Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 May 1.
Published in final edited form as: Alcohol Clin Exp Res (Hoboken). 2024 Apr 1;48(5):955–966. doi: 10.1111/acer.15307

Is the 21st birthday a turning point for alcohol and cannabis use? A monthly study of young adults

Isaac C Rhew 1, Michael S Gilson 1, Charles B Fleming 1, Katherine Walukevich-Dienst 1, Katarina Guttmannova 1, Megan E Patrick 2, Christine M Lee 1
PMCID: PMC11260108  NIHMSID: NIHMS2009772  PMID: 38558408

Abstract

Background:

An important life-course event with respect to alcohol and cannabis use is turning 21 years of age, which may be associated with increases in use of these substances due to celebrations during the month and easier access to them on and following this birthday. We examined the trajectories of alcohol and cannabis use behaviors in the months leading up to, during, and following the 21st birthday month. We also examined whether the use trajectories vary by college status and baseline levels of use.

Methods:

We used data from 203 young adults recruited from the Greater Seattle region who turned 21 during the course of the study. Surveys were administered each month for 24 consecutive months. Measures included the typical number of drinks per week for the past month, the frequency of heavy episodic drinking, the number of cannabis use days, and any simultaneous alcohol and cannabis use. Multilevel spline models were run that estimated linear slopes over time at four intervals: (1) up to 1 month before the 21st birthday month; (2) from 1 month before to the month of the 21st birthday; (3) from the 21st birthday month to 1 month following; and (4) from 1 month following the 21st birthday month through all following months.

Results:

Alcohol use, generally, and simultaneous alcohol and cannabis use showed sharp increases from the month before the 21st birthday month to the 21st birthday month and decreases following the 21st birthday month. For cannabis use, there were significant increases in the months leading up to the 21st birthday and no other significant changes during other time intervals. Patterns differed by baseline substance use and college status.

Conclusions:

Findings from the current study have implications for the timing and personalization of prevention and intervention efforts. Event-specific 21st birthday interventions may benefit from incorporating content targeting specific hazardous drinking behaviors in the month prior to the 21st birthday.

Keywords: 21st birthday, alcohol, cannabis, minimum legal age, young adulthood

INTRODUCTION

Alcohol use, including heavier forms of use (i.e., heavy episodic drinking [also referred to as binge drinking] and high intensity drinking), is widespread among young adults. Findings from the 2022 national Monitoring the Future (MTF) study suggests that 30.5% of young adults ages 19–30 reported heavy episodic drinking in the past 2 weeks and 4.6% reported drinking alcohol daily (Patrick et al., 2023). MTF data also suggest that the prevalence of cannabis use among young adults has reached its highest levels ever recorded with an over 12% increase in prevalence of any past month cannabis use (16.6% in 2012 to 28.8% in 2022) and doubling of daily use (5.6% in 2012 to 11.3% in 2022) over the past decade. Increases in cannabis use among young adults have been observed in other national studies as well (Substance Abuse and Mental Health Services Administration, 2021). The use of alcohol and cannabis tend to peak within individuals in their early 20s (Patrick et al., 2019; Substance Abuse and Mental Health Services Administration, 2021).

An important life-course event with respect to alcohol and cannabis use is turning 21 years of age. Since 1984, age 21 has been the legal drinking age across the United States and, since 2012, nonmedical cannabis use has also been legalized for adults age 21+ in some states. Turning 21 may be associated with increases in alcohol and cannabis use due to celebrations during the 21st birthday month as well as easier access on and following the 21st birthday.

Multiple studies have shown that 21st birthday celebrations are occasions for “event-specific drinking” (e.g., Neighbors et al., 2005) defined as drinking occasions where an individual often exceeds their typical maximum number of drinks and is at heightened risk for experiencing alcohol-related negative consequences (Brister et al., 2011; Lewis et al., 2009; Rutledge et al., 2008). Although this “21st birthday effect” (i.e., alcohol use frequency and quantity increasing as part of the birthday celebration) has been most commonly studied and verified in college populations, a recent study found that this effect was similar in noncollege populations as well (Gilson et al., 2022). Thus, despite the fact that alcohol use is more prevalent and heavier in college compared to noncollege populations (Patrick et al., 2023), use surrounding 21st birthdays for the two populations is of considerable interest.

Less is known about whether a similar “21st birthday effect” exists for cannabis, particularly with nonmedical (or “recreational”) use now legal in many states. Previous research has suggested that cannabis use is significantly higher on certain high-risk events such as 4/20 (April 20th), Mardi Gras, and St. Patrick’s Day (Bravo et al., 2017; Buckner et al., 2018; Walukevich-Dienst & Buckner, 2019). Among states where cannabis has been legalized, 21st birthdays represent the first opportunity for young adults to legally purchase and use cannabis. To our knowledge, only one study has examined cannabis use specifically during 21st birthday celebrations, and no “21st birthday effect” was observed (Gilson et al., 2022).

Although there have been multiple studies specifically on drinking on the 21st birthday as a high-risk event, little is known about changes in alcohol use in the months leading up to and, importantly, following the 21st birthday when it is legal to purchase and use alcohol. Even less is known about changes in cannabis use surrounding the 21st birthday despite the fact that cannabis is now legal in many states. Many studies have shown that prevalence of alcohol and cannabis use and heavy use tend to peak in the early 20s and then decline thereafter (Patrick et al., 2019, 2021; Terry-McElrath et al., 2018). More specific to the 21st birthday, there is some evidence that turning 21 may be a turning point in hazardous drinking behaviors. A longitudinal study of college students by Fromme and colleagues showed increases in alcohol frequency and quantity between the ages of 18 and 21, but decreases in quantity from 21 to 23 (Fromme et al., 2010). They also found prevalence of alcohol-related health risk behaviors (such as driving after drinking) increased significantly after individuals turned 21, suggesting that despite drinking a lower quantity of alcohol over time, young adults may be more willing to engage in hazardous alcohol-involved practices that could put them at risk for severe negative consequences. Another study found that the 21st birthday celebration itself may impact subsequent alcohol use such that greater alcohol consumption during 21st birthday celebrations was prospectively associated with more alcohol use and negative drinking consequences 1 year later (Geisner et al., 2017). Of note, this study also found that post-21st-birthday increases were greatest for those who reported less frequent alcohol use prior to 21st birthday, which may speak to differential trajectories depending on pre-21st-birthday use. Taken together, these studies suggest that the importance of 21st birthdays for alcohol use is not limited solely to the potential negative consequences experienced during the event itself but may also have important implications for subsequent trajectories of alcohol use.

Few existing studies have examined whether turning 21 alters trajectories of cannabis use. Most studies used repeated cross-sectional designs with data collected prior to when most states legalized nonmedical cannabis, and findings were mixed (Crost & Guerrero, 2012; Crost & Rees, 2013; Yoruk & Yoruk, 2013). Importantly, these studies may not accurately reflect current individual-level cannabis trajectories among young adults, especially in states where nonmedical cannabis is legalized. As with alcohol, cannabis use—particularly in states where it is legally used starting at age 21—may increase sharply around the time of the 21st birthday celebration and elevated use may extend into subsequent months due to easier access. More recent research suggests that legalization may be associated with increased prevalence of cannabis use among young adults, which may be particularly true when restricted to those aged 21 years or older (Bae & Kerr, 2020; Kilmer et al., 2022). More research is needed to understand whether cannabis trajectories when young adults age into legal access to cannabis. Further, it will be important to consider how other specific high-risk substance use behaviors change around the time of the 21st birthday, including both heavy episodic drinking and simultaneous alcohol and cannabis use (often referred to as simultaneous alcohol and marijuana use, or SAM). Importantly, SAM use may have negative consequences beyond the additive effects of the two substances (Fleming et al., 2021; Patrick, Cronce, et al., 2016).

The effects of turning 21 on alcohol and cannabis use may vary for different subpopulations of young adults. A better understanding of how trajectories of use following the 21st birthday may differ may by sociodemographic characteristics may help identify individuals at higher risk of increasing and maintaining alcohol and cannabis use after turning 21. An important social environmental context for alcohol use may be 4-year college, where alcohol use is common and often part of campus social culture (Patrick, Terry-McElrath, et al., 2016; White & Hingson, 2013). Although the gap is closing (Patrick et al., 2022), research has consistently shown elevated prevalence and frequency of alcohol use among young adults attending 4-year college compared to those not in college (Merrill & Carey, 2016). It is possible that the college context may also shape trajectories of use following the 21st birthday. Further, 21st birthday may be a particular turning point for use for those engaged in lower levels of substance use prior to the 21st birthday as suggested in the study from Geisner and colleagues described earlier (Geisner et al., 2017).

To better understand how turning 21 shapes trajectories of alcohol and cannabis use and their simultaneous use and assess whether there are both short-term “celebratory” elevations in use in the 21st birthday month and longer-term elevations in use due to increased access, the present study takes advantage of a longitudinal study of young adults from the Greater Seattle, Washington region who were surveyed monthly for two consecutive years. Cannabis use has been legalized in Washington State for adults over age 21 since 2012 and, thus, this setting provides an opportunity to examine associations with turning 21 in the context of legalized nonmedical cannabis. Using monthly data from participants who turned 21 during the course study period, the present study examines variation in trajectories leading up to, during, and following the 21st birthday month. Further, we assess whether these trajectories vary by college status at the time of 21st birthday and earlier levels of substance use.

MATERIALS AND METHODS

Study sample and procedures

Data for this current study were from Project Transitions, a monthly study of social role transitions and substance use across early adulthood. Between February 2015 and January 2016, young adults living in the greater Seattle, Washington area, were recruited for participation via online and social media advertising (e.g., Craigslist, Facebook), flyering, advertisements in local and school newspapers, and other outreach activities (Patrick et al., 2018). Interested individuals completed a brief online eligibility survey. Eligibility criteria for the parent study included being 18 to 23 years old, living within the greater Seattle metropolitan area (i.e., 60 miles from study offices), reporting drinking alcohol at least once in the last year, having a valid email address and being willing to meet in-person at the study office for an initial appointment. During this 1.5–2 h in-person appointment at the research project site, young adults completed age verification and informed consent with study staff and then completed a baseline assessment on study computers. For the next consecutive 24 months, beginning the 1st day of the month, participants were invited to complete an online monthly survey about social role transitions and alcohol use during the previous calendar month.

Out of 1644 who met the eligibility criteria, 778 young adults completed their initial appointment and baseline survey and started the monthly assessments. Retention was high throughout the 24 months of the study, with monthly completion rates ranging from 78.4% to 97.7%. As compensation for their time, participants received $40 in Amazon gift cards for the baseline assessment and between $20 and $45 for monthly surveys depending on survey length. The study had university Institutional Review Board approval and no adverse events occurred.

Because of interest in changes in alcohol and cannabis use before and after the 21st birthday, the analytic sample for the current study only included those participants who turned 21 over the course of the study (n = 212). Further, primary analyses were restricted to those who reported the use of respective substances during at least 1 month of the study period (n = 203 for typical drinking and heavy episodic drinking; n = 145 for cannabis; n = 141 for SAM). Distributions of demographic and other characteristics for the sample are shown in Table 1.

TABLE 1.

Distribution of participant sociodemographic characteristics and past month substance outcomes.

Mean (SD) or %
N = 203		 Range
Age at baseline 20.0 (0.6) 19.1, 20.9
Female sex assigned at birth 53.2
Race
 White 60.9
 Asian 17.3
 Non-White and non-Asiana 21.8
Hispanic ethnicity 5.5
Minoritized sexual identity 22.8
Attending 4-year college at 21st birthday 57.1
Typical drinks per weekb 5.1 (8.1) 0, 50
Number of days of cannabis usec 4.4 (8.7) 0, 30
Frequency of heavy episodic drinkingb
 Never 40.6
 Once per month 22.7
 2–3days per month 16.4
 At least weekly 20.3
Any simultaneous alcohol and cannabis usec 31.0
Open in a new tab
a

Includes participants identifying as Black, Native Hawaiian or Pacific Islander, Middle Eastern, Multiracial, and Other race.

b

Averaged across all monthly observations among those who reported at least one typical drink per week during at least 1 month of the study period (n = 203).

c

Averaged across all monthly observations among those who reported cannabis use during at least 1 month of the study period (n = 145).

Monthly study measures

21st birthday

At screening, participants provided their date of birth. Using this information, we were able to assess the participant’s age at each study assessment, including the calendar month of the participant’s 21st birthday.

Typical drinks per week

Participants were administered a modified Daily Drinking Questionnaire (Collins et al., 1985). In reference to the prior calendar month, participants were asked, “How much alcohol (measured in number of standard drinks), on average, did you drink each day of a typical week?” The total number of drinks in a typical week was calculated as a sum score. For analyses, we recoded the maximum possible value to 50.

Heavy episodic drinking

To assess the frequency of heavy episodic drinking, female participants were asked how often they had 4 or more drinks and male participants were asked how often they had 5 or more drinks during a 2-h period in the past calendar month. Response options were ordinal categories and ranged from 0 (never) to 7 (every day). Due to relatively sparse numbers of participants reporting 1 or more heavy episodic drinking episodes per week, the variable was recategorized from 0 to 3, where 0 was never, 1 was 1 day, 2 was 2 to 3 days, and 3 was 1 or more days a week in the past month.

Cannabis use frequency

In reference to the prior calendar month, participants were asked to report the number of days in the past 30 days that they used marijuana.

Simultaneous alcohol and cannabis use

For participants who reported any cannabis use in the prior calendar month, they were asked to report the number of times during that prior calendar month that they used cannabis at the same time as alcohol “so that their effects overlapped.” Response options were “not at all” (0), “a few of the times” (1), “some of the times” (2), “most of the times” (3), and “every time” (4). For ease of interpretation, we dichotomized the variable to any simultaneous use (1) versus no simultaneous use (0); the latter category also included participants who reported no cannabis use.

Demographic characteristics

We obtained demographic information at baseline, including sex assigned at birth (0: female, 1: male), sexual identity (0: heterosexual, 1: minoritized sexual identity [bisexual, gay, lesbian, queer, questioning, or other]), race (White [ref], Asian, not White or Asian [Black, American Indian or Alaska Native, Pacific Islander, Middle Eastern, Multiracial, other race]), Hispanic ethnicity (0: non-Hispanic, 1: Hispanic), and parents’ highest level of education (0: less than 4-year college degree, 1: 4-year college degree, 2: graduate or professional degree). Although not expected to be associated with timing around 21st birthday, these covariates were included to improve precision of estimates because of their associations with substance use due to various social factors. Further, we assessed participants’ 4-year college status at time of 21st birthday (0: not currently attending 4-year college, 1: attending 4-year college) in monthly surveys as a time-varying covariate.

Data analytic plan

To examine trajectories of alcohol and cannabis use outcomes preceding and following 21st birthday, we ran generalized linear mixed effects spline (or piecewise) models with random intercepts to account for clustering of observations within individuals. Because of the generalized form of the mixed effects models, estimates are conditional on the random intercept. Time was modeled as age in months, which was centered on the 21st birthday month (e.g., 1 month prior to 21st birthday month = −1, birthday month = 0, month after birthday month = 1). For the spline aspect of the model, three separate knots were specified that allowed for estimation of linear slopes at four distinct intervals: (1) up to 1 month before the 21st birthday month; (2) the slope from 1 month before to the month of 21st birthday; (3) the slope from the 21st birthday month to 1 month following; and (4) the slope from 1 month following the 21st birthday month and all following months. Thus, time interval 2 captures short-term changes during the 21st birthday month that may reflect celebratory increases, and time intervals 3 and 4 capture short- and long-term changes, respectively, following the 21st birthday month. All models included time-fixed baseline covariates for the demographic characteristics, calendar year at time of 21st birthday to account for potential historical differences (0: 2015, 1: 2016, 2: 2017 or 2018), and a time-varying covariate for 4-year college status.

Analyses included all available months for participants who turned 21 during the study. Thus, the possible range for prebirthday months was −23 (i.e., 23 months before the 21st birthday) to −1, and the possible range for postbirthday months was +1 to +23. Sensitivity analyses were conducted that restricted the range of months analyzed from 12 months prior to 12 months following the 21st birthday and substantive results were similar.

Because the drinks per week and past 30 day cannabis use outcomes were non-negative discrete integers showing positive skew with evidence of overdispersion where the variance exceeded the mean, we used negative binomial forms of the mixed effects models. Because the drinks per week and past 30-day cannabis use outcomes also showed evidence of excess zeros, these models were restricted to participants that at one or more of the monthly surveys reported at least one typical drink per week (n = 203) or at least 1 day of cannabis use (n = 145), respectively. Thus, the population of inference of those who at least engaged in occasional use. Frequency of heavy episodic drinking was an ordinal outcome and modeled using an ordinal logistic (also known as cumulative probability) form of the mixed effects model; and any past month SAM use, a dichotomous outcome, was modeled using a logistic regression form of the model. We also restricted analyses of heavy episodic drinking to those reporting alcohol use during at least one of the study months (n = 203); and analyses of SAM use were restricted to those who reported alcohol use during at least one study month and cannabis use during at least one study month (n = 141). Analyses of all four outcomes were also run in the full sample (i.e., including those that did not report any typical drinks per week or any days of cannabis use) and results were very similar.

We examined whether trajectories of all outcomes varied according to college status (attending 4-year college vs. not attending 4-year college) specifically at the 21st birthday month. For typical drinking and heavy episodic drinking outcomes, we further examined moderation by baseline level of typical drinking. For ease of interpretation, baseline drinking was dichotomized as <4 typical drinks per week (0) versus 4 or more typical drinks per week (1), where roughly half were in each category. For the cannabis use frequency outcome, we assessed moderation by any past month use reported at baseline; and for SAM use, moderation was assessed by both baseline drinking and any cannabis use. To test moderation, we assessed interaction terms between each of the time intervals and the moderators (e.g., time pre-21st birthday × 4-year college). Separate models were run for each outcome and moderator of interest.

All analyses were conducted using R statistical software version 3.6.2. We used the “glmmTMB” package for negative binomial mixed models (Magnusson et al., 2017), the “bbmle” package for ordinal logistic mixed models (Bolker, 2022), and the “lme4” package for logistic mixed models (Bates et al., 2015). Finally, to aid in interpretation, the “ggeffects” package was used to obtain model-predicted values across the study months, holding covariates constant at their mean (Lüdecke, 2018), and “ggplot2” was used to visualize these estimates (Wickham, 2016).

RESULTS

Typical drinks per week

Analyses of typical drinking utilized 4031 observations from 203 study participants who reported one or more typical drinks per month during at least one of the study months. Across monthly observations, zero typical drinks per week were reported in 36% of months. Table 2 shows results from the mixed effects models for the monthly changes over time in outcomes during the specific pre- and post-21st birthday time intervals, adjusted for covariates. There was no statistically significant change in typical drinking up to 1 month prior to the 21st birthday month, but a significant 50% increase in the typical number of drinks per week from Month −1 to the 21st birthday month (Month 0). There was no significant change from the 21st birthday month to the following month, and then a statistically significant, but modest, decrease for the remainder of the study. To aid in interpretation, Figure 1 shows model-predicted estimates for typical drinking over time. As is more apparent from the figure, although there was a decrease in typical drinking following the 21st birthday, the level of drinking was generally higher following compared to prior to the 21st birthday. When examining moderators, the trajectories appeared to vary according to 4-year college status at the time of the 21st birthday and by baseline levels of drinking. As shown in Figure 2A, for the trajectory up to Month −1, those in college showed a decrease in typical weekly drinking while those not in college showed an increase (interaction-p = 0.003), and there was also a steeper increase from Month −1 to the 21st birthday month among non-college participants relative to those in college (interaction-p = 0.011). There was no statistically significant difference in trajectories for the other time intervals. When considering the role of baseline drinking (Figure 2B), although those with lower levels of baseline drinking had lower levels of typical drinking over the course of the study compared to those with higher baseline typical drinking, they showed a stronger relative increase in typical drinking up to Month −1 (interaction p = 0.009). There were no other statistically significant interactions for the other time intervals.

TABLE 2.

Count or odds ratios from spline mixed effects models for monthly change in substance use outcomes at different time intervals before and after the 21st birthday month, adjusted for covariates.a

Typical drinking Heavy episodic drinking frequency Cannabis frequency Simultaneous alcohol and cannabis use
Covariate CR 95% Cl P OR 95% Cl P CR 95% Cl P OR 95% Cl P
Time intervalb
 Up to Month −1 1.00 0.99, 101 0.6661 0.98 0.96, 1.01 0.152 1.03 1.01, 1.04 <0.001 1.00 0.98, 1.03 0.778
 Month −1 to 0 1.50 1.31, 1.71 <0.001 2.25 1.58, 3.20 <0.001 1.07 0.88, 1.30 0.500 1.83 1.10, 3.07 0.021
 Month 0 to 1 1.03 0.91, 1.17 0.637 0.56 0.40, 0.80 0.001 0.96 0.80, 1.17 0.715 0.72 0.42, 1.20 0.206
 Month 1 to end 0.98 0.97, 0.99 <0.001 0.97 0.95, 0.99 0.011 1.00 0.99, 1.02 0.461 0.98 0.95, 1.02 0.306
Open in a new tab

Abbreviations: CR, Count Ratio; OR, Odds Ratio.

a

Adjusted for calendar year at 21st birthday (0: 2016; 1: 2017; 2: 2018), race (white [ref], Asian, other race [Black, American Indian or Alaska Native, Pacific Islander, Middle Eastern, Multiracial, other race]), ethnicity (0: non-Hispanic; 1: Hispanic), sex assigned at birth (0: female; 1: male), parents’ highest level of education (0: less than 4-year college degree, 1: 4-year college degree, 2: graduate or professional degree), sexual identity (0: heterosexual, 1: minoritized sexual identity [bisexual, gay, lesbian, queer, questioning, or other]), age at baseline (lowest age = 0), and 4-year college status (0: not currently attending 4-year college, 1: attending 4-year college) as time-varying covariate.

b

Month −1 is the month before the 21st birthday month; Month 0 is the month of the 21st birthday; and Month 1 is the month following the 21st birthday month.

FIGURE 1.

FIGURE 1

Open in a new tab

Model-predicted typical drinking over time.

FIGURE 2.

FIGURE 2

Open in a new tab

Model-predicted typical drinking over time by (A) college status and (B) baseline level of drinking.

Heavy episodic drinking frequency

Among those reporting at least one typical drink per week during at least one study month, there was a sharp significant increase in heavy episodic drinking frequency from Month −1 to the 21st birthday month, a sharp significant decrease from the birthday month to Month 1, and a continued significant decline following Month 1 (Table 2). There was no statistically significant change over time during the first time interval (up to Month −1). Because of the challenges with visualizing changes in a multi-category outcome, Figure 3 shows the model-predicted likelihood of any past month heavy episodic drinking, which shows a similar pattern of results as the ordinal categorical outcome. There was no evidence of differences in trajectories for any of the time intervals by college status. When examining moderation by baseline typical drinking, there was a steeper decline in the frequency of heavy episodic drinking over the months following the 21st birthday among those with high compared to low baseline typical drinking (interaction-p < 0.001). However, those engaging in high baseline drinking reported more frequent heavy episodic drinking over the course of the study. Figure 4 shows the predicted prevalence in the sample of any past month heavy episodic drinking (again, dichotomized for simplicity and also showing a similar pattern as the ordinal categorical outcome) by baseline drinking levels across the spline time intervals.

FIGURE 3.

FIGURE 3

Open in a new tab

Model-predicted likelihood of any past month heavy episodic drinking over time.

FIGURE 4.

FIGURE 4

Open in a new tab

Model-predicted likelihood of any past month heavy episodic drinking over time by baseline drinking.

Past month cannabis use days

Analyses of past month cannabis use utilized 2730 monthly observations from 145 participants who reported at least 1 day of cannabis use during at least one of the study months. Across the monthly observations, no cannabis use was reported in 50% of the months. As shown in Table 2 and visualized in Figure 5, in contrast to the alcohol outcomes, we observed statistically significant increasing trajectories of cannabis use frequency in the months leading up to the 21st birthday among those who reported cannabis use at least once during the study period (n = 145); however, there were no statistically significant changes during the other time intervals. We did not observe strong evidence for moderation of trajectories at any of the time intervals by college status. However, there was evidence of moderation by baseline cannabis use (Figure 6). The increase in cannabis use relative to where one started was stronger among those who did not report any use at baseline compared to baseline users during the months leading up to (interaction p < 0.001) and following the 21st birthday (interaction p = 0.005). However, across the course of the study, cannabis use was elevated among those using at baseline compared to those not using at baseline.

FIGURE 5.

FIGURE 5

Open in a new tab

Model-predicted number of past month cannabis use days over time.

FIGURE 6.

FIGURE 6

Open in a new tab

Model-predicted cannabis use days over time by baseline cannabis use.

SAM use

Among the 141 participants who reported use of alcohol and cannabis during at least one study month (though not necessarily at the same month), there was an increased likelihood of SAM use from the month prior to the month of the 21st birthday, but there were no other statistically significant changes during other time intervals (Table 2; Figure 7). There was no strong evidence of moderation in trajectories by college status or baseline drinking or cannabis use during any of the time intervals.

FIGURE 7.

FIGURE 7

Open in a new tab

Model-predicted likelihood of simultaneous and cannabis use over time.

DISCUSSION

For young adults, the 21st birthday marks a developmental milestone in the transition to adulthood, including legal access to alcohol and (in some states) cannabis. Consistent with prior literature, the typical number of drinks consumed per week, frequency of heavy episodic drinking, and likelihood of SAM use increased during the 21st birthday month in this study sample, which may reflect celebratory drinking. However, these 21st birthday increases did not appear to lead to sustained increases in the following months for alcohol. For cannabis use, while there was no clear increase in frequency of use during the 21st birthday month, specifically, there appeared to be increases over time leading up to the 21st birthday.

The study findings build on prior work (Fromme et al., 2010) by further identifying two additional hazardous drinking practices (i.e., SAM use, heavy episodic drinking) that also increase during the 21st birthday month. Even short-term increases in SAM use and heavy episodic drinking are cause for concern, as both are associated with more negative drinking consequences among young adults (Fleming et al., 2021; Patrick, Cronce, et al., 2016). In contrast, although cannabis use increased during the months leading up to the 21st birthday and remained fairly stable, patterns were not notably different during the 21st birthday month. This finding differs from research on cannabis use patterns in a state where cannabis was not legalized, such that past-year cannabis use increased over time among college students from ages 18 to 20, but peaked and declined after age 20 (Arria et al., 2017). It is possible, then, that although legal access to cannabis may not lead to subsequent sharp increases in cannabis use following the 21st birthday when individuals can legally purchase it, it may be associated with sustained increases in cannabis use over time. This is supported by recent data indicating the prevalence of cannabis use among adults aged 35 to 50 is at historic highs (Patrick et al., 2023). These findings suggest that prevention messaging past young adulthood may be of considerable public health benefit.

Considering the lack of increase in cannabis use during the 21st birthday month, one potential factor could be the lack of a legal public venue for celebratory cannabis use that is analogous to on-premise outlets available at age 21 for alcohol (i.e., bars). In Washington, retail on-site locations for consumption of cannabis are illegal. However, this is not the case in all states that have legalized nonmedical cannabis. A number of states permit public cannabis use (e.g., “cannabis lounges”; Nicole, 2021) and, if this practice were to proliferate, could potentially create an environment that could contribute to elevated use around 21st birthday celebrations. Another factor that may contribute to the lack of a significant increase in cannabis use during the 21st birthday month may involve motivations for cannabis use among young adults. While for some young adults, cannabis use is used for celebrations in general (Lee et al., 2007, 2009), this may not be the primary motivation for cannabis use. For many young adults, cannabis use is associated with using it to cope with social anxiety, loneliness, boredom, and/or distress (Garrison et al., 2021; Patrick et al., 2024; Rhew et al., 2021) rather than celebratory or social motivations.

Our results showed that, in the lead up to the 21st birthday, considerable differences were found in typical drinking behaviors by college status. Consistent with national datasets collected during the same time as the present study (Schulenberg et al., 2020), college students drank significantly more than non-college students. However, the trajectory for college students up to their 21st birthday was negative, while the trajectory for the same time period for noncollege students was positive, reducing the differences between the two groups. These findings may suggest that the influence of college wanes as students settle into the latter years of school (Arria et al., 2017). The fact that typical drinking changed at similarly dramatic rates for both college and noncollege students alike during the 21st birthday month highlights the universality of the 21st birthday effect. This paper, along with earlier work (Gilson et al., 2022), may suggest the importance of event-specific alcohol interventions around the 21st birthday for the noncollege population, in which such interventions have not been tested to our knowledge, as well as college populations. In contrast to typical drinking, there was no strong evidence that college and noncollege young adults showed different patterns of changes over time in cannabis use in the months surrounding 21st birthdays. However, it is important to note that the average number of days of cannabis use was significantly higher among noncollege than college young adults across the time period. Future studies that explore whether this difference extends longer-term would be informative.

Our study also showed that the 21st birthday effect on alcohol use was present for both those who reported lower and higher drinking at baseline. The 21st birthday effect for lower baseline drinkers, while smaller, is nevertheless significant. As pointed out by Lewis and colleagues, 21st birthday drinking among previously lower frequency drinkers was associated with greater negative consequences as the population may lack tolerance for alcohol, which mitigates some of the effects (Lewis et al., 2009). Due to their limited experience, they may also be less familiar with and/or not prepared to implement protective behavioral strategies that could mitigate risk. A small 21st birthday effect was noted among those who reported cannabis use at baseline, but not for those who reported no cannabis use at baseline, which may suggest that turning 21 in a state that has legalized non-medical cannabis is not associated with high levels of cannabis initiation. At least in these data, those who had refrained from use prior to reaching legal age largely continued to do so once reaching the minimum age for legal use.

This study had a number of strengths, including, most notably, the intensive longitudinal data collection that was able to capture monthly substance use prior to, during, and following the 21st birthday month, which reduces recall bias and allows for better estimation of change over time. However, there were important limitations that must also be considered. This study recruited a community convenience sample from a single urban area in the Pacific Northwest where cannabis is legal for those 21 and over. Thus, it is not clear whether the results generalize to other contexts. The dichotomous operationalization of simultaneous alcohol and cannabis use in analyses was not ideal. Future research should use measures that quantify the frequency of simultaneous alcohol and cannabis use. Relatedly, it is possible that there are other alcohol and cannabis measures (e.g., number of days of alcohol use, typical hours high on cannabis) that may have shown different trajectories around the 21st birthday, and additional studies should consider examining these outcomes. The distributions of the drinks per week and days of cannabis use outcomes showed high proportions of zeros, and, thus, the negative binomial form of the models may not have provided the optimal fit. However, zero-inflated count models for longitudinal data do not provide a straightforward interpretation, and, to our knowledge, there is not yet software widely available for other approaches, such as the marginalized zero-inflated model (Long et al., 2014; Mun et al., 2023), within a multilevel modeling framework.

In this study sample, there were heightened levels of substance use, particularly alcohol use and simultaneous alcohol and cannabis use, during the 21st birthday months relative to other months. However, these increases appeared to be generally restricted to the time around the 21st birthday and not necessarily indicative of longer-term, sustained escalations in the full sample. When exploring variation in trajectories by participant characteristics, it appeared that those who were not in college showed greater increases in typical drinking leading up to and during the 21st birthday month compared to those in college. Also, those originally showing lower levels of drinking at baseline did show more sustained levels of typical drinking and heavy episodic drinking post-21st birthday. Similarly, those not originally using cannabis at the beginning of the study showed greater relative increases in cannabis use leading up to and following the 21st birthday month. Further research into variability in trajectories across other important characteristics, including gender, will be important. Findings from the current study may have direct implications for the timing and personalization of prevention and intervention efforts. Preventive event-specific 21st birthday interventions may want to incorporate content targeting specific hazardous drinking behaviors in the month prior to the 21st birthday for both college and noncollege populations. However, more research is needed to determine ideal intervention time points to reduce increases in cannabis use among young adults.

FUNDING INFORMATION

Data collection and manuscript preparation were supported by grants from the National Institute on Alcohol Abuse and Alcoholism (R01 AA022087, R01 AA027496). Manuscript preparation was also supported by grant F32 AA029589. The content is solely the responsibility of the authors and does not necessarily represent the views of the National Institutes of Health.

Footnotes

CONFLIC T OF INTEREST STATEMENT

The authors have no conflict of interest to disclose.

DATA AVAILABILITY STATEMENT

Research data are not shared.

REFERENCES

  1. Arria AM, Caldeira KM, Allen HK, Bugbee BA, Vincent KB & O’grady KE (2017) Prevalence and incidence of drug use among college students: an 8-year longitudinal analysis. The American Journal of Drug and Alcohol Abuse, 43, 711–718. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bae H & Kerr DCR (2020) Marijuana use trends among college students in states with and without legalization of recreational use: initial and longer-term changes from 2008 to 2018. Addiction, 115, 1115–1124. [DOI] [PubMed] [Google Scholar]
  3. Bates D, Machler M, Bolker B & Walker S (2015) Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67, 1–48. [Google Scholar]
  4. Bolker BM (2022) bbmle: Tools for General Maximum Likelihood Estimation.
  5. Bravo AJ, Pearson MR, Conner BT & Parnes JE (2017) Is 4/20 an event-specific marijuana holiday? A daily diary investigation of marijuana use and consequences among college students. Journal of Studies on Alcohol and Drugs, 78, 134–139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brister HA, Sher KJ & Fromme K (2011) 21st birthday drinking and associated physical consequences and behavioral risks. Psychology of Addictive Behaviors, 25, 573–582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Buckner JD, Walukevich KA & Henslee AM (2018) Event-specific cannabis use and cannabis use motives. Substance Use & Misuse, 53, 1093–1098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Collins RL, Parks GA & Marlatt GA (1985) Social determinants of alcohol consumption: the effects of social interaction and model status on the self-administration of alcohol. Journal of Consulting and Clinical Psychology, 53, 189–200. [DOI] [PubMed] [Google Scholar]
  9. Crost B & Guerrero S (2012) The effect of alcohol availability on marijuana use: evidence from the minimum legal drinking age. Journal of Health Economics, 31, 112–121. [DOI] [PubMed] [Google Scholar]
  10. Crost B & Rees DI (2013) The minimum legal drinking age and marijuana use: new estimates from the NLSY97. Journal of Health Economics, 32, 474–476. [DOI] [PubMed] [Google Scholar]
  11. Fleming CB, Duckworth JC, Rhew IC, Abdallah DA, Guttmannova K, Patrick ME et al. (2021) Young adult simultaneous alcohol and marijuana use: between- and within-person associations with negative alcohol-related consequences, mental health, and general health across two-years. Addictive Behaviors, 123, 107079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fromme K, Wetherill RR & Neal DJ (2010) Turning 21 and the associated changes in drinking and driving after drinking among college students. Journal of American College Health, 59, 21–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Garrison E, Gilligan C, Ladd BO & Anderson KG (2021) Social anxiety, cannabis use motives, and social Context’s impact on willingness to use cannabis. International Journal of Environmental Research and Public Health, 18, 4882. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Geisner IM, Lewis MA, Rhew IC, Mittmann AJ, Larimer ME & Lee CM (2017) Does one day of drinking matter? 21st birthday drinking predicts subsequent drinking and consequences. Addictive Behaviors, 64, 57–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gilson MS, Cadigan JM, Fleming CB, Fairlie AM, Lewis MA & Lee CM (2022) Young adult birthday celebrations as windows of risk for alcohol and cannabis use: 21st birthdays compared to other young adult birthdays. Psychology of Addictive Behaviors, 36, 798–803. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kilmer JR, Rhew IC, Guttmannova K, Fleming CB, Hultgren BA, Gilson MS et al. (2022) Cannabis use among young adults in Washington state after legalization of nonmedical cannabis. American Journal of Public Health, 112, 638–645. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Lee CM, Neighbors C, Hendershot CS & Grossbard JR (2009) Development and preliminary validation of a comprehensive marijuana motives questionnaire. Journal of Studies on Alcohol and Drugs, 70, 279–287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lee CM, Neighbors C & Woods BA (2007) Marijuana motives: young adults’ reasons for using marijuana. Addictive Behaviors, 32, 1384–1394. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lewis MA, Lindgren KP, Fossos N, Neighbors C & Oster-Aaland L (2009) Examining the relationship between typical drinking behavior and 21st birthday drinking behavior among college students: implications for event-specific prevention. Addiction, 104, 760–767. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Long DL, Preisser JS, Herring AH & Golin CE (2014) A marginalized zero-inflated Poisson regression model with overall exposure effects. Statistics in Medicine, 33, 5151–5165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Lüdecke D (2018) Ggeffects: tidy data frames of marginal effects from regression models. Journal of Open Source Software, 3, 772. [Google Scholar]
  22. Magnusson A, Skaug HJ, Nielsen A, Berg CW, Kristensen K, Maechler M et al. (2017) glmmTMB: generalized linear mixed models using template model builder. R Package Version 0.1.3 [Google Scholar]
  23. Merrill JE & Carey KB (2016) Drinking over the lifespan: focus on college ages. Alcohol Research: Current Reviews, 38, 103–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mun EY, Zhou Z, Huh D, Tan L, Li D, Tanner-Smith EE et al. (2023) Brief alcohol interventions are effective through 6 months: findings from marginalized zero-inflated Poisson and negative binomial models in a two-step IPD meta-analysis. Prevention Science, 24, 1608–1621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Neighbors C, Spieker CJ, Oster-Aaland L, Lewis MA & Bergstrom RL (2005) Celebration intoxication: an evaluation of 21st birthday alcohol consumption. Journal of American College Health, 54, 76–80. [DOI] [PubMed] [Google Scholar]
  26. Nicole W (2021) Cannabis consumption in dispensaries: public health implications of an emerging practice. Environmental Health Perspectives, 129, 84001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Patrick ME, Cronce JM, Fairlie AM, Atkins DC & Lee CM (2016) Day-to-day variations in high-intensity drinking, expectancies, and positive and negative alcohol-related consequences. Addictive Behaviors, 58, 110–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Patrick ME, Kloska DD, Mehus CJ, Terry-McElrath Y, O’Malley PM & Schulenberg JE (2021) Key subgroup differences in age-related change from 18 to 55 in alcohol and marijuana use: U.S. National Data. Journal of Studies on Alcohol and Drugs, 82, 93–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Patrick ME, Miech RA, Johnston LD & O’Malley PM (2023) Monitoring the future study annual report: national data on substance use among adults 19 to 60, 1976–2022. Arbor Ann, MI: University of Michigan Institute for Social Research. [Google Scholar]
  30. Patrick ME, Peterson SJ, Terry-McElrath YM, Rogan SEB & Solberg MA (2024) Trends in coping reasons for marijuana use among U.S. adolescents from 2016 to 2022. Addictive Behaviors, 148, 107845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Patrick ME, Rhew IC, Lewis MA, Abdallah DA, Larimer ME, Schulenberg JE et al. (2018) Alcohol motivations and behaviors during months young adults experience social role transitions: microtransitions in early adulthood. Psychology of Addictive Behaviors, 32, 895–903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Patrick ME, Schulenberg JE, Miech RA, Johnston LD, O’Malley PM & Bachman JG (2022) Monitoring the future study annual report: national data on substance use among adults 19 to 60. Ann Arbor, MI: University of Michigan Institute for Social Research. [Google Scholar]
  33. Patrick ME, Terry-McElrath YM, Kloska DD & Schulenberg JE (2016) High-intensity drinking among young adults in the United States: prevalence, frequency, and developmental change. Alcoholism, Clinical and Experimental Research, 40, 1905–1912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Patrick ME, Terry-McElrath YM, Lanza ST, Jager J, Schulenberg JE & O’Malley PM (2019) Shifting age of peak binge drinking prevalence: historical changes in normative trajectories among young adults aged 18 to 30. Alcoholism, Clinical and Experimental Research, 43, 287–298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rhew IC, Cadigan JM & Lee CM (2021) Marijuana, but not alcohol, use frequency associated with greater loneliness, psychological distress, and less flourishing among young adults. Drug and Alcohol Dependence, 218, 108404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Rutledge PC, Park A & Sher KJ (2008) 21st birthday drinking: extremely extreme. Journal of Consulting and Clinical Psychology, 76, 511–516. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Schulenberg JE, Johnston LD, O’Malley PM, Bachman JG, Miech RA & Patrick ME (2020) Monitoring the future national survey results on drug use, 1975–2019: volume II, college students and adults ages 19–60. Arbor Ann, MI: Institute for Social Research, The University of Michigan. [Google Scholar]
  38. Substance Abuse and Mental Health Services Administration. (2021) Key substance use and mental health indicators in the United States: results from the 2020 National Survey on drug use and health. NSDUH Series H-56 Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. [Google Scholar]
  39. Terry-McElrath YM, Patrick ME, O’Malley PM & Johnston LD (2018) The end of convergence in developmental patterns of frequent marijuana use from ages 18 to 30: an analysis of cohort change from 1976–2016. Drug and Alcohol Dependence, 191, 203–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Walukevich-Dienst K & Buckner JD (2019) 4/20 cannabis use is greater than other high-risk events: identification of psychosocial factors related to 4/20 use. Cognitive Therapy and Research, 43, 852–860. [Google Scholar]
  41. White A & Hingson R (2013) The burden of alcohol use: excessive alcohol consumption and related consequences among college students. Alcohol Research: Current Reviews, 35, 201–218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Wickham H (2016) Elegant graphics for data analysis. New York, NY: Springer-Verlag. [Google Scholar]
  43. Yoruk BK & Yoruk CE (2013) The impact of minimum legal drinking age laws on alcohol consumption, smoking, and marijuana use revisited. Journal of Health Economics, 32, 477–479. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Research data are not shared.

RESOURCES