.97]*	0.53 [0.38, 0.76]***	1.48 [0.93, 2.34]
Social weekend day, γ300	2.91 [2.29, 3.69]***	1.05 [0.75, 1.46]	1.46 [0.96, 2.22]
Semester level
Semester-mean pregame HED, γ010	1.43 [0.80, 2.54]	1.45 [0.57, 3.73]	0.75 [0.14, 3.98]
Semester-mean no pregaming, γ020	1.09 [0.73, 1.64]	0.49 [0.25, 0.95]*	0.99 [0.29, 3.40]
Semester number, γ030	0.68 [0.63, 0.74]***	1.04 [0.91, 1.18]	0.87 [0.69, 1.11]
Person level
Person-mean pregame HED, γ001	1.84 [0.86, 3.93]	1.93 [0.65, 5.72]	2.80 [0.23, 33.46]
Person-mean no pregaming, γ002	0.36 [0.21, 0.62]***	0.73 [0.33, 1.62]	0.60 [0.09, 3.82]
Male, γ003	1.39 [1.08, 1.79]*	1.15 [0.79, 1.66]	2.17 [0.87, 5.43]
Fraternity/sorority participant, γ004	1.28 [0.96, 1.73]	1.17 [0.77, 1.77]	1.46 [0.51, 4.16]
Student athlete, γ005	0.98 [0.49, 1.97]	4.04 [1.71, 9.56]**	-
Honors student, γ006	0.75 [0.47, 1.20]	0.38 [0.18, 0.80]*	1.17 [0.21, 6.49]
Race/ethnicity (ref.: White NHL)
Asian NHL, γ007	0.89 [0.65, 1.23]	0.97 [0.61, 1.55]	2.69 [0.83, 8.68]
Black NHL, γ008	0.38 [0.24, 0.59]***	1.38 [0.75, 2.54]	4.60 [1.08, 19.64]*
Multiracial NHL, γ009	1.20 [0.83, 1.72]	1.64 [0.98, 2.77]	1.00 [0.23, 4.26]
Hispanic/Latinx, γ0010	0.66 [0.40, 1.08]	1.39 [0.70, 2.75]	2.18 [0.37, 12.77]

Notes: n = 520-521 students. Student athlete status was not included in the model predicting any marijuana use as only student athlete reported using marijuana. OR = odds ratio; CI = confidence interval; HED = heavy episodic drinking; social weekend day = Thursday, Friday, or Saturday (in comparison to Sunday through Wednesday); pregame HED = women/men consuming 4+/5+ drinks while pregaming; ref. = reference; NHL = non-Hispanic/Latinx.

^aReference group is moderate pregaming days, or women/men consuming 1-3/1-4 drinks before going out.

^*p < .05;

^**p < .01;

^***p < .001.

Between-person hypotheses: Were there group differences in counts of pregame HED?

The number of Pregame HED days reported across all sampled days did not differ between men and women, fraternity/sorority participants and non-participants, student athletes and non-athletes, or honors and non–honors students (Supplemental Table C).

Discussion

This article introduced the concept of Pregame HED and found that students were more likely to engage in HID, experienced more negative consequences, and were more likely to use marijuana on Pregame HED days than on moderate and non-pregaming drinking days. Descriptively, most drinking occurred while pregaming on Pregame HED days and after pregaming on moderate pregaming days. Thus, Pregame HED days appear to be fundamentally different from moderate pregaming and non-pregaming drinking days.

Pregame HED may be common among college students

Approximately two fifths of drinkers engaged in Pregame HED at least once across sampled days, and Pregame HED occurred on more than one third of pregaming days and approximately one sixth of drinking days. This indicates that Pregame HED was common among drinkers in this sample. Consistent with past research (Pedersen, 2016; Zamboanga & Olthuis, 2016), any pregaming was relatively normative in this sample, being reported by more than two thirds of drinkers and on more than two fifths of drinking days.

Amount of alcohol consumed while pregaming matters

This study’s most important contribution is showing that on days in which students consume 4+/5+ drinks before going out, they are more likely to drink at the clearly hazardous level of 8+/10+ drinks and to report more negative consequences compared with days they pregame more moderately and with non-pregaming drinking days. Any pregaming is consistently linked with greater alcohol intake and increased negative consequences (e.g., Pedersen, 2016; Zamboanga & Olthuis, 2016), and the quantity of alcohol consumed per drinking occasion is positively linked with the number of consequences experienced (Gruenewald & Mair, 2015). To our knowledge, this study provides the first evidence that the amount of alcohol consumed while pregaming predicts consuming more alcohol overall and experiencing more consequences that day. It also connects Pregame HED with a particularly risky drinking behavior, HID, that is of increasing interest to researchers studying adolescent and young adult drinking (e.g., Patrick et al., 2016; Patrick & Terry-McElrath, 2017).

Critics might argue that it is circular reasoning to predict an overall level of drinking on a given day based on the level of drinking in a particular context the same day. However, in prior work, some students reported pregaming as a substitute for drinking at the primary event because alcohol might be expensive, difficult to obtain, or unsafe (LaBrie et al., 2012; Read et al., 2010). If these pre-drinking motives are common, they should lead to less drinking after going out, on average. Yet, two studies suggest that pregaming tends not to substitute for later drinking but rather is characteristic of heavy drinking days (Pedersen & LaBrie, 2007; Read et al., 2010). The present results support the latter findings. Although substitution pregaming motives do occur, it seems to be more common for drinking to continue after pregaming and for Pregame HED days to resemble event-specific drinking days like 21st birthdays and certain holidays (Neighbors et al., 2006; Tremblay et al., 2010). Thus, Pregame HED may be a method that drinkers use (consciously or not) to achieve very high levels of alcohol intake.

Pregaming co-occurs with other risky substance use behaviors

Pregaming was associated with three risky substance use behaviors, albeit in various ways. First, although Pregame HED was not specifically associated with playing drinking games or mixing alcohol with energy drinks, students were more likely to engage in both on any pregaming days than on non-pregaming drinking days, consistent with previous studies (Fairlie et al., 2015; Linden-Carmichael & Lau-Barraco, 2017; Read et al., 2010). Second, students were more likely to use marijuana on Pregame HED than on moderate pregaming and non-pregaming drinking days. Although we could not differentiate co-use from simultaneous alcohol and marijuana use days, our findings are consistent with research showing that drinking tends to be greater on days both alcohol and marijuana are used (e.g., Gunn et al., 2018; Lee et al., 2020).

Prevention/intervention implications

One prevention/intervention implication of these findings is that substantial amounts of alcohol may be consumed before students go out. This means students may be at risk for harms when traveling to or arriving at a bar or party on Pregame HED days. It could be helpful for prevention scientists to develop protective behavioral strategies that focus on the transitions between contexts. Similarly, these findings, along with the rest of the pregaming literature, highlight the dynamic nature of drinking episodes that may occur in multiple physical and social contexts. Changes in context may both pose risks and be potential targets for prevention efforts.

Limitations

This study has several strengths compared with prior pregaming work. First, probability-based sampling aimed to obtain a sample that was balanced regarding sex and race/ethnicity. Second, within-person associations were tested using a large number of days per person and data collected daily; most past pregaming research has included significantly fewer daily measurements (e.g., Fairlie et al., 2015) or has used TLFB (e.g., Radomski et al., 2016), which may be more prone to recall and social desirability bias and have lower measurement validity (Dulin et al., 2017; Patterson et al., 2019).

Limitations are also present. First, the external validity is unclear. Data came from a large, public university in the Northeast United States; it is unknown how findings generalize to students in other types of institutions and locations. Data were collected in 2-week bursts during the academic year; it is unclear to what extent these weeks generalize across the year (e.g., winter/summer break). Second, data were collected from late sophomore through early senior year. Notably, these analyses do not include data from the first year, when many risky drinking behaviors occur. Most participants submitted daily surveys in semesters when they were underage and in others when they were of legal drinking age. Some past work has shown that age is negatively associated with odds of pregaming among young adults (e.g., Read et al., 2010; Rutledge et al., 2016), and some young adults pregame because alcohol may be difficult to obtain (e.g., because they are underage) at the primary event (LaBrie et al., 2012). Thus, it is possible that the associations reported here may be moderated by age. Examination of such moderating effects is beyond the scope of this work but is warranted in future research. Third, although the response rate was high, students may have skipped some surveys following very heavy drinking days (e.g., HID days). Finally, when drinking at such high levels, students may have had difficulty with accurate recall of total drinks, pregame drinks, and other risky substance use behaviors.

Future directions

Many important questions regarding pregaming and Pre-game HED remain unanswered. First, students experienced more acute consequences on Pregame HED versus moderate pregaming days. Future research should ask whether Pregame HED predicts greater longer-term consequences (e.g., Zucker et al., 2006). Second, the focus on pregaming highlights how some drinking episodes consist of a progression of drinking locations. Future work could better track changes in physical location, modes of transportation to and from each location, and changes in social context. This work could identify effectively targetable intervention points. Third, evidence-based interventions like BASICS (Dimeff et al., 1999) encourage students to use protective behavioral strategies (e.g., go home with friends, designated driver). Whether students use these strategies on Pregame HED days, what barriers exist, and whether Pregame HED amplifies or reduces their effectiveness is unknown.