Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Mar 26.
Published in final edited form as: Alcohol Clin Exp Res (Hoboken). 2024 May 25;48(6):1107–1121. doi: 10.1111/acer.15323

Reasons for alcohol use from 1976 to 2020 in the United States among those ages 18 to 30: historical changes and mediation of cohort effects in binge drinking

Katherine M Keyes 1, Caroline Rutherford 1, Megan E Patrick 2, Jonathan M Platt 3, Deborah D Kloska 2, Justin Jager 4
PMCID: PMC13014394  NIHMSID: NIHMS2147121  PMID: 38795320

Abstract

Background:

Alcohol use is declining among US adolescents/early young adults and increasing among adults, and increases in adult binge drinking have been more concentrated in females than in males. Reasons for drinking are historically patterned by age and sex, and if they are historically variant it would suggest that changes over time could in part explain age- and sex-differential cohort effects.

Methods:

We analyzed longitudinal Monitoring the Future data for individuals born 1958 to 1990; these individuals were adults aged 29/30 from 1987 to 2020, and first surveyed at age 18 from 1976 to 2008 (N=14,190). Five reasons for drinking were analyzed (social, enhancement, avoid problems, relax, boredom); social reasons and to relax were most prevalent. Total effects of birth cohort predicting past-two-week binge drinking (5+ drinks) were estimated with polynomial regression models by age; indirect effects through mediators were estimated.

Results:

Drinking reasons exhibited dynamic time trends across birth cohort and sex. Notable increases were observed in social reasons to drink: among women aged 29/30, social reasons increased from 53% to 87% from 1987 to 2020. Social reasons to drink had prominent positive indirect effects at adult ages (age 23/24 and above among men; age 19 and above among women, indicating that increases in binge drinking would have been more negative if social reasons to drink had not increased. Social reasons also mediated male/female differences at adult ages, indicating that part of the reason sex differences are converging is more rapid increases in social reasons among adult women. Indirect effects were also observed for drinking to relax and drinking for boredom, and limited indirect effects were observed for enhancement and avoid problems.

Conclusion:

Changing endorsement of drinking reasons, especially social reasons, among US adult drinkers mediate cohort effects in binge drinking in the US adult population, and explain in part why binge drinking is converging by sex.

Keywords: binge drinking, cohort effects, reasons for drinking, gender differences

Introduction

Excessive alcohol use continues to be a central contributor to preventable morbidity and mortality in the United States (CDC, 2023; NIAAA, 2022), and per capita consumption of alcohol has increased ~3% annually over the last two decades (Slater and Alpert, 2020). The COVID-19 pandemic has accelerated these longer-term trends; alcohol-related deaths increased >25% just from 2019 to 2020 (White et al., 2022). Increased consumption, however, is starkly differential by age. Indeed, between 1976 and 2019, alcohol use and binge drinking declined more than 30% among adolescents/early young adults, while alcohol use and binge drinking increased more than 15% among adults by age 30, across the same time period (Jager et al., 2022). Increases in adult binge drinking have been more concentrated in females than in males, resulting in a narrowing of sex differences across adult cohorts (Keyes et al., 2019). The reasons underlying the observed age- and sex-differential cohort effects are multifactorial, and evidence supports a contribution of societal trends in education and family formation, religiosity and marriage, and gender roles and attitudes (Jager et al., 2022; Keyes et al., 2021). Yet these factors do not explain the totality of the observed trends, and the role of more proximal factors specifically related to the role of alcohol in society have been under-explored.

A long line of research has identified how motivations to drink (Cooper, 1994; Cox and Klinger, 1988), including reasons for drinking, and the perceived function that alcohol serves in a particular drinking context, are associated with decisions to drink, the level of alcohol use, and consequences of alcohol use (Bresin and Mekawi, 2021; Kuntsche et al., 2006; Skrzynski and Creswell, 2020). Reasons for alcohol use are typically distinguished along four central domains, including social (to obtain social rewards, to have fun), enhancement (to enhance positive affect, get high or drunk), conformity (to avoid social rejection), and coping (to forget about problems). Social reasons are most frequently endorsed across studies of both adolescents and adults (Kuntsche et al., 2005; Patrick and Schulenberg, 2011). Drinking for enhancement is associated with heavier drinking episodes (Bresin and Mekawi, 2021), and drinking to cope with stress is a particularly sensitive indicator for the development of alcohol problems (Kuntsche et al., 2005; Merrill et al., 2014). Understanding reasons for drinking provides a window into approaches for prevention, with tailored prevention interventions focused on occasion-specific contexts and providing specific protective strategies based on the motives for a particular drinking occasion.

Reasons for drinking are differ by age and sex (Kuntsche et al., 2006; Patrick and Schulenberg, 2011), and if they are historically variant as well, it would suggest that changes in reasons for drinking over time could in part explain age- and sex-differential cohort effects. Drinking for enhancement and coping motives increase with age (Patrick et al., 2017, 2011b; Patrick and Schulenberg, 2011), suggesting that as populations age, the role of alcohol in their lives and the purpose that it serves varies (Patrick et al., 2018). By sex (Holmila and Raitasalo, 2005), studies of both adolescents/early young adults and adults have generally found that boys and men endorse more social and conformity drinking reasons, including for social mobility and peer connection, whereas women endorse conformity and coping drinking reasons (Kuntsche et al., 2005). However, few studies have examined whether reasons to drink, among either adolescents, early young adults, or adults, are shifting across cohort. Available data indicate that since 2015, some coping reasons (relax or relieve tension) for alcohol use declined whereas others (drinking to get through the day) increased, but both accelerated after the onset of the COVID-19 pandemic (Patrick et al., 2022b). Thus, while prior literature has established age and sex associations with reasons for drinking, the question of whether those associations vary by cohort (i.e. age by cohort interactions, and age by cohort by sex interactions) remains untested. In sum, a broader historical analysis of the reasons and function of alcohol is lacking. This is a notable gap in the literature particularly given the historical increases in alcohol use among adult women; given that alcohol motivations are relevant for onset of heavy drinking and alcohol use disorders, assessing whether and how historical changes in alcohol motives mediate cohort differences in alcohol use by gender would provide clinically relevant science for intervention efforts.

The present study had two aims. First, we estimated long-term time trends in reasons for drinking alcohol by age and sex in general population samples for individuals born 1958 to 1990; these individuals were adults aged 29/30 from 1987 to 2020, and first surveyed at age 18 from 1976 to 2009. Participants who had not turned 29/30 by 2020 were not included. Second, we tested whether changes over historical time in reasons for drinking alcohol mediate cohort effects in binge drinking over that time period for men and women. In sum, this analysis examines whether part of the reason that sex differences in binge drinking across historical and developmental time are changing is that reasons for drinking among men and women in the US are shifting, given the role of drinking motivations in development of heavy drinking patterns and long-term drinking outcomes.

Materials and Methods

Study sample.

Data are drawn from the longitudinal Monitoring the Future (MTF) panel study (Patrick et al., 2022a). The pool of eligible respondents included those who participated in nationally representative samples of approximately 15,000 US high school seniors (12th grade) surveyed annually since 1976 (Miech et al., 2022). Among that pool, 2,450 students are randomly selected for longitudinal follow-up, with oversampling for students who report drug use. Students selected for follow-up are randomly assigned to begin assessments either one year after baseline (modal age 19) or two years after baseline (modal age 20) and are then followed every two years (for those who start at modal age 19, they are surveyed at modal age 21, 23, 25, 27, and 29; for those who start at modal age 20, they are surveyed at modal age 22, 24, 26, 28, and 30). Supplemental Table 1 provides an overview of the birth year, and years of data collection for each age, for all participants included in the present study. Respondents are randomly assigned survey forms with a set of core questions common across all forms, and form-specific questions; the present study includes only survey respondents assigned to the form that queries reasons to drink alcohol. Further, we include all longitudinal data on participants who were high school seniors between 1976 and 2008, as those cohorts had the opportunity to reach age 29/30 by 2020, the final year of data collection included in our analytic sample. In total, 14,190 respondents comprised the analytic sample. An Institutional Review Board of University of Michigan approved the study.

Attrition from the study cohort increases across age, and has increased across historical time. Between baseline and first follow-up, attrition among in the subset included in these analyses ranged from 15.00% for those who were in 12th grade in 1981 to 50.74% in for those who were in 12th grade in 2006. By age 29/30, attrition ranged from 26.74% in for those who were in 12th grade in 1976 to 65.49% in for those who were in 12th grade in 2003. Previous analyses indicate that attrition has a limited impact on longitudinal estimates of binge drinking in the MTF panel (Keyes et al., 2020), and we include comprehensive attrition weights in all analyses to account for known covariates of attrition. Attrition weights were calculated as the inverse probability of participation at each follow up based on the following covariates measured at age 18: gender, race/ethnicity, college plans, truancy, high school grades, number of parents in the home, religiosity, parental education, alcohol use, cigarette use, marijuana use, other illicit drug use, region, cohort, and sampling weight correcting for over-sampling of age 18 substance users. Nonresponse missingness ranged from 0.2% of respondents (sex) to 4.26% of respondents (past 12-month drinking at baseline).

Measures.

Birth cohort.

Cohorts are grouped and labeled by their modal birth year; for example, those who were high school seniors (modal age 18) in 1976 were born in modal year 1958, and those who were high school seniors (modal age 18) in 2008 were born in modal year 1990. The measure cohort ranged from 0 to 32 and was based on modal birth year (i.e., 1958 = 0, 1959 =1, and so forth). Using the variable cohort and the ORPOL function within SAS/IML (SAS Institute Inc., 2011), we generated orthogonal polynomials ranging from the first (linear), second (quadratic), and third (cubic) degree. Collectively, within our aim 2 mediation analyses these three polynomials were used to examine linear, quadratic, and cubic effects of cohort on both the mediators (reasons to drink) and the outcome (binge drinking).

Binge drinking.

Binge drinking was defined at each wave based on the question, “How many times have you had five or more drinks in a row over the past two weeks?” Due to heavily skewed distributions, responses at each wave were dichotomized as any occasion versus none. Note that binge drinking for women is often defined as 4+ drinks (NIAAA, 2020), thus this measure may have less sensitivity for detecting binge drinking among women. However, because any potential differential sensitivity by gender introduced by this measure would be systematic or consistent across age and cohort, our measure of binge drinking should not obscure estimates of age and cohort trends.

Abstaining.

Respondents were classified as abstaining from alcohol based on responses to “On how many occasions (if any) have you had any alcoholic beverage to drink—more than just a few sips—during the last 12 months?” Respondents who reported 0 times were classified as abstaining. Questions on reasons to drink alcohol were only asked among those who reported any drinking in the past 12 months. Descriptions and historical trends in alcohol abstaining are provided in Supplemental Figure 2.

Reasons for drinking.

Respondents who reported past-year alcohol use were asked to mark all of the 14 reasons for drinking alcohol that applied. Of the possible 14, we focus on the five most commonly endorsed reasons, following the same coding as Patrick and Schulenberg (2011). These reasons included: for social reasons (i.e. ‘to have a good time with my friends’), to relax (i.e. ‘to relax or relieve tension’), for enhancement (i.e. ‘to feel good or get high’), to avoid problems (i.e., “to get away from problems”), and boredom (i.e. “because of boredom, nothing else to do”). Responses were not mutually exclusive, thus respondents could endorse multiple reasons to drink. Descriptions and historical trends in each reason to drink item are provided in Figures 1 and 2, and Supplemental Figures 3 to 5.

Figure 1.

Figure 1.

Open in a new tab

Cohort trends in social reasons* for drinking alcohol among men and women born 1958 to 1990, by age from age 18 (surveyed from 1976 to 2008) to 29/30 (surveyed from 1987 to 2020)

* Social reasons based on endorsement of the item “to have a good time with my friends”

Figure 2.

Figure 2.

Open in a new tab

Cohort trends in drinking to relax* among men and women born 1958 to 1990, by age from age 18 (surveyed from 1976 to 2008) to 29/30 (surveyed from 1987 to 2020)

* Drinking to relax based on endorsement of the item ‘to relax or relieve tension’

Sex.

Respondents were classified as male and female based on self-report. While MTF surveys have included non-binary response options since 2019, binary response options only were included in most of the years that we analyzed.

Statistical analysis.

All analyses used appropriate weights to adjust for attrition as well as the oversampling of drug users at baseline. For aim 1, cohort trends in reasons to drink were estimated by age and sex using SAS Version 9.4 (SAS Institute Inc., 2012). For aim 2, whether cohort effects on binge drinking are mediated by cohort trends in reasons to drink was estimated for men and women via a path model in Mplus Version 8.9. Within this path model we employ an analytical approach that we have previously used (Jager et al., 2022) to test for unique indirect effects of multiple mediators when examining age-varying mediation of cohort effects on binge drinking.

To capture age-varying effects of cohort on each mediator (i.e., “a” paths from predictor to mediator), we specified the linear, quadratic, and cubic cohort polynomials as predictors of each mediator within the path model. To capture age-varying effects of each mediator on binge drinking (i.e., “b” paths from mediator to outcome), we specified each mediator for a given age as a predictor of that age’s binge drinking indictor within the path model. To capture age-varying effects of cohort on binge drinking (i.e., direct effects or the “c” path), we specified the linear, quadratic, and cubic cohort polynomials as predictors of each binge drinking indicator within the path model. We used the model constraint command available in Mplus (Muthén & Muthén, 2022) to calculate indirect effects by specifying the product of corresponding “a” and “b” paths and then calculate for each age the specific indirect effects for each mediator.

To model sex differences within the path model, we conducted multiple group analyses with sex as a grouping variable. To estimate sex differences, we again used the model constraint command in Mplus to create new model parameters that subtracted corresponding estimates of women from men. The estimates for these new model parameters as well as their standard errors were calculated within Mplus.

When estimating the effects of the cohort year polynomials and mediators on binge drinking within the path model, we used an identity link (as opposed to a non-linear logit link) in conjunction with a robust maximum likelihood sandwich estimator that is robust to non-normality. We did so because when comparing effects across age and across sex we are interested in absolute differences in binge drinking – or risk differences – which are obtained via an identity link function, as opposed to relative differences in binge drinking prevalence – or odds ratios and risk ratios – which are obtained via a non-linear logit link function. Example Mplus syntax for the Aim 2 path model is provided in Appendix 1.

Results

Supplemental Figure 1 shows the prevalence of binge drinking, by birth year and age and stratified by sex, in our analytic sample. Consistent with our prior publications (Jager et al., 2022; Keyes et al., 2019), cohort trends in binge drinking are differential by age. Among men at age 18, the prevalence of binge drinking declined from an estimated 43% among those born 1958 to 26% among those born in 1990. Yet by age 29/30, the prevalence of binge drinking increased across cohorts for men, particularly from a low of an estimated 28% among those born in 1980 to 46% among those born in 1990. These differential patterns by age are more pronounced among women. Among women at age 18, the prevalence of binge drinking declined from an estimated 23% among those born 1958 to 16% among those born in 1990. Yet by age 29/30, the prevalence of binge drinking increased across cohorts for women, from an estimated 14% among those born in 1958 to 20% among those born in 1990.

All drinking reasons exhibited dynamic time trends across birth cohort and sex. We present two drinking reasons in the main figures (Figures 1–2) and the remainder of drinking reasons in supplemental figures (Supplemental Figures 3–5). The main figures present trends for drinking for social reasons and drinking to relax/relieve tension, as these were the two most prevalently endorsed reasons of those analyzed.

Loess smoothing was used to show trends over time in Figures 1–2 and Supplemental Figures 1–5, and exact percentages can be found in Supplemental Tables 2–7. Figure 1 shows the time trend in social reasons among those who reported any past year alcohol use, by age and sex. Among early young adult men at age 18, social reasons have remained relatively stable across time, from 74% among those born in 1958 to 73% among those born in 1990 (see Supplemental Table 3). Among adult men, however, social reasons to drink have increased over time, with consistent increases at age 23/24, 25/26, and 29/30. Social reasons were more curvilinear at age 27/28; social reasons increased from 64% among those born in 1958 to 89% in among those born in 1980, and then declined to 59% among those born in 1990 (see Supplemental Table 3).

Trends were generally similar among women (Figure 1); however, among early young adult women at age 18, social reasons increased across time, from 60% among those born in 1958 to 72% among those born in 1990 (see Supplemental Table 3). Increases in social reasons among adult women were pronounced. For example, among those aged 29/30, social reasons increased by cohort, from 53% among those born in 1958 to a maximum of 87% among those born in 1987 (see Supplemental Table 3).

Figure 2 shows the time trend in drinking to relax among those who reported past year alcohol use, by age and sex. Among men aged 18, drinking to relax increased from 32% among those born in 1958 to a maximum of 54% among those born in 1984, and declined thereafter (see Supplemental Table 4). Trends in drinking to relax were similarly curvilinear for most young adult males. Increases in drinking to relax were consistently observed for men aged 29/30, who reported among the highest prevalence of drinking to relax, increasing from 64% among those born in 1958 to 86% among those born in 1990 (see Supplemental Table 4).

Among women (Figure 2), drinking to relax among those aged 18 remained flat, at 40% among women born in both 1958 in 1990 (see Supplemental Table 4). Drinking to relax increased across all other age groups, with those aged 29/30 most likely to endorse drinking to relax and increasing from 65% among those born in 1958 to 73% among those born in 1990 (see Supplemental Table 4).

Supplemental Figures 2 (past-year abstention), 3 (enhancement reason), 4 (avoid problems reason), and 5 (boredom reason) display trends in the additional mediators included in our analysis. Alcohol abstention sharply increased for those aged 18 and 19/20 among both men and women, and then leveled off and began to decrease at young adult ages. The declines in alcohol abstention were more pronounced for adult women than men. For example, among adult women aged 29/30, alcohol abstention declined from a height of 26% among those born 1966 to 13% among those born in 1990 (see Supplemental Table 2).

Table 1 describes, among men, the linear, quadratic, and cubic relationship between birth cohort and binge drinking prevalence in terms of total effect, direct effect (i.e. effect not mediated by reasons for drinking), and indirect effects through each motivation category as well as abstention from alcohol. Significant total cohort effects were observed, consistent with our other analyses (Jager et al., 2022), with linear declines at age 18 (B=−0.130, p<0.01), 19/20 (B=10.122, p<0.01), and 21/22 (B=−0.072, p<0.01), shifting to no change across cohort by age 23/24.

Table 1.

Total effects, indirect effects, and direct effects of linear, quadratic, and cubic associations between birth cohort (those born between 1958 and 1990) and binge drinking, by age, among men1

Total effect Indirect effect - Abstaining Indirect effect - Social Indirect effect - Relax Indirect effect – Enhancement Indirect effect – Avoid problems Indirect effect - Boredom Direct effect of cohort
Linear cohort effects
Age 18 (surveyed from 1976 to 2008) −0.130 (<.01) −0.027 (<.01) −0.013 (<.01) 0.006 (0.02) −0.021 (<.01) 0.002 (0.15) 0.001 (0.19) −0.068 (<.01)
Age 19/20 (surveyed from 1978 to 2010) −0.122 (<.01) −0.041 (<.01) 0.014 (0.02) 0.000 (0.93) −0.007 (0.29) −0.001 (0.55) −0.004 (0.15) −0.075 (<.01)
Age 21/22 (surveyed from 1980 to 2012) −0.072 (<.01) −0.024 (0.04) 0.014 (0.08) 0.001 (0.66) −0.007 (0.16) 0.000 (0.77) 0.000 (0.90) −0.051 (<.01)
Age 23/24 (surveyed from 1982 to 2014) −0.034 (0.09) −0.005 (0.17) 0.014 (<.01) 0.003 (0.17) −0.010 (0.02) 0.001 (0.45) 0.004 (0.03) −0.037 (0.03)
Age 25/26 (surveyed from 1984 to 2016) −0.004 (0.84) −0.005 (0.30) 0.012 (<.01) 0.004 (0.13) −0.004 (0.25) 0.004 (0.14) 0.001 (0.31) −0.015 (0.42)
Age 27/28 (surveyed from 1986 to 2018) −0.029 (0.16) 0.000 (0.93) 0.007 (<.01) 0.005 (0.04) −0.003 (0.61) 0.000 (0.62) 0.005 (0.02) −0.040 (0.02)
Age 29/30 (surveyed from 1988 to 2020) 0.001 (0.96) 0.002 (0.71) 0.011 (<.01) 0.006 (0.01) 0.002 (0.61) 0.003 (0.18) 0.007 (0.01) −0.028 (0.15)
Quadratic cohort effects
Age 18 (surveyed from 1976 to 2008) −0.022 (0.17) −0.003 (0.39) −0.003 (0.18) −0.003 (0.12) −0.002 (0.52) −0.002 (0.06) −0.002 (0.08) −0.007 (0.60)
Age 19/20 (surveyed from 1978 to 2010) 0.008 (0.67) 0.004 (0.42) −0.001 (0.63) −0.002 (0.26) −0.002 (0.60) 0.000 (0.75) 0.000 (0.86) 0.009 (0.58)
Age 21/22 (surveyed from 1980 to 2012) −0.007 (0.72) 0.004 (0.42) −0.004 (0.18) −0.001 (0.58) −0.001 (0.85) 0.000 (0.74) −0.001 (0.65) −0.004 (0.80)
Age 23/24 (surveyed from 1982 to 2014) −0.007 (0.71) 0.001 (0.65) 0.001 (0.65) −0.001 (0.71) 0.004 (0.35) −0.001 (0.43) −0.001 (0.72) −0.011 (0.53)
Age 25/26 (surveyed from 1984 to 2016) −0.006 (0.76) 0.001 (0.85) 0.000 (0.89) 0.001 (0.65) 0.002 (0.53) −0.002 (0.28) 0.001 (0.33) −0.008 (0.65)
Age 27/28 (surveyed from 1986 to 2018) −0.026 (0.21) 0.000 (0.93) −0.003 (0.13) 0.001 (0.59) 0.002 (0.68) −0.001 (0.39) 0.000 (0.87) −0.024 (0.18)
Age 29/30 (surveyed from 1988 to 2020) 0.031 (0.15) 0.002 (0.74) 0.000 (0.78) −0.001 (0.73) 0.005 (0.19) −0.002 (0.36) 0.003 (0.08) 0.023 (0.23)
Cubic cohort effects
Age 18 (surveyed from 1976 to 2008) 0.002 (0.89) 0.001 (0.77) 0.003 (0.19) −0.001 (0.39) −0.001 (0.73) 0.000 (0.96) −0.001 (0.34) 0.001 (0.94)
Age 19/20 (surveyed from 1978 to 2010) −0.002 (0.93) −0.002 (0.76) −0.003 (0.37) −0.001 (0.70) −0.007 (0.09) 0.000 (0.77) −0.002 (0.25) 0.011 (0.49)
Age 21/22 (surveyed from 1980 to 2012) 0.002 (0.91) −0.002 (0.75) −0.004 (0.13) 0.000 (0.94) −0.004 (0.23) 0.000 (0.78) −0.001 (0.51) 0.012 (0.49)
Age 23/24 (surveyed from 1982 to 2014) −0.054 (0.01) −0.001 (0.76) 0.000 (0.97) −0.002 (0.25) −0.001 (0.77) −0.001 (0.42) −0.004 (0.03) −0.041 (0.02)
Age 25/26 (surveyed from 1984 to 2016) −0.023 (0.25) −0.001 (0.73) −0.002 (0.45) 0.001 (0.76) −0.001 (0.81) −0.002 (0.24) −0.001 (0.51) −0.016 (0.38)
Age 27/28 (surveyed from 1986 to 2018) −0.047 (0.03) 0.000 (0.94) −0.005 (0.11) −0.004 (0.20) −0.007 (0.28) 0.000 (0.52) −0.002 (0.15) −0.025 (0.16)
Age 29/30 (surveyed from 1988 to 2020) 0.017 (0.41) −0.003 (0.60) −0.002 (0.18) 0.002 (0.35) 0.002 (0.59) −0.001 (0.70) 0.002 (0.20) 0.017 (0.37)
Open in a new tab
1

Abstaining was coded as 1 if a respondent reported 0 times to “On how many occasions (if any) have you had any alcoholic beverage to drink—more than just a few sips during the last 12 months?” Reasons for drinking were coded as 1 if a respondent endorsed them when answering: “What have been the most important reasons for your drinking alcoholic beverages? (Mark all that apply).” Social: “To have a good time with my friends”; Relax: “To relax or relieve tension”; Enhancement: “To feel good or get high”; Avoid problems: “To get away from my problems or troubles”; Boredom: “Because of boredom, nothing else to do”

These total cohort linear effects were in part mediated by abstention as well as reasons for drinking. Abstaining had a significant negative mediating cohort effect at age 18 (B=−0.027, p<0.01), 19/20 (B=−0.041, p<0.01) and 21/22 (B=−0.024, p=0.04), meaning that part of the reason that binge drinking linearly declined by cohort at these ages was that men at these ages are less likely to engage in any alcohol use in more recently-born cohorts. Further, the overall magnitude of the indirect cohort effect declines across age, consistent with alcohol abstention as a partial explanation of the reversal.

Over and above abstention, changing alcohol reasons also mediated linear cohort change in binge drinking. Social and enhancement reasons had a negative indirect cohort effect at age 18 (see Table 1), meaning that if social and enhancement reasons were more invariant, the linear relationship between cohort and binge drinking would have declined less than it did. In other words, social and enhancement reasons to drink mediated, in part, declines in binge drinking among age 18 men.

Also shown in Table 1, positive indirect cohort effects through reasons for drinking were observed in adult ages, especially social reasons for which there were significant positive indirect effects at ages 19/20, 23/24, 25/26, 27/28 and 29/30. The most straightforward interpretation of these effects is to consider the counterfactual: if social reasons were invariant by cohort, the relationship between cohort and binge drinking would have been more negative for adult men. Indeed, the direct effect of cohort is negative at these adult ages, indicating that we would have observed declines in binge drinking among adult men in more recently-born cohorts if social reasons were invariant across cohort. However, because these social reasons increased in prevalence in adults, the total linear cohort effect is null. In other words, increases in social reasons mediated historical variation in binge drinking in adults, and binge drinking would have continued to decline in adult ages (as it did at age 18), but did not because social reasons increased in prevalence. There were also positive indirect cohort effects at some adult ages for drinking due to boredom and drinking to relax. The interpretation of these effects is similar to the interpretation for drinking for social reasons: at these ages the cohort effect in binge drinking would have been more negative if drinking due to boredom and drinking to relax had not increased. There were negative indirect cohort effects for drinking for enhancement at age 18 and 23/24, and no indirect effects for drinking to avoid problems. There were no systematic patterns for quadratic or cubic cohort effects.

Table 2 shows a parallel analysis among women. Total linear cohort effects for cohort among women are generally less pronounced than among men; while there is a negative linear cohort effect at age 18 (B=−0.044, p<0.01) and age 19/20 (B=−0.031, p=0.04) for women, it is less pronounced than among men. Total linear cohort effects become positive at older ages for women (Table 2), with a significant positive linear effect for women at age 25/26 (B=0.049, p=0.01), and age 29/30 (B=0.049, p=0.01).

Table 2.

Total effects, indirect effects, and direct effects of linear, quadratic, and cubic associations between birth cohort (those born between 1958 and 1990) and binge drinking, by age, among women1

Total effect Indirect effect - Abstaining Indirect effect – Social Indirect effect - Relax Indirect effect – Enhancement Indirect effect – Avoid Problems Indirect effect - Boredom Direct effect of cohort
Linear cohort effects
Age 18 (surveyed from 1976 to 2008) −0.044 (<.01) −0.015 (<.01) 0.001 (0.71) 0.002 (0.56) −0.025 (<.01) 0.002 (0.15) 0.002 (0.36) −0.009 (0.50)
Age 19/20 (surveyed from 1978 to 2010) −0.031 (0.04) −0.024 (0.03) 0.021 (<.01) 0.004 (0.19) −0.005 (0.52) 0.003 (0.12) 0.005 (0.13) −0.032 (0.02)
Age 21/22 (surveyed from 1980 to 2012) 0.030 (0.07) −0.003 (0.51) 0.026 (<.01) 0.006 (<.01) −0.007 (0.02) 0.003 (0.03) 0.003 (0.02) 0.000 (0.99)
Age 23/24 (surveyed from 1982 to 2014) 0.024 (0.15) 0.000 (0.92) 0.021 (<.01) 0.007 (<.01) −0.005 (0.03) 0.003 (0.06) 0.004 (0.02) −0.005 (0.76)
Age 25/26 (surveyed from 1984 to 2016) 0.049 (0.01) 0.002 (0.80) 0.022 (<.01) 0.005 (0.02) −0.006 (0.07) 0.002 (0.09) 0.001 (0.64) 0.021 (0.18)
Age 27/28 (surveyed from 1986 to 2018) 0.010 (0.58) 0.011 (0.04) 0.018 (<.01) 0.005 (0.01) −0.002 (0.41) 0.005 (0.02) −0.006 (0.12) −0.020 (0.22)
Age 29/30 (surveyed from 1988 to 2020) 0.049 (0.01) 0.004 (0.62) 0.013 (<.01) 0.005 (0.03) 0.000 (0.99) 0.004 (0.27) 0.008 (0.04) 0.013 (0.45)
Quadratic cohort effects
Age 18 (surveyed from 1976 to 2008) 0.015 (0.29) 0.000 (0.85) −0.001 (0.30) 0.001 (0.75) 0.000 (0.87) −0.001 (0.12) −0.001 (0.66) 0.017 (0.17)
Age 19/20 (surveyed from 1978 to 2010) 0.013 (0.39) 0.002 (0.60) −0.001 (0.75) 0.002 (0.15) 0.001 (0.76) −0.001 (0.45) 0.001 (0.48) 0.008 (0.56)
Age 21/22 (surveyed from 1980 to 2012) 0.003 (0.84) 0.005 (0.22) −0.002 (0.33) 0.001 (0.36) −0.002 (0.56) −0.002 (0.19) 0.002 (0.06) 0.000 (0.97)
Age 23/24 (surveyed from 1982 to 2014) 0.017 (0.30) 0.003 (0.35) −0.002 (0.40) 0.004 (0.04) 0.000 (0.84) −0.001 (0.42) 0.002 (0.06) 0.011 (0.46)
Age 25/26 (surveyed from 1984 to 2016) 0.050 (<.01) 0.002 (0.80) 0.001 (0.56) 0.003 (0.33) −0.001 (0.78) 0.001 (0.40) 0.002 (0.53) 0.040 (0.01)
Age 27/28 (surveyed from 1986 to 2018) 0.029 (0.09) 0.010 (0.05) −0.004 (0.20) 0.003 (0.04) 0.002 (0.48) 0.003 (0.14) −0.005 (0.19) 0.018 (0.25)
Age 29/30 (surveyed from 1988 to 2020) 0.030 (0.10) 0.003 (0.63) −0.003 (0.18) 0.002 (0.29) −0.001 (0.80) 0.003 (0.28) 0.010 (0.01) 0.016 (0.36)
Cubic cohort effects
Age 18 (surveyed from 1976 to 2008) 0.005 (0.73) 0.005 (0.03) 0.002 (0.09) 0.001 (0.66) 0.000 (0.95) 0.002 (0.08) −0.001 (0.55) −0.004 (0.74)
Age 19/20 (surveyed from 1978 to 2010) −0.035 (0.03) 0.000 (0.99) 0.002 (0.29) −0.001 (0.68) −0.007 (0.02) 0.000 (0.75) −0.003 (0.05) −0.026 (0.06)
Age 21/22 (surveyed from 1980 to 2012) −0.012 (0.45) 0.002 (0.66) −0.002 (0.44) 0.001 (0.62) −0.004 (0.13) 0.002 (0.10) 0.000 (0.76) −0.011 (0.44)
Age 23/24 (surveyed from 1982 to 2014) −0.031 (0.06) −0.001 (0.66) −0.001 (0.67) −0.001 (0.52) −0.004 (0.12) −0.002 (0.16) −0.004 (0.02) −0.017 (0.24)
Age 25/26 (surveyed from 1984 to 2016) −0.013 (0.44) −0.001 (0.82) −0.006 (0.01) −0.003 (0.04) 0.001 (0.53) −0.002 (0.11) −0.001 (0.50) −0.001 (0.93)
Age 27/28 (surveyed from 1986 to 2018) 0.012 (0.48) −0.007 (0.11) −0.006 (0.01) −0.002 (0.09) 0.001 (0.50) −0.004 (0.04) 0.002 (0.52) 0.027 (0.09)
Age 29/30 (surveyed from 1988 to 2020) −0.004 (0.82) −0.005 (0.62) 0.00 (0.99) 0.000 (0.82) 0.002 (0.72) −0.002 (0.72) −0.004 (0.40) 0.004 (0.80)
Open in a new tab
1

Abstaining was coded as 1 if a respondent reported 0 times to “On how many occasions (if any) have you had any alcoholic beverage to drink—more than just a few sips during the last 12 months?” Reasons for drinking were coded as 1 if a respondent endorsed them when answering: “What have been the most important reasons for your drinking alcoholic beverages? (Mark all that apply).” Social: “To have a good time with my friends”; Relax: “To relax or relieve tension”; Enhancement: “To feel good or get high”; Avoid problems: “To get away from my problems or troubles”; Boredom: “Because of boredom, nothing else to do”

Alcohol abstention, in part, mediated linear cohort effects. Negative indirect effects for alcohol abstention at age 18 (B=−0.015, p<0.01) and 19/20 (B=−0.024, p=0.03) indicate that part of the reason that binge drinking declined in more recently-born cohorts at these ages is that women at these ages were more likely to abstain from alcohol across historical time. A significant positive indirect effect of abstaining was observed at age 27/28, meaning that the total effect of cohort on binge drinking would have been more positive if alcohol abstaining had been more invariant.

Over and above indirect effects of abstention, social reasons were significant positive mediators of linear cohort effects in drinking across all ages save for age 18. These significant positive indirect cohort effects have different interpretations at different ages, because the total cohort effects are in different directions at different ages. Similar to positive indirect effects among men, the interpretation for these indirect effects is that part of the reason why the linear relationship between cohort and binge drinking is negative at age 19/20, and positive at age 25/26 and 29/30, is that across these age groups social reasons increased. That is, the cohort effects would have trended downward (i.e. negative effects would have been more negative and positive effects would have been less positive), but they did not trend downward because social reasons increased in prevalence for these age groups in more recently-born cohorts. While smaller in magnitude, positive indirect cohort effects were observed across all adult ages beginning at age 21/22 for drinking to relax. The interpretation of these effects is similar to the interpretation for drinking for social reasons: part of the reason that there is a linear increase in binge drinking at adult ages (significant increases for ages 25/26 and ages 29/30, but positive in direction for other ages as well) is because there were increases among women in drinking to relax. At select ages, positive indirect effects through drinking for boredom and to avoid problems were also observed. However, negative indirect effects for drinking for enhancement were observed at ages 18, 12/22 and 23/24, indicating that cohort effects would have trended upward if drinking for enhancement was invariant over time. There were no systematic patterns for quadratic or cubic cohort effects.

Table 3 shows the difference between men and women in linear, quadratic, and cubic total cohort effects, direct cohort effects (i.e. effect not mediated by motivations to drink), indirect cohort effects through each motivation category. These effects can be interpreted as cohort by sex interactions. The linear decline in binge drinking in more recently-born cohorts was significantly more pronounced for men, with faster linear declines at age 18 (B=−0.086, p<0.01), 19/20 (B=−0.091, p<0.01), 21/22 (B=−0.102, p<0.01), 23/24 (B=−0.058, p=0.02), and 25/26 (B=−0.053, p=0.05). There were few sex differences in the indirect effects of drinking reasons. However, comparing the indirect effect estimate for social reasons at ages 18 (B=−0.014, p<0.01), 25/26 (B=−0.011, p=0.03), and 27/28 (B=−0.011, p=0.02) between men and women, the indirect effect is less positive for men than women, thus creating a negative indirect effect for the sex difference. We can thus interpret this negative sex difference for the indirect effect to mean that if social reasons were invariant by cohort and sex, the sex difference in binge drinking at these ages would have been more positive (i.e. men even more likely to binge drink than women). However, because social reasons changed, and increased faster for women than for men at adults ages, this faster increase led to a sex convergence in binge drinking between men and women at adult ages. In other words, part of the explanation for sex convergence in adulthood is that women are endorsing social reasons to drink more, and differentially more than men across cohort. Thus, sex convergence in social reasons mediates gender convergence in binge drinking. There were no systematic patterns for quadratic or cubic cohort effects.

Table 3.

Male/female differences in total effects, indirect effects, and direct effects of linear, quadratic, and cubic associations between birth cohort (those born between 1958 and 1990) and binge drinking, by age1

Total effect Indirect effect - Abstaining Indirect effect - Social Indirect effect - Relax Indirect effect - Enhancement Indirect effect – Avoid Problems Indirect effect - Boredom Direct effect of cohort
Linear cohort effects
Age 18 (surveyed from 1976 to 2008) −0.086 (<.01) −0.012 (0.07) −0.014 (<.01) 0.004 (0.24) 0.005 (0.43) 0.000 (0.83) 0.000 (0.93) −0.068 (<.01)
Age 19/20 (surveyed from 1978 to 2010) −0.091 (<.01) −0.017 (0.24) −0.007 (0.32) −0.004 (0.30) −0.002 (0.82) −0.003 (0.11) −0.008 (0.04) −0.049 (0.04)
Age 21/22 (surveyed from 1980 to 2012) −0.102 (<.01) −0.022 (0.08) −0.012 (0.15) −0.005 (0.21) −0.001 (0.89) −0.003 (0.13) −0.003 (0.19) −0.056 (0.02)
Age 23/24 (surveyed from 1982 to 2014) −0.058 (0.02) −0.005 (0.30) −0.007 (0.15) −0.004 (0.20) −0.005 (0.32) −0.002 (0.20) 0.001 (0.70) −0.036 (0.14)
Age 25/26 (surveyed from 1984 to 2016) −0.053 (0.05) −0.006 (0.42) −0.011 (0.03) −0.002 (0.60) 0.002 (0.61) 0.001 (0.64) 0.000 (0.89) −0.038 (0.14)
Age 27/28 (surveyed from 1986 to 2018) −0.039 (0.15) −0.011 (0.04) −0.011 (0.02) 0.000 (0.89) −0.001 (0.90) −0.005 (0.03) 0.011 (0.01) −0.023 (0.38)
Age 29/30 (surveyed from 1988 to 2020) −0.048 (0.09) −0.002 (0.84) −0.003 (0.56) 0.001 (0.79) 0.002 (0.74) −0.001 (0.77) −0.002 (0.75) −0.043 (0.11)
Quadratic cohort effects
Age 18 (surveyed from 1976 to 2008) −0.037 (0.09) −0.003 (0.53) −0.002 (0.50) −0.003 (0.18) −0.001 (0.72) −0.001 (0.60) −0.001 (0.52) −0.026 (0.19)
Age 19/20 (surveyed from 1978 to 2010) −0.005 (0.83) 0.002 (0.74) −0.001 (0.86) −0.004 (0.07) −0.003 (0.54) 0.000 (0.59) −0.001 (0.54) 0.001 (0.95)
Age 21/22 (surveyed from 1980 to 2012) −0.011 (0.68) −0.001 (0.93) −0.001 (0.69) −0.003 (0.34) 0.001 (0.83) 0.002 (0.38) −0.003 (0.10) −0.005 (0.82)
Age 23/24 (surveyed from 1982 to 2014) −0.024 (0.35) −0.001 (0.78) 0.003 (0.37) −0.004 (0.09) 0.004 (0.36) 0.001 (0.72) −0.003 (0.14) −0.023 (0.34)
Age 25/26 (surveyed from 1984 to 2016) −0.056 (0.03) −0.002 (0.87) −0.001 (0.77) −0.002 (0.56) 0.003 (0.55) −0.003 (0.17) −0.001 (0.71) −0.051 (0.05)
Age 27/28 (surveyed from 1986 to 2018) −0.055 (0.04) −0.010 (0.06) 0.000 (0.90) −0.002 (0.53) 0.001 (0.93) −0.003 (0.09) 0.005 (0.20) −0.045 (0.08)
Age 29/30 (surveyed from 1988 to 2020) 0.001 (0.98) −0.001 (0.90) 0.003 (0.37) −0.002 (0.36) 0.006 (0.25) −0.005 (0.16) −0.007 (0.09) 0.008 (0.77)
Cubic cohort effects
Age 18 (surveyed from 1976 to 2008) −0.003 (0.90) −0.004 (0.32) 0.001 (0.75) −0.002 (0.36) −0.001 (0.77) −0.002 (0.17) 0.000 (0.96) 0.006 (0.78)
Age 19/20 (surveyed from 1978 to 2010) 0.034 (0.17) −0.002 (0.81) −0.005 (0.17) 0.000 (0.99) 0.000 (0.97) 0.000 (0.89) 0.001 (0.63) 0.040 (0.08)
Age 21/22 (surveyed from 1980 to 2012) 0.015 (0.57) −0.003 (0.61) −0.002 (0.50) −0.001 (0.84) 0.000 (0.97) −0.003 (0.14) −0.001 (0.48) 0.025 (0.31)
Age 23/24 (surveyed from 1982 to 2014) −0.023 (0.38) 0.000 (0.97) 0.001 (0.74) −0.001 (0.71) 0.002 (0.60) 0.001 (0.40) 0.000 (0.88) −0.026 (0.27)
Age 25/26 (surveyed from 1984 to 2016) −0.010 (0.70) −0.001 (0.86) 0.004 (0.27) 0.004 (0.09) −0.002 (0.57) 0.000 (0.96) 0.001 (0.74) −0.016 (0.54)
Age 27/28 (surveyed from 1986 to 2018) −0.059 (0.03) 0.007 (0.27) 0.002 (0.62) −0.002 (0.55) −0.008 (0.22) 0.004 (0.07) −0.005 (0.22) −0.056 (0.03)
Age 29/30 (surveyed from 1988 to 2020) 0.022 (0.45) 0.002 (0.87) −0.002 (0.51) 0.003 (0.41) 0.000 (0.97) 0.001 (0.89) 0.006 (0.21) 0.013 (0.62)
Open in a new tab
1

Abstaining was coded as 1 if a respondent reported 0 times to “On how many occasions (if any) have you had any alcoholic beverage to drink—more than just a few sips during the last 12 months?” Reasons for drinking were coded as 1 if a respondent endorsed them when answering: “What have been the most important reasons for your drinking alcoholic beverages? (Mark all that apply).” Social: “To have a good time with my friends”; Relax: “To relax or relieve tension”; Enhancement: “To feel good or get high”; Avoid problems: “To get away from my problems or troubles”; Boredom: “Because of boredom, nothing else to do”

Discussion

We aimed to estimate the extent to which reasons for drinking are changing across time and by age, and whether those changes mediate changes over time in binge drinking and explain shifting sex differences. The context for this investigation is the observed reversal in binge drinking across historical time and development (Jager et al., 2022; Patrick et al., 2019). Declines in binge drinking among recent early young adult (age 18) cohorts reverse in adulthood, resulting in no change or even increases in prevalence among recent adult cohorts. This reversal is more prominent for women than for men. Given these differential patterns in binge drinking by age, mediators that would explain the reversal should also be differentially changing in their indirect effects across age, differentially by sex. We add four central findings. First, we observed that overall alcohol abstention is a mediator of the reversal; we observe negative indirect effects in early young adulthood that become null or positive at older ages. This is consistent with the reversal in which binge drinking declines in early young adulthood, and then becomes null or increases in older ages. Second, over and above abstention, reasons for drinking among those who report alcohol use in the past 12 months are shifting across historical time, differentially by age in ways that are consistent with explanations for the reversal. The most prominent example is social reasons to drink. Among both males and females there were indirect effects of social reasons, and the magnitude generally increased (i.e., got less negative/more positive) as age increased. Thus, absent any historical variation in social reasons, cohort effects would be adjusted downward to a greater extent at older ages, thus mitigating the magnitude of the reversal. Third, indirect effects for drinking to relax also increased as age increased for both genders, indicating that historical shifts in drinking to relax also partially mediated the reversal. Fourth, we find that differential changes over time in social reasons to drink partially explain sex convergence in binge drinking; indeed, women who drink increased social reasons at a faster rate than men in adulthood, relative to previous cohorts. Taken together, these results suggest the reversal is, at least in part, mediated by less overall alcohol use in early young adulthood shifting to more overall alcohol use in later young adulthood, and that increases in binge drinking in later young adulthood is in part explained by increases in drinking for social reasons and to relax among those who drink. These reasons for drinking are differentially increasing among females, indicating that reasons to drink are becoming more similar between males and females. While available literature indicates that social and relaxation reasons to drink have always been prominent motivations (Kuntsche et al., 2005; Patrick and Schulenberg, 2011), and less associated with alcohol problems than other motivations such as coping with problems (Kuntsche et al., 2006; Patrick et al., 2011a), these data indicate that they are becoming even more prominent especially for adult drinking women; promotion of social and recreational activities that decenter alcohol may be part of a public health approaches to reduce alcohol-related harms in adult populations.

We show that abstention rates increased historically at younger ages but remained flat at older ages. Indeed, increasingly, early young adults (particularly those ages 18 to 20) are simply not drinking at all. This is a major shift in the epidemiology of alcohol use in the United States; prior studies (Johnston et al., 2023) have demonstrated that in the late 1970s and early 1980s, more than 80% of high school seniors drank in the past year and more than 90% had tried alcohol, and by 2008, less than half of high school seniors used alcohol in the last year. By 2013 less than half ever tried alcohol, and alcohol use has further declined since then among adolescents (Johnston et al., 2023). The reasons why alcohol abstention is increasing in adolescence, and the peak ages of binge drinking extending during young adulthood (Patrick et al., 2019) are multifactorial. In the 1980s, prior work has shown that increases in the minimum legal drinking age (MLDA) across states mediated part of the decline in alcohol use among young people (Jager et al., 2022), yet alcohol use continued to decline in the decades since major shifts in the MLDA in many states. Developmental studies have demonstrated a general shift of adulthood starting later in recent decades (Patrick et al., 2019), and there is a broader shift downward in numerous behaviors typically associated with the externalizing spectrum (Kreski et al., 2022a). The prevalence of numerous other drugs has also declined (Miech et al., 2020), and conduct problems such as interpersonal violence, shoplifting and other crime has declined, as has sexual activity in adolescence (Keyes et al., 2018; Twenge and Park, 2017). Adolescents and early young adults are spending fewer evenings out without supervision, and general participation in peer activities such as sports and clubs has remarkably declined in recent years (Kreski et al., 2022b). Thus, the increases in abstention that are localized to the youngest end of the age band examined here may be better contextualized as part of this broader shift, which itself is inadequately understood in terms of underlying causation. Some have hypothesized that shifts in social network preferences to digital formats may underlie general declines in youth unsupervised social time (Twenge, 2020), yet it is incomplete; these shifts began prior to the widespread use of digital technology for engagement, and the association between digital technology in recent trends in youth behavior and mental health is generally of small magnitude (Ferguson et al., 2021; Orben, 2020). Further, declines in drug use are not universal; cannabis use prevalence has remained relatively invariant (Miech et al., 2022; Patrick et al., 2020), and novel drug products such as e-cigarettes and vapes surged in prevalence in the late 2010s (Miech et al., 2021)

Yet over and above shifts towards alcohol abstention, social reasons to drink mediate both declines in binge drinking in early young adulthood (age 18) among women, and the plateau/increases in adulthood among both men and women. Reasons that early young adults (age 18) who drink are endorsing fewer social reasons may be contextualized within the broader frame of less adolescent/early young adult unsupervised social time in general (Keyes et al., 2018; Kreski et al., 2023); if adolescent/early young adults are socializing less, then they have fewer social/recreational motives to drink. College attendance, especially among women, has increased over the past several decades (Carnevale and Rose, 2003; NCES, 2003), thus future orientations and academic pressure (Högberg, 2021) associated with college preparation may reduce the desire for social/recreational activities with alcohol involved due to concerns about academic performance and potential punitive consequences of underage drinking. In turn, the increases in social reasons to drink during the transition to adulthood may result from delay in adult role acquisition that is typically associated with declines in drinking (Casey et al., 2011; Sawyer et al., 2018). Indeed, today’s young adults are delaying partnership and family formation to later years (Jager et al., 2015; Lesthaeghe, 2014; Mayer, 2004); we have previously documented that delays in marriage in particular mediate the reversal in binge drinking (Jager et al., 2022), which may in turn increase social reasons to drink if young adults have fewer responsibilities. Future research incorporating multiple levels of mediation, from birth cohort to social roles and social time to downstream mediation through reasons for drinking and alcohol use behaviors, may be an important area to fully analyze these myriad pathways.

We also find that increases in social reasons to drink in adulthood explain, at least for some ages, convergence in sex differences in drinking. While men remain more likely to binge drink than women, increases in social reasons to drink among adult women have been faster than increases for men, and because endorsing social reasons for drinking is associated with higher rates of binge drinking, this faster increase had led to convergence between men and women. Available evidence indicates that roles associated with more binge drinking, such as college attendance (Bailey and DiPrete, 2016), are increasing more among women than men across the last four decades, and roles associated with less binge drinking are declining, such as marriage/co-habitation and having children (Adams et al., 2023; McKetta and Keyes, 2019; Mortimer, 2015). Previous analyses have also indicated that increases in binge drinking among adult women are concentrated in those with the highest levels of income, education, and occupational prestige (McKetta et al., 2021; McKetta and Keyes, 2020), which together come with more resources to purchase products such as alcohol. With these changes and delays in roles, it may be that adult women have more time and resources for going out and socializing with alcohol than in previous decades.

Importantly, alcohol use among women is often framed as a coping response to sex-specific or sex-concentrated stressors (Pollard et al., 2020), and indeed we find that increases in drinking to relax or relieve tension have increased among adult women and partially mediate the plateau, and for some ages increases in binge drinking among adult women. Yet social reasons, not reasons around tension relief, explain why sex differences are converging among adults. The alcohol industry has centered marketing and advertising for women on positioning alcohol as an adjunct to a healthy social life (Fullwood et al., 2016; Jung and Hovland, 2016), capitalizing on the increases in social reasons to drink. Public health campaigns that provide accurate information about potential health consequences of alcohol use are needed to counter these messages, especially targeted to women. Further, effective public health prevention initiatives such as price and taxation (Burton et al., 2017; Guindon et al., 2022; Wagenaar et al., 2010) may be particularly effective given the trends that we observe towards the centering of alcohol for social/recreational purposes.

Limitations of the study are noted. Measures of reasons for drinking are not based on multidimensional validated scales, and have relatively low inter-item correlation compared with other measures. However, the advantage of the MTF study design is that measures are consistently captured across 40 years of data collection, thus this is the only data to our knowledge that captures broad historical variation in motivations to use alcohol. The sampling frame of MTF does not include those not in high school, thus results may not generalize beyond those who attend high school. Further, there is substantial attrition in the MTF panel, and attrition has increased across historical time as it has for many population-based longitudinal studies; we use comprehensive attrition weights to account for loss to follow-up, and our own simulation analyses indicate that the amount of attrition has limited impact on longitudinal analyses of binge drinking (Keyes et al., 2020). Nevertheless, the attrition in the sample is a limitation.

In conclusion, we document in this paper three salient processes that underlie, in part, historical trends in binge drinking that systematically vary by age in the U.S. First, part of the reason that binge drinking is declining in adolescence/early young adulthood is that more young people are abstaining from alcohol altogether, but the historical increase in abstention wanes in the transition to adulthood as age peaks in binge drinking shift to older ages (Patrick et al., 2019). Second, among those who do drink, social reasons to drink are increasing in adulthood, leading to plateaus and reversals in binge drinking across developmental time. Third, there are differential reasons to drink by sex; young adult women are endorsing both more social and tension related reasons for drinking compared with young adult women in previous decades, and increases in social reasons for drinking are increasing historically among young adult women more than young adult men, leading to sex convergence in alcohol use.

Taken together, these results indicate that providing attractive alcohol-free alternatives for socializing, while a prominent strategy among college-attending young adults (NIAAA, 2023), may be a promising avenue for reducing heavy drinking among all adults. The need for a public health approach is urgent; from 2015–2019, almost 700,000 US adults died from excessive alcohol use per year, indicating that 1 in 8 deaths among working-age adults is attributable to excessive alcohol use (which is likely an underestimate of the total contribution of alcohol to US deaths) (Esser et al., 2022). In 2020, deaths accelerated during the COVID-19 pandemic, indicating a growing and urgent need for intervention and prevention efforts focused on those groups exhibiting increases in alcohol use, particularly adults entering midlife (White et al., 2022). Continuing to advance and promote other population-level interventions with known efficacy in reducing adult binge drinking such as price and sales controls, and addressing misleading advertising messages are all important avenues to reduce harms related to alcohol in the US population.

Supplementary Material

Supplementary Material
Appendix 1

Acknowledgements:

The present work was funded by R01AA026861 (PI: Keyes and Jager), R01 DA016575 (PI: Patrick), R01DA001411 (PI: Miech).

Footnotes

Financial disclosure/conflict of interest: The authors report no conflicts of interest and have no financial relationships with commercial interests related to the material in this manuscript.

References

  1. Adams R, McKetta S, Jager J, Stewart M, Keyes K (2023) Cohort effects of women’s midlife binge drinking and alcohol use disorder symptoms in the United States: Impacts of changes in timing of parenthood. Addiction. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bailey MJ, DiPrete TA (2016) Five Decades of Remarkable but Slowing Change in U.S. Women’s Economic and Social Status and Political Participation. RSF 2:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bresin K, Mekawi Y (2021) The “Why” of Drinking Matters: A Meta-Analysis of the Association Between Drinking Motives and Drinking Outcomes. Alcohol Clin Exp Res 45:38–50. [DOI] [PubMed] [Google Scholar]
  4. Burton R, Henn C, Lavoie D, O’Connor R, Perkins C, Sweeney K, Greaves F, Ferguson B, Beynon C, Belloni A, Musto V, Marsden J, Sheron N (2017) A rapid evidence review of the effectiveness and cost-effectiveness of alcohol control policies: an English perspective. Lancet 389:1558–1580. [DOI] [PubMed] [Google Scholar]
  5. Carnevale A, Rose S (2003) Socioeconomic Status, Race/Ethnicity, and Selective College Admissions. A Century Foundation Paper. [Google Scholar]
  6. Casey BJ, Jones RM, Somerville LH (2011) Braking and Accelerating of the Adolescent Brain. J Res Adolesc 21:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. CDC (2023) Deaths from Excessive Alcohol Use in the United States. National Center for Chronic Disease Prevention and Health Promotion. [Google Scholar]
  8. Cooper ML (1994) Motivations for alcohol use among adolescents: Development and validation of a four-factor model. Psychol Assess 6:117–128. [Google Scholar]
  9. Cox WM, Klinger E (1988) A Motivational Model of Alcohol Use. J Abnorm Psychol 97:168–180. [DOI] [PubMed] [Google Scholar]
  10. Esser MB, Leung G, Sherk A, Bohm MK, Liu Y, Lu H, Naimi TS (2022) Estimated Deaths Attributable to Excessive Alcohol Use Among US Adults Aged 20 to 64 Years, 2015 to 2019. JAMA Netw Open 5:e2239485–e2239485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Ferguson CJ, Kaye LK, Branley-Bell D, Markey P, Ivory JD, Klisanin D, Elson M, Smyth M, Hogg JL, McDonnell D, Nichols D, Siddiqui S, Gregerson M, Wilson J (2021) Like This Meta-Analysis: Screen Media and Mental Health. Prof Psychol Res Pr 53:205–214. [Google Scholar]
  12. Fullwood MD, Mongiovi J, Hillyer G, Basch CH, Ethan D, Hammond R (2016) An advertisement analysis of alcohol products in popular women’s magazines. Front Womens Health 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Guindon GE, Zhao K, Fatima T, Garasia S, Quinn N, Baskerville NB, Paraje G (2022) Prices, taxes and alcohol use: a systematic umbrella review. Addiction (Abingdon, England) 117:3004–3023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Högberg B (2021) Educational stressors and secular trends in school stress and mental health problems in adolescents. Soc Sci Med 270:113616. [DOI] [PubMed] [Google Scholar]
  15. Holmila M, Raitasalo K (2005) Gender differences in drinking: why do they still exist? Addiction (Abingdon, England) 100:1763–1769. [DOI] [PubMed] [Google Scholar]
  16. Jager J, Keyes KM, Schulenberg JE (2015) Historical variation in young adult binge drinking trajectories and its link to historical variation in social roles and minimum legal drinking age. Dev Psychol 51:962–974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jager J, Keyes KM, Son D, Patrick ME, Platt J, Schulenberg JE (2022) Age 18–30 trajectories of binge drinking frequency and prevalence across the past 30 years for men and women: Delineating when and why historical trends reversed across age. Dev Psychopathol 1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Johnston LD, Miech RA, Patrick ME, O’Malley PM, Schulenberg JE, Bachman JG (2023) Monitoring the Future national survey results on drug use 1975–2022: Overview, key findings on adolescent drug use. Institute for Social Research, University of Michigan. [Google Scholar]
  19. Jung AR, Hovland R (2016) Targeting gender: A content analysis of alcohol advertising in magazines. http://dx.doi.org/101080/0735968320161199210 33:221–238. [DOI] [PubMed] [Google Scholar]
  20. Keyes KM, Gary DS, Beardslee J, Prins SJ, O’Malley PM, Rutherford C, Schulenberg JE (2018) Joint effects of age, period, and cohort on conduct problems among American adolescents from 1991 through 2015. Am J Epidemiol 187:548–557. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Keyes KM, Jager J, Mal-Sarkar T, Patrick ME, Rutherford C, Hasin DS (2019) Is there a recent epidemic of women’s drinking? A critical review of national studies. Alcohol Clin Exp Res 43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Keyes KM, Jager J, Platt J, Rutherford C, Patrick ME, Kloska DD, Schulenberg J (2020) When does attrition lead to biased estimates of alcohol consumption? Bias analysis for loss to follow-up in 30 longitudinal cohorts. Int J Methods Psychiatr Res 29:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Keyes KM, Platt J, Rutherford C, Patrick ME, Kloska DD, Schulenberg J, Jager J (2021) Cohort effects on gender differences in alcohol use in the United States: How much is explained by changing attitudes towards women and gendered roles? SSM Popul Health 15:100919. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kreski NT, Cerdá M, Chen Q, Hasin DS, Martins SS, Mauro PM, Olfson M, Keyes KM (2022a) Adolescents’ Use of Free Time and Associations with Substance Use from 1991 to 2019. Subst Use Misuse 57:1893–1903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kreski NT, Chen Q, Olfson M, Cerdá M, Hasin DS, Martins SS, Mauro PM, Keyes KM (2022b) Time use and associations with internalizing symptoms from 1991 to 2019 among US adolescents. SSM Popul Health 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kreski NT, Riehm KE, Cerdá M, Chen Q, Hasin DS, Martins SS, Mauro PM, Olfson M, Keyes KM (2023) Parenting Practices and Adolescent Internalizing Symptoms in the United States, 1991–2019. J Adolesc Health 72:189–196. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kuntsche E, Knibbe R, Gmel G, Engels R (2006) Who drinks and why? A review of socio-demographic, personality, and contextual issues behind the drinking motives in young people. Addictive behaviors 31:1844–1857. [DOI] [PubMed] [Google Scholar]
  28. Kuntsche E, Knibbe R, Gmel G, Engels R (2005) Why do young people drink? A review of drinking motives. Clin Psychol Rev 25:841–861. [DOI] [PubMed] [Google Scholar]
  29. Lesthaeghe R (2014) The second demographic transition: a concise overview of its development. Proc Natl Acad Sci U S A 111:18112–18115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mayer KU (2004) Whose lives? How history, societies, and institutions define and shape life courses. Res Hum Dev 1:161–187. [Google Scholar]
  31. McKetta S, Prins SJ, Bates LM, Platt JM, Keyes KM (2021) US trends in binge drinking by gender, occupation, prestige, and work structure among adults in the midlife, 2006–2018. Ann Epidemiol 62:22–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. McKetta SC, Keyes KM (2020) Trends in U.S. women’s binge drinking in middle adulthood by socioeconomic status, 2006–2018. Drug Alcohol Depend 212:108026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. McKetta SC, Keyes KM (2019) Heavy and binge alcohol drinking and parenting status in the United States from 2006 to 2018: An analysis of nationally representative cross-sectional surveys. PLoS Med 16:e1002954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Merrill JE, Wardell JD, Read JP (2014) Drinking Motives in the Prospective Prediction of Unique Alcohol-Related Consequences in College Students. J Stud Alcohol Drugs 75:93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Miech R, Keyes KM, O’Malley PM, Johnston LD (2020) The great decline in adolescent cigarette smoking since 2000: consequences for drug use among US adolescents. Tob Control 29:638–643. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Miech R, Leventhal A, Johnston L, O’Malley PM, Patrick ME, Barrington-Trimis J (2021) Trends in Use and Perceptions of Nicotine Vaping Among US Youth From 2017 to 2020. JAMA Pediatr 175:185–190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Miech RA, Johnston LD, O’malley PM, Bachman JG, Schulenberg JE, Patrick ME (2022) Monitoring the Future national survey results on drug use, 1975–2021: Volume I, Secondary school students. Institute for Social Research, The University of Michigan. [Google Scholar]
  38. Mortimer JT (2015) Social change and entry to adulthood. Emerging Trends in the Social and Behavioral Sciences, Major Reference Works 1–17. [Google Scholar]
  39. Muthén & Muthén (2022) Mplus 8.9.
  40. NCES (2003) Digest of Education Statistics 2002. National Center for Education Statistics. [Google Scholar]
  41. NIAAA (2023) CollegeAIM Alcohol Intervention Matrix. College Drinking Prevention. [Google Scholar]
  42. NIAAA (2022) Alcohol Facts and Statistics. NIH. [Google Scholar]
  43. NIAAA (2020) Drinking Levels Defined. NIH. [Google Scholar]
  44. Orben A (2020) Teenagers, screens and social media: a narrative review of reviews and key studies. Soc Psychiatry Psychiatr Epidemiol 55:407–414. [DOI] [PubMed] [Google Scholar]
  45. Patrick ME, Evans-Polce R, Kloska DD, Maggs JL, Lanza ST (2017) Age-Related Changes in Associations Between Reasons for Alcohol Use and High-Intensity Drinking Across Young Adulthood. J Stud Alcohol Drugs 78:558–570. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Patrick ME, Miech RA, Kloska DD, Wagner AC, Johnston LD (2020) Trends in Marijuana Vaping and Edible Consumption From 2015 to 2018 Among Adolescents in the US. JAMA Pediatr 174:900–902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Patrick ME, Rhew IC, Lewis MA, Abdallah DA, Larimer ME, Schulenberg JE, Lee CM (2018) Alcohol motivations and behaviors during months young adults experience social role transitions: Microtransitions in early adulthood. Psychol Addict Behav 32:895–903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Patrick ME, Schulenberg JE (2011) How Trajectories of Reasons for Alcohol Use Relate to Trajectories of Binge Drinking: National Panel Data Spanning Late Adolescence to Early Adulthood. Dev Psychol 47:311–317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Patrick ME, Schulenberg JE, Miech RA, Johnston LD, O’Malley PM, Bachman JG (2022a) Monitoring the Future Panel Study annual report: National data on substance use among adults ages 19 to 60, 1976–2021.
  50. Patrick ME, Schulenberg JE, O’Malley PM, Johnston LD, Bachman JG (2011a) Adolescents’ Reported Reasons for Alcohol and Marijuana Use as Predictors of Substance Use and Problems in Adulthood. J Stud Alcohol Drugs 72:106–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Patrick ME, Schulenberg JE, O’Malley PM, Maggs JL, Kloska DD, Johnston LD, Bachman JG (2011b) Age-Related Changes in Reasons for Using Alcohol and Marijuana From Ages 18 to 30 in a National Sample. Psychology of Addictive Behaviors 25:330–339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. 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. Alcohol Clin Exp Res 43:287–298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Patrick ME, Terry-McElrath YM, Miech RA, Keyes KM, Jager J, Schulenberg JE (2022b) Alcohol use and the COVID-19 pandemic: Historical trends in drinking, contexts, and reasons for use among US adults. Soc Sci Med In press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Pollard MS, Tucker JS, Green HD Jr (2020) Changes in adult alcohol use and consequences during the COVID-19 pandemic in the US. JAMA Netw Open 3:e2022942–e2022942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. SAS Institute Inc. (2012) SAS 9.4 System Options: Reference, Second Edition. Cary, NC. [Google Scholar]
  56. Sawyer SM, Azzopardi PS, Wickremarathne D, Patton GC (2018) The age of adolescence. Lancet Child Adolesc Health 2:223–228. [DOI] [PubMed] [Google Scholar]
  57. Skrzynski CJ, Creswell KG (2020) Associations between solitary drinking and increased alcohol consumption, alcohol problems, and drinking to cope motives in adolescents and young adults: a systematic review and meta-analysis. Addiction (Abingdon, England) 115:1989–2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Slater ME, Alpert HR (2020) Surveillance report #117: Apparent per capita alcohol consumption: National, state, and regional trends, 1977–2019. National Institute on Alcohol Abuse and Alcoholism. [Google Scholar]
  59. Twenge JM (2020) Why increases in adolescent depression may be linked to the technological environment. Curr Opin Psychol 32:89–94. [DOI] [PubMed] [Google Scholar]
  60. Twenge JM, Park H (2017) The decline in adult activities among U.S. adolescents, 1976–2016. Child Dev. [DOI] [PubMed] [Google Scholar]
  61. Wagenaar AC, Tobler AL, Komro KA (2010) Effects of alcohol tax and price policies on morbidity and mortality: A systematic review. Am J Public Health 100:2270–2278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. White AM, Castle I-JP, Powell PA, Hingson RW, Koob GF (2022) Alcohol-related deaths during the COVID-19 pandemic. JAMA. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Material
NIHMS2147121-supplement-Supplementary_Material.docx (860.1KB, docx)
Appendix 1
NIHMS2147121-supplement-Appendix_1.docx (100.8KB, docx)

RESOURCES