From Risk to Resilience? Hazardous Drinking Trajectories in and Beyond the Last Years of University Life

Abstract

In this study, we examined the effects of loneliness, social support, and stress resilience on alcohol consumption and problems among university students in their final years of education during the COVID-19 pandemic. We surveyed 437 students with a pre-pandemic history of heavy episodic drinking across five waves from February 2021 to May 2023. Our findings showed that stress resilience significantly reduced alcohol-related problems over time. Those who frequently drank before the pandemic experienced a slower decline in problems, suggesting a delay in maturing out. Men reported higher hazardous drinking, yet gender did not influence trajectories. Loneliness initially correlated with increased drinking problems, without long-term effects, and social support had no significant impact. Our results highlight that stress resilience is essential for preventing alcohol problems, reveal the persistence of hazardous drinking into later university years, and suggest that the COVID-19 pandemic shifted typical drinking patterns in the Netherlands, marked by significant post-lockdown rebounds.

Keywords

loneliness stress resilience social support alcohol use/abuse longitudinal

Introduction

Hazardous drinking is the quantity or pattern of alcohol use that increases the risk of adverse health consequences (Saunders et al., 1993). Heavy episodic drinking (HED) is a form of hazardous drinking characterized by consuming a large quantity of alcohol on a single occasion (Jackson, 2008; Wechsler & Nelson, 2001). This behaviour is alarmingly prevalent in university settings and is deeply rooted in student culture (NIAAA, 2023). For instance, a detailed 2022 survey at a Dutch university found that 43–47% of students engage in these drinking behaviours, highlighting the urgency to address this issue (Vink, 2023). In the Netherlands, guidelines to identify HED suggest a cut-off of 4 or more drinks for women and 6 or more drinks for men in a single day, occurring at least weekly (Trimbos Instituut, 2022). University students who display HED are at increased risk for alcohol-related problems, including accidents, memory problems, academic setbacks, and alcohol overdoses, and have a greater likelihood of developing dependence compared to their peers (Babor et al., 2001; Herrero-Montes et al., 2022; White & Hingson, 2014). Given the serious risks associated with HED, it is important to identify indicators of persistent or increased hazardous drinking behaviours in critical life transitions.

The transition out of university life is a critical period for investigating hazardous drinking, coinciding with age-related reductions in this behaviour (Hingson et al., 2017; Trimbos Instituut, 2022). While existing studies indicate a general decrease, often referred to as maturing out of hazardous drinking in young adulthood, some individuals persist in or escalate their drinking behaviour into more problematic patterns (Lee & Sher, 2018; Petker et al., 2019). Well-known risk factors for persisting/increasing hazardous drinking in the period of young adulthood are family history of alcohol use disorders, male gender, specific personality traits (e.g., greater extraversion, impulsivity, and sensation-seeking tendencies), ADHD symptoms, and psychological distress, such as depressive symptoms, along with early drinking initiation (Ashenhurst et al., 2015; Gotham et al., 1997; Haardörfer et al., 2021; Jackson et al., 2001). Considering decreases in hazardous drinking over time, environmental influences, including transitional role changes (e.g., from student to employee) and life events related to family dynamics (e.g., marriage), play pivotal roles in reducing hazardous drinking. Moreover, individual characteristics such as greater emotional stability and conscientiousness have been associated with greater decreases (for reviews see Lee & Sher, 2018; Petker et al., 2019).

Despite these insights, our understanding of hazardous drinking trajectories among university students, particularly during their last years and beyond, remains incomplete. The COVID-19 pandemic significantly influenced drinking patterns among students between early 2020 and 2022. Measures to curb the virus’s spread, such as the closure of nightlife venues, sports clubs, and universities, along with social distancing interventions, disrupted daily routines. Initially, the lockdown reduced alcohol consumption and hazardous drinking among Dutch university students (Rubio et al., 2023; van Hooijdonk et al., 2022). However, by May-June 2022, when restrictions were lifted, hazardous drinking levels seemed to surge beyond pre-pandemic levels, particularly among men (Vink, 2023). This highlights the need for longitudinal research to understand hazardous drinking trajectories during and after the pandemic.

Furthermore, it is important to note that the pandemic’s impact extended beyond drinking habits to significantly affect students’ mental health, resulting in increased distress and loneliness (Werner et al., 2021; Wood et al., 2024). These mental health challenges, prevalent among students and exacerbated by the pandemic, may have further complicated the trajectories of hazardous drinking. Therefore, it is crucial to examine how specific independent predictors—loneliness, social support, and stress resilience—affect students' hazardous drinking during times of stress and transitions. Within hazardous drinking trajectories, we will focus on two significant aspects: alcohol consumption and alcohol-related problems. Given the robust evidence of higher risk progression to problematic alcohol use among men, we will also explore the interplay between gender and predictors of hazardous drinking (Ashenhurst et al., 2015; Haardörfer et al., 2021; Jackson et al., 2001).

One key predictor in our study is loneliness. Loneliness, characterized by the mismatch between expected and attained social relationships, in either quality, quantity, or both, is a growing health concern among students with significant implications for mental and physical well-being (Holt-Lunstad, 2017; Perlman & Peplau, 1981). While previous research consistently links loneliness and alcohol-related issues in clinical samples (Akerlind & Hörnquist, 1992) and the general population (Wakabayashi et al., 2022), the relationship in university students is complex and warrants further longitudinal study. Longitudinal research has generally found that higher levels of loneliness predict alcohol-related problems (Herchenroeder et al., 2022; Segrin et al., 2018), especially among students with hazardous alcohol and eating-related behaviours (Herchenroeder et al., 2022). Interestingly, some studies found that loneliness did not directly influence alcohol consumption (Segrin et al., 2018) or hazardous drinking (Richardson et al., 2017).

During the early pandemic period (first lockdown, March 2020), a prospective study by Pocuca et al. (2022) observed that higher loneliness was associated with increased alcohol consumption frequency in emerging adults. This finding contrasts with previous studies that primarily linked loneliness to alcohol problems rather than (hazardous) consumption, suggesting that the relationship might be specific to the pandemic period. Given the mixed results of prior research, the long-term impact of heightened loneliness on hazardous drinking in students remains uncertain. No study has specifically explored these relationships in students with HED during times of stress and transitions. Significant research gaps include the consistency of loneliness’s effects on hazardous drinking trajectories across studies and whether these effects extend beyond the pandemic context. Fostering social connections and providing support during transitions could mitigate the negative effects of loneliness on hazardous drinking.

On the flip side of the perceived absence of social connections, the availability of support networks during challenging periods (Zimet et al., 1988) has been associated with lower rates of alcohol use and higher rates of abstinence (Groh et al., 2007; Jason et al., 2012). However, the relationship between social support and hazardous drinking in university students is nuanced and varies with the source of support—family, friends, or a significant other (i.e., romantic partner). For example, being in a committed romantic relationship, which is often seen as part of maturing out of hazardous drinking, has been linked to fewer alcohol problems among university students (Egerton & Read, 2019). These findings could be explained by the support from a partner or lifestyle changes associated with being in a relationship. In contrast, peer support can influence higher hazardous drinking among students, as HED is often seen as part of social status and adherence to peer norms (Tinajero et al., 2019). Furthermore, while family support is protective during adolescence, its effectiveness may not extend into university years (Groh et al., 2007). The mentioned examples highlight the complex interplay between different social support sources and hazardous drinking in students.

Most prior studies have focused on social connections rather than direct measures of social support, with some exceptions. A study by Lechner et al. during the COVID-19 pandemic found that university students who felt more supported by family, peers, and romantic partners reported lower alcohol use during the first lockdown compared to those who felt less supported (Lechner et al., 2020). This study, however, only explored the combined influence of these support sources during this specific period. To our knowledge, no study has yet examined whether social support from these sources individually (or collectively) protects against the persistence or increase of hazardous drinking in students with HED during times of stress and transitions. A significant gap remains in understanding how specific sources of social support can positively influence hazardous drinking trajectories in students, particularly those at increased risk for dependence. This can direct prevention efforts towards stimulating specific sources of support during times of stress and transitions.

In addition to the influence of social connections, factors such as stress resilience play a critical role in protecting against negative mental health outcomes, both in general and during the pandemic (Smith et al., 2008; Veer et al., 2021). Resilience can be defined in various ways. We adopted Smith et al.'s (2008) definition of stress resilience, focusing on the trait-level capacity to rebound from stress. Research has indicated that heightened trait resilience may mitigate the risk of alcohol use disorders (AUD) and problematic drinking across different populations, spanning from older adults (van Gils et al., 2022) to university students (Lyvers et al., 2020; Sanchez et al., 2022). Furthermore, a recent scoping review by Cusack et al. (2023) stated that trait resilience was a robust protective factor against several alcohol use outcomes besides AUD, such as excessive alcohol use, dependence symptoms, and alcohol problems.

To our knowledge, only one study has explored the longitudinal effects of trait resilience specifically on hazardous drinking among university students, finding a protective effect against alcohol dependence (Sheerin et al., 2021). However, it is not yet clear to what extent it could be protective for students with HED specifically. During the COVID-19 pandemic, studies on stress resilience and alcohol consumption among university students have shown mixed results. While some research indicates that students with lower stress resilience consumed more alcohol (Rubio et al., 2023; Zysset et al., 2022), other studies report no significant impact on alcohol use (Myntti & Armstrong, 2022). Despite evidence that trait stress resilience broadly mitigates alcohol-related risks, further understanding is needed on how it could positively influence hazardous drinking trajectories, particularly during times of stress and transitions. Students with HED, who are particularly vulnerable to dependence, would benefit from targeted prevention efforts and resilience training (Herrero-Montes et al., 2022).

General Aims

The primary goal of this pre-registered study was to longitudinally investigate the predictive effects of loneliness, perceived social support, and trait stress resilience, measured during a lockdown period (Wave 1), on hazardous drinking in university students, measured five times over two years (Wave 1–5). This research is particularly innovative in its focus on students in their last years of university and beyond, emphasizing those with HED. We aim to explore the long-term effects of loneliness, social support, and resilience on two significant outcomes: alcohol consumption and alcohol-related problems among students in their final university years and as they transition out of university. Due to gender-specific variations in how men and women experience the progression to problematic alcohol use, we will assess how our predictors interact with gender in influencing hazardous drinking trajectories.

To our knowledge, this dual-outcome approach has not been comprehensively explored in existing literature, with limited exceptions such as the study by Segrin et al. (2018), which primarily focused on loneliness as the main predictor. Our 2-year longitudinal study seeks to enhance the theoretical understanding of risk factors for hazardous drinking specifically among students with HED, during this critical period of stress and transition. Furthermore, the current investigation aims to identify predictors for interventions that improve social connections and stress management, preventing hazardous drinking in students with HED from escalating into their professional lives.

Methods

The procedures and statistical analyses for this study were pre-registered on the Open Science Framework (OSF), accessible via: https://osf.io/675zd.

Participants

University students were recruited for the study. Eligibility criteria included: (1) attending a Dutch research or applied university, (2) having a good command of the Dutch language, (3) being in the last year of a bachelor programme or any year of a master programme, and (4) reporting HED before the COVID-19 outbreak (i.e., retrospective self-report). HED was defined as consuming at least once a week six or more drinks in a day for men and four or more drinks in a day for women. This definition is based on guidelines from the Trimbos Institute (2017) in the Netherlands. Participants who engaged in this behaviour the month prior to the COVID-19 outbreak were invited to participate in the longitudinal project.

A total of 1176 students were initially screened during the fall semester of 2020. Of these, 634 classified for HED and were invited to participate; 443 (70%) agreed to participate and completed the baseline measure (Wave 1) in mid-February 2021. Retention rates of the 443 participants across the follow-up measures were very high: 95%, 90%, 84%, and 87%, for Waves 2 to 5, respectively. A final sample of 437 participants (M_age at baseline = 22.5 years, SD = 1.8 years; 67% women) was analysed after excluding six individuals for distinct reasons (see Table 1 for more detailed information about the demographic characteristics of the sample at baseline).

Table 1.

Sociodemographic Characteristics of Participants at Baseline.

Baseline characteristics	n	%
Gender^a
Women	291	66.6%
Men	146	33.4%
Programme level^b
Last year university students (bachelor’s)	126	28.8%
Last year applied university (bachelor’s)	62	14.2%
Master’s university students
Year 1	127	29.1%
Year 2	75	17.2%
Year 3	9	2.1%
Pre-Master’s	21	4.8%
Master’s applied university students	4	0.9%
Other education	2	0.5%
Not currently university student^c	11	2.5%
Relationship status
Single	150	34.3%
Committed relationship (cohabitating)	73	16.7%
Committed relationship (living apart)	181	41.4%
Married/partnered	2	0.5%
Dating	26	5.9%
Non-monogamous relationship	2	0.5%
Other	3	0.7%
Living Situation
In a student house	222	50.8%
Alone	59	13.5%
With parents	73	16.7%
With partner	63	14.4%
Other	20	4.6%
Employment Status^d
Student	333	76.2%
Internship	139	31.8%
Full-time employed	5	1.1%
Part-time employed	152	34.8%
Self-employed	19	4.3%
Not working currently	40	9.2%

Note. N = 437. Participants were on average 22.54 years old (SD = 1.8) at Wave 1. Participants were excluded due to technical issues (n = 1) and needing help for drinking problems (n = 3).

^aTwo participants who identified as “diverse” were not included in the analyses because the group was too small.

^bIn the Netherlands, bachelor’s studies in research universities generally last three years, while master’s studies from one (most specializations) to two (technical natural sciences and research masters) years, with some exceptions such as medicine. Between bachelor’s and master’s, some students have to undergo a pre-masters’ if they attended applied universities for their bachelor’s education.

^cThese participants were students in the screening, but no longer so by Wave 1. By Wave 5, 54.2% participants had graduated.

^dParticipants could respond to more than one category for employment status.

Procedure

Potential participants were invited to complete a short screening survey for a two-year online study on lifestyle changes during the transition out of university and into graduation. Everyone who completed the screening was entered into a lottery for €50 coupons, redeemable at a popular Dutch web shop. Recruitment took place from September 2020 to early March 2021 via social media, flyers in student-related venues, university organizations, the participant recruitment system of the university, and snowball sampling.

Eligible participants were invited to the longitudinal study and received personalized links to the online survey in their email at each data wave, using LimeSurvey (Limesurvey GmbH, 2021). For each completed wave, participants received a €10 coupon (with the last wave offering a €5 coupon due to its shorter duration). Data collection took place during different periods of the COVID-19 pandemic in the Netherlands. See Figure 1 for a visualization of the study timeline and COVID-19-related restrictions. Further details on COVID-19 preventive measures during the years of the study can be found in the timeline provided by the Dutch National Institute for Public Health and the Environment (Rijksinstituut voor Volksgezondheid en Milieu, 2022).

Figure 1.

Visualization of the study timeline and COVID-19 related restrictions during data collection.

The data collection for the al-RISCO Lifestyle Project received approval from the Ethics Committee of the Faculty of Social Sciences of Radboud University, Nijmegen, the Netherlands, Approval Code: ECSW-2020-059, Approval Date: 18-June-2020.

Measures

Outcomes

In each wave, the outcome measures were assessed with the Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993), translated to Dutch (Schippers & Broekman, 2010), a 10-item screening test for hazardous alcohol use. The AUDIT has two subscales: 3 items for alcohol consumption (AUDIT-C) and 7 items for problem drinking (AUDIT-P; e.g., Sanchez-Roige et al., 2019). The psychometric properties of the AUDIT have been well-documented (Saunders et al., 1993).

Alcohol consumption

The questions of the AUDIT-C assess the frequency and quantity of drinking, and binge drinking frequency (Bush et al., 1998; Sanchez-Roige et al., 2019; Saunders et al., 1993; Verhoog et al., 2020). Participants’ answers are rated on a scale from 0 to 4 (e.g., 0 = never; 4 = 4 or more times a week). The responses to the three items were summed to a total alcohol consumption score (0–12). Higher scores indicate greater alcohol consumption. The reliability of the AUDIT-C in our study (Cronbach’s α) was .67, .69, .70, .71, and .73 for Waves 1 to 5, respectively.

Alcohol Problems

The questions of the AUDIT-P assess the frequency of problems associated with drinking behaviour, for example: “How often during the last year have you found that you were not able to stop drinking once you had started?, “Have you or someone else been injured as a result of your drinking?”, and “How often during the last year have you been unable to remember what happened the night before because you had been drinking?” (Sanchez-Roige et al., 2019; Saunders et al., 1993). Participants can answer from never (=0) to (almost) daily (=4). The responses to the seven items were summed to a total alcohol problems score (0–28). Higher scores indicate more severe alcohol problems. The reliability of the AUDIT-P in our study (Cronbach’s α) was .63, .67, .64, .67, and .61 for Waves 1 to 5, respectively.

Predictors

Loneliness

The Three-Item Loneliness Scale (UCLA-revised: Hughes et al., 2004; Russell et al., 1980), translated to Dutch (Distel et al., 2010), was employed to measure participants’ loneliness. This shortened version of the Revised UCLA Loneliness scale has three items: “How often do you feel that you lack company?”, “How often do you feel left out?”, and “How often do you feel isolated from others?”. Items were rated on a scale from 1 (almost never) to 3 (often). A mean score of the three items was calculated to indicate the level of loneliness. Higher scores indicated more loneliness. The good reliability and validity of the scale have been established in previous research (Hughes et al., 2004; Russell et al., 1980). The reliability of UCLA-revised in our study (Cronbach’s α) was .69 for Wave 1.

Perceived social support

The Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al., 1988) was used to assess perceived social support. The scale has 12 items, measuring support from family, friends, and a significant other. Participants rated how much they agreed with each item on a scale from 1 (totally disagree) to 7 (totally agree). Scale scores were calculated for support from family, friends, and significant other by averaging the relevant items. Higher scores indicated more social support from each source. The MSPSS has demonstrated good psychometric properties (translated to Dutch by De Clerck, 2009; original version Zimet et al., 1988). The reliability of MSPSS in our study (Cronbach’s α) was .92 for the social support from family subscale, .89 for the friends subscale, and .93 for the significant other subscale, all for Wave 1.

Trait Stress Resilience

Trait stress resilience was assessed with the Brief Resilience Scale (BRS) (Smith et al., 2008), translated to Dutch (Leontjevas et al., 2014), a 6-item scale designed to measure individuals’ ability to bounce back from stress. Participants rated how much they agreed with each statement on a 5-point scale from 1 (strongly disagree) to 5 (strongly agree). Negatively phrased statements were reverse-coded, and a mean score was calculated across the six items. Higher scores indicated more stress resilience. The BRS has demonstrated satisfactory reliability and validity (Smith et al., 2008). The reliability of BRS in our study (Cronbach’s α) was .81 for Wave 1.

Covariates

Gender

Participants were asked to indicate their gender identity. The response options included man (1), woman (2), and diverse (3). Due to the low prevalence of participants identifying as diverse (n = 2), this group was not included in the analyses. Gender was effect coded for the purpose of the analyses (woman = −1, man = +1).

Frequency of alcohol use before COVID-19

The frequency of alcohol use in the 12 months preceding the start of the COVID-19 restrictions was assessed retrospectively at Wave 1. Participants were asked: “How often did you usually drink alcoholic beverages in the 12 months before the COVID-19 measures were applied?”. The question and response categories were adapted based on research guidelines from the National Institute of Alcohol Abuse and Alcoholism (NIAAA) in 2003 (Nugawela et al., 2016). The response options ranged from (1) every day to (9) 1 or 2 times in the year before the COVID-19 restrictions. The answer options were reverse-coded so higher scores indicated more frequent alcohol use before COVID-19. The category “never” was not included as the sample consisted of heavy episodic drinkers before COVID-19. This measure of frequency was used as a covariate in our main analyses to control for drinking behaviour in the year before the COVID-19 outbreak.

Psychometric properties of the Study’s measures

In expanding our assessment beyond reliability metrics like Cronbach’s alpha, we comprehensively evaluated the validity of our measures through theoretical grounding and empirical correlations. Notably, the negative correlations between social support (from family, friends, and a significant other) and loneliness confirmed our constructs’ convergent validity, aligning with theoretical predictions about their relationship. Similarly, stress resilience’s moderate and negative correlation with loneliness demonstrated its theoretical role as a buffer against adverse emotional states. Measure validity was further supported by the positive relationship between stress resilience and social support from friends, highlighting the interconnectedness of these constructs.

For the single-item measure assessing the frequency of alcohol use, its validity was established through moderate to strong positive correlations with the outcomes, indicating a consistent pattern of behaviour. Additionally, the correlation between the outcomes of alcohol consumption and related problems across successive waves demonstrated not only convergent validity but also provided evidence of test-retest reliability, underscoring the robustness of our measures.

Statistical Analyses

To examine changes in alcohol consumption and alcohol problems over two years, Latent Growth Models (LGM) were run on the alcohol consumption and problem scores from Wave 1 to Wave 5 (McArdle & Epstein, 1987; Meredith & Tisak, 1990). The analyses were conducted by using the software R (R Core Team, 2021) and the lavaan package (Rosseel, 2012). Linear and quadratic changes for both dependent variables were explored. Missing data from follow-up waves were handled by using full information likelihood estimation (Enders, 2010). Maximum likelihood with robust standard errors (MLR) was used to account for non-normal distributions. Model fit was evaluated by the comparative fit index (CFI), the Tucker–Lewis fit index (TLI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). Values >.90 for the CFI and TLI and values <.10 for the RMSEA and SRMR are considered acceptable (e.g., Bollen & Curran, 2006; Maciejewski et al., 2015).

The unconditional growth models for alcohol consumption and alcohol problems were run first. They were then extended to examine whether the independent variables predicted linear and quadratic changes in both dependent variables. The first step was to include the effects of the covariates gender, drinking frequency before COVID-19, and the predictors loneliness, trait stress resilience, and social support from from family, friends, and a significant other. In a second step, the interactions between gender and the predictors were incorporated. Variables were first mean-centered, and a product term between the predictors and gender was calculated for the model with interactions (loneliness × gender; social support (for each source of support) × gender; trait stress resilience × gender). A Bonferroni correction adjusted the standard p < .05 criterion for testing two main models (alcohol use and alcohol problems), setting alpha at .05/2 = .025.

Deviations from Pre-registration

In the pre-registration, we planned to analyse data from four waves of the longitudinal project. As an additional wave of data was collected during this research project, a total of five waves of data have now been included in the analyses. This adjustment in the number of waves analysed ensures that the available data are maximized and provides a more comprehensive and robust analysis of the research questions.

Results

Descriptive Statistics

All participating university students engaged in HED in the month before the COVID-19 outbreak. By Wave 1 of the study, 46% of men and 36% of women maintained this hazardous drinking behaviour. By Wave 2, the proportion of heavy episodic drinkers increased to 60% of men and 51% of women, coinciding with a period when COVID-19 measures began to relax. For male students, from Wave 3 (60%), to Wave 4 (53%), to Wave 5 (53%), percentages remained higher than at Wave 1. Fewer women engaged in HED by Wave 3 (46%), Wave 4 (44%), up to Wave 5 (34%), with lower engagement than in Wave 2 (51%). Men and women significantly differed across several waves, with women demonstrating significantly lower HED than men at Wave 1 (p < .04), Wave 3 (p < .015), and Wave 5 (p < .001) (tested using Pearson’s X2 test for categorical variables).

Table 2 presents descriptive statistics for the study outcomes—alcohol consumption and alcohol problems—across all study waves, alongside predictors including loneliness, stress resilience, and social support from from family, friends, and a significant other. It also incorporates the covariate of drinking frequency before COVID-19. The table further outlines the correlations between the predictors and alcohol outcomes. Our analysis highlighted stable, statistically significant positive correlations between alcohol consumption and alcohol-related problems across waves, indicating a consistent relationship of moderate strength. While loneliness showed a negative correlation with alcohol use at specific waves, suggesting that higher loneliness is associated with lower alcohol consumption, stress resilience mainly displayed a positive correlation with alcohol consumption, albeit inconsistently across waves. Notably, significant associations between social support and alcohol outcomes were limited and varied, with only occasional positive correlations observed. For detailed correlation strengths and significance levels across waves, refer to Table 2.

Table 2.

Descriptive Statistics and Correlations for Outcomes and Predictors.

Variable correlations	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. Alcohol consumption (W1)	—
2. Alcohol consumption (W2)	0.68**	—
3. Alcohol consumption (W3)	0.63**	0.79**	—
4. Alcohol consumption (W4)	0.58**	0.72**	0.77**	—
5. Alcohol consumption (W5)	0.57**	0.66**	0.69**	0.79**	—
6. Alcohol problems (W1)	0.42**	0.38**	0.33**	0.26**	0.27**	—
7. Alcohol problems (W2)	0.47**	0.49**	0.43**	0.37**	0.31**	0.71**	—
8. Alcohol problems (W3)	0.40**	0.45**	0.45**	0.36**	0.33**	0.60**	0.73**	—
9. Alcohol problems (W4)	0.40**	0.42**	0.37**	0.42**	0.36**	0.61**	0.69**	0.73**	—
10. Alcohol problems (W5)	0.38**	0.44**	0.41**	0.43**	0.45**	0.60**	0.69**	0.66**	0.71**	—
11. Drink pre-COVID-19^a	0.38**	0.36**	0.37**	0.34**	0.37**	0.24**	0.28**	0.31**	0.30**	0.28**	—
12. Loneliness	−0.12*	−0.07	−0.07	−0.11*	−0.15**	0.05	0.06	0.07	0.08	0.02	−0.06	—
13. Stress resilience	0.16*	0.10*	0.09	0.11*	0.19**	0.09*	−0.01	−0.05	−0.07	−0.02	0.11*	−0.33**	—
14. SS-Significant Other	−0.04	−0.15**	−0.06	−0.05	−0.06	−0.04	−0.11*	−0.02	0.1	0.01	0.07	−0.13**	−0.06	—
15. SS-Family	0.02	0.04	0.05	0.1	0.11*	−0.03	−0.02	0.04	0.01	0.05	0.09	−0.26**	0.06	0.20**	—
16. SS-Friends	0.09*	0.01	−0.02	0.03	0.02	0.02	0.03	−0.01	−0.05	0.01	0.09	−0.38**	0.18**	0.15**	0.30**	—
Variable descriptives
Mean	5.33	6.15	6.03	5.72	5.23	3.99	4.42	4.21	3.90	3.62	5.36	1.73	3.31	5.60	5.58	5.95
SD	2.26	2.2	2.18	2.24	2.24	3.26	3.55	3.39	3.36	2.94	1.02	0.52	0.71	1.35	1.23	0.88
Range	0–11	0–12	0–11	0–11	0–11	0–14	0–16	0–18	0–16	0–17	1–8	1–3	1.5–5	1–7	1–7	2–7

Note. *p < .05. **p < .01

^aFrequency of alcohol use in the 12 months preceding the start of the COVID-19 restrictions.

SS-Significant Other, SS-Family, and SS-Friends represent social support from each source.

Alcohol Consumption and Alcohol Problems Trajectories

First, an unconditional growth model with intercept and linear change only was run for the outcome of alcohol consumption. This model did not fit the data well. Adding the quadratic term significantly improved model fit (p < .001), with satisfactory indexes, CFI = .985, TLI = .975, RMSEA = .089, and SRMR = .028. This model predicted an estimated average alcohol consumption score (AUDIT-C) at the initial measurement (Wave 1) of 5.4, with a significant quadratic decrease over time (q = −0.200, p < .001). The intercept variance of 3.4 points in alcohol consumption suggested individual differences at initial levels (p < .001). Significant variation in the slope indicated individual differences in participants’ change in alcohol consumption over time (p = .001).

Likewise, for alcohol problems, the linear model did not fit the data well. The quadratic curve model had a significantly better fit (p < .001) in all evaluated indexes, CFI = .991, TLI = .985, RMSEA = .066, and SRMR = .027. This model predicted an estimated average alcohol problems score (AUDIT-P) at the initial measurement (Wave 1) of 4.0, with a significant quadratic decrease over time (q = −0.125, p < .001). The intercept variance of 8.3 points and significant variance in the slope for alcohol problems suggest high individual differences at initial levels (p < .001) as well as in participants’ change in alcohol problems over time (p = .003). See Figure 2 for the predicted trajectories of alcohol consumption (2a) and alcohol problems (2b).

Figure 2.

Unconditional growth model for alcohol consumption and alcohol problems: Quadratic growth visualization.

The predictors were then added to the unconditional growth models for alcohol consumption and alcohol problems. The intercept, and the linear and quadratic slopes, were regressed on all the predictors.

Alcohol Consumption Model: Loneliness, Social Support, and Trait Stress Resilience, Drinking Frequency Before COVID-19, and Gender (Interactions) as Predictors of Alcohol Consumption Trajectories

The simplest model is reported, as significant interactions with gender were not observed. Table 3 contains parameter estimates, standard errors, and p-values for the Alcohol Consumption Model. The Alcohol Consumption Model analyses revealed effects of gender and drinking frequency before COVID-19 on the intercept of alcohol consumption. Specifically, being a male student and drinking more frequently before COVID-19 was associated with more alcohol consumption at Wave 1, but not with changes in alcohol consumption over time. No effects of the other predictors were found, neither for the prediction of the intercept nor for the prediction of changes in alcohol consumption over time.

Table 3.

Results of Growth Curve Model of Alcohol Consumption as Self Reported by University Students From Wave 1 to Wave 5 With Main Predictors.

Predictors	Alcohol consumption (AUDIT-C scores)
	Intercept factor			Linear factor			Quadratic factor
	β	SE	p	β	SE	p	β	SE	p
Drinking pre-COVID-19	0.77	0.09	<.001	−0.002	0.07	0.977	−0.002	0.02	0.927
Gender	0.60	0.11	<.001	0.01	0.10	0.906	−0.004	0.02	0.854
Loneliness	−0.02	0.21	0.920	−0.02	0.18	0.916	−0.02	0.04	0.709
Stress resilience	0.05	0.15	0.726	−0.13	0.12	0.256	0.04	0.03	0.100
SS - Significant Other	−0.13	0.07	0.045	−0.05	0.06	0.350	0.01	0.01	0.293
SS - Family	−0.02	0.08	0.819	0.06	0.07	0.356	−0.01	0.02	0.644
SS - Friends	0.23	0.14	0.104	−0.17	0.11	0.109	0.02	0.02	0.338

Note. Significant difference at p ≤ .025.

For more details about the full Alcohol Consumption Model with gender interactions, please see Table S1, in supplemental material, containing parameter estimates, standard errors, and p-values.

Alcohol Problems Model: Loneliness, Social Support, and Trait Resilience, Drinking Frequency Before COVID-19, and Gender (Interactions) as Predictors of Alcohol Problems Trajectories

The simplest model is reported for alcohol problems as well, as significant interactions with gender were not observed. Table 4 contains parameter estimates, standard errors, and p-values for the Alcohol Problems Model. The Alcohol Problems Model revealed effects of gender, drinking frequency before COVID-19, and loneliness, on the intercept of alcohol problems. Specifically, being a man, drinking more frequently before COVID-19, and higher loneliness were associated with more alcohol problems at Wave 1, but not with changes in alcohol problems over time.

Table 4.

Results of Growth Curve Model of Alcohol Problems as Self-Reported by University Students From Wave 1 to Wave 5 With Main Predictors.

Predictors	Alcohol problems (AUDIT-P scores)
	Intercept factor			Linear factor			Quadratic factor
	β	SE	p	β	SE	p	β	SE	p
Drinking pre-COVID-19	0.72	0.16	<.001	0.31	0.11	0.005	−0.07	0.03	0.003
Gender	0.68	0.17	<.001	−0.06	0.15	0.676	0.01	0.03	0.781
Loneliness	0.86	0.34	0.01	−0.12	0.24	0.620	0.01	0.06	0.846
Stress resilience	0.15	0.22	0.494	−0.51	0.18	0.004	0.10	0.04	0.014
SS - Significant Other	−0.09	0.11	0.400	−0.15	0.08	0.078	0.04	0.02	0.038
SS - Family	−0.07	0.13	0.616	0.08	0.11	0.433	−0.01	0.02	0.594
SS - Friends	0.33	0.18	0.065	−0.11	0.14	0.416	0.01	0.03	0.664

Note. Significant difference at p ≤ .025.

For the predictors of alcohol problems over time, the analyses revealed main effects of drinking frequency before COVID-19 and stress resilience on the linear and quadratic terms. The effects for drinking frequency before COVID-19 were positive on the linear term and negative on the quadratic term. This quadratic trajectory suggests that, while all university students showed a gradual decrease in alcohol problems, those with greater drinking frequency before COVID-19 stayed higher than those with average and low frequency (for a visualization of this trajectory, please see Figure 3(a)). The effects for stress resilience were negative on the linear term and positive on the quadratic term. This quadratic trajectory indicates that, for university students with higher resilience, the initial decline in problems is steeper than for those with average or lower resilience (for a visualization of this trajectory, please see Figure 3(b)).

Figure 3.

(a). Line Graph Depicting Average Trajectories of University Students’ Alcohol Problems from Wave 1 to Wave 5 with Drinking Frequency Before COVID-19 Scores 1 Standard Deviation Above and Below the Mean. Note. The original model intercept represents the predicted value of problem drinking when continuous predictors are mean-centered (set to 0), with a model intercept of 4.3 reflecting standard conditions. Conversely, the plot’s intercept portrays outcomes under specific predictor values, where predictors' regression coefficients are multiplied by their respective averages for visualization. Therefore, difference between model intercept and plot intercept arises from plotting the impact of Drinking Frequency Before COVID-19 on problem drinking across diverse scenarios (mean, one standard deviation above, and one standard deviation below the mean). (b). Line Graph Depicting Average Trajectories of University Students’ Alcohol Problems from Wave 1 to Wave 5 with Trait Stress Resilience Scores 1 Standard Deviation Above and Below the Mean. Note. The original model intercept represents the predicted value of problem drinking when continuous predictors are mean-centered (set to 0), with an original intercept of 4.3 reflecting standard conditions. Conversely, the plot’s intercept portrays outcomes under specific predictor values, where predictors’ regression coefficients are multiplied by their respective averages for visualization. Thereby, difference between model intercept and plot intercept arises from plotting the impact of Trait Stress Resilience on problem drinking across diverse scenarios (mean, one standard deviation above, and one standard deviation below the mean).

For more details about the full Alcohol Problems Model with gender interactions, please see Table S2, in supplemental material, containing parameter estimates, standard errors, and p-values.

Discussion

The present pre-registered 5-wave longitudinal study aimed to unravel the associations between loneliness, perceived social support (from family, friends, and a significant other), and trait stress resilience as predictors of changes in hazardous drinking trajectories among university students in the Netherlands. This study specifically focused on final-year students who engaged in heavy episodic drinking (HED) before the COVID-19 outbreak. Our research examined the longitudinal impact of these factors on two separate hazardous drinking outcomes: alcohol consumption and alcohol problems. This period, marked by significant stress and transitions, not least due to the disruptions caused by COVID-19 restrictions, provides a unique backdrop for investigating how these predictors influence students’ hazardous drinking behaviours.

Results revealed that, at the beginning of the study, male students and those with a history of more frequent drinking before the COVID-19 outbreak exhibited higher levels of alcohol consumption and related problems. Elevated loneliness scores were initially linked to increased alcohol problems. Notably, higher trait stress resilience emerged as the only protective factor associated with a decline in alcohol problems over time. Social support did not significantly influence the trajectories of alcohol consumption or problems. Additionally, higher drinking frequency before the COVID-19 outbreak predicted smaller decreases in alcohol problems over time.

It should be noted that although the participants in our study were not engaging in HED as frequently by the end of data collection as before the outbreak, the high prevalence remains a cause for concern. A considerable percentage of the student sample, more than half of the men and 34% of the women, still engaged in these behaviours once a week or more after two years of data collection. Considering that, by the last wave of our study, participants were on the path towards graduation, this elevated prevalence of HED demand attention.

Alcohol Consumption and Alcohol Problem Trajectories

The latent growth curve analyses revealed similar trajectories for alcohol consumption and alcohol problems among university students. Following the second lockdown in Winter 2021 (Wave 2), both measures initially increased, but later showed only modest decreases. Previous studies offer mixed insights; while Bewick et al. (2008) noted declines in alcohol consumption through undergraduate studies, they also reported consistently high levels of weekly drinking among a sizable number of participants. Arria et al. (2016) observed a decrease in the amount but an increase in the frequency of drinking as students approached graduation—a trend that often continued post-graduation, challenging the expected maturing out pattern. Conversely, Lee et al. (2013) identified a maturing out trend from late adolescence to young adulthood, primarily reflected in high-risk drinkers shifting to moderate-risk drinking.

Our study, conducted during and after the onset of COVID-19, indicates that the most significant decreases in hazardous drinking among university students occurred early in the pandemic, with only small changes observed in subsequent phases spanning both the pandemic and post-pandemic periods. These initial reductions may align with significant reductions in hazardous drinking observed among students in the Netherlands during the first lockdown (Rubio et al., 2023; van Hooijdonk et al., 2022). However, with the gradual relaxation of pandemic measures, these reductions have stabilized and, in some instances, have even reversed. Vink (2023) observed a ‘rebound effect' among university students in the post-restriction period. This suggests that students might have decreased the most during the lockdown, but further decreases could have been disrupted by the easing of restrictions. Similar to the findings of Lee et al. (2013), the current study’s high-risk drinkers might have transitioned to moderate-risk drinking.

Our findings lead to two hypotheses regarding the impact of COVID-19 on hazardous drinking trajectories among students in the last years of university and beyond. The first suggests the pandemic may have obscured steeper declines in hazardous drinking, potentially delaying expected maturing out trends. The second proposes that modest reductions might be a temporary return to the mean rather than a true beginning of maturing out. Distinguishing between these outcomes is crucial. If we are witnessing a (delayed) maturing out trend, this insight would allow prevention efforts to be more finely tuned, possibly by targeting younger demographics or by strengthening the positive transitions of young adults as they leave university. On the other hand, if the initial reduction in hazardous drinking does not stem from natural developmental changes but rather a return to the mean, it suggests that public health strategies to mitigate hazardous drinking among university students have not yet reached their intended effectiveness. This realization underscores the necessity for innovative prevention efforts. Additionally, it points to the potential need for further support and resources for young adults as they adapt to post-pandemic life, with a particular focus on enhancing resilience and coping strategies.

Predictors of Alcohol Consumption Trajectories

The predictors we assessed, including gender interactions, did not significantly impact alcohol consumption trajectories. Interestingly, despite observing differences between genders in two of the studied predictors, namely loneliness and stress resilience, these variations did not translate into substantial effects on hazardous drinking trajectories. This suggests that influences from the university environment and social norms might overshadow the role of gender in shaping drinking behaviours, even as students approach graduation (Neighbors et al., 2007). Future studies could explore gender-based differences in subsequent developmental phases to gain significant insights into these patterns.

Our findings indicate that loneliness, similar to what was observed in Segrin et al. (2018), is not directly linked to alcohol consumption. Regarding social support, we found that none of the sources significantly impacted alcohol consumption trajectories, contrasting with the findings of Lechner et al. (2020), who noted a protective effect against increased drinking from social support during the first COVID-19 lockdown. This discrepancy calls for caution in generalizing Lechner et al.'s findings to other pandemic phases or the immediate post-lockdown periods.

Furthermore, our examination of trait stress resilience did not predict changes in alcohol consumption trajectories. The only other longitudinal study on this topic, by Sheerin et al. (2021), focused on dependence rather than alcohol use specifically. Thus, while more research is needed for definitive conclusions, trait stress resilience might not protect students against consumption but may guard against more problematic outcomes. Other factors, such as family history, personality traits like impulsivity and sensation seeking, and greater emotional stability, appear to be more strongly associated with changes in alcohol consumption over time (Ashenhurst et al., 2015; Gotham et al., 1997; Haardörfer et al., 2021; Jackson et al., 2001). Looking ahead, it is worthwhile to explore potential mediating and moderating mechanisms between these risk factors, social relationships, and stress resilience to understand their interactions with alcohol consumption trajectories.

Predictors of Alcohol Problems’ Trajectories

In contrast to alcohol consumption, our longitudinal analysis revealed that loneliness and trait stress resilience influenced alcohol problems, but perceived social support did not. Similar to the alcohol consumption model, the interactions between loneliness, stress resilience, social support, and gender did not influence alcohol problems. While previous research found evidence of loneliness predicting alcohol problems (Herchenroeder et al., 2022; Segrin et al., 2018), our study did not support this finding. However, we did find that loneliness was specifically associated with alcohol problems at baseline. Loneliness and alcohol problems may be closely linked through a general negative perception of one’s self and relationships, as suggested by Akerlind & Hörnquist (1992).

These negative feelings could potentially trigger or sustain alcohol-related problems as individuals turn to alcohol to cope. Interestingly, loneliness did not predict problem drinking trajectories. Possibly, feelings of loneliness at the beginning of the study were heightened due to COVID-19-related measures, but this effect did not persist throughout the duration of the study. Given these observations, future studies could benefit from a more detailed approach to examine how fluctuations in loneliness affect problem drinking over time. To achieve this, incorporating daily diary methods or ecological momentary assessment (EMA) would be particularly useful. These methods can capture day-to-day variations in loneliness and its immediate impacts on hazardous drinking, offering deeper insights into the dynamic relationship between emotional states and substance use.

Conversely, trait stress resilience emerged as a protective factor against continued problem drinking among university students in the last years of university and beyond. This protective effect held even when accounting for loneliness, social support, gender, and pre-COVID-19 drinking frequency. Students with higher stress resilience exhibited a steeper decline in problem drinking compared to their counterparts with average or lower stress resilience. These findings align with previous research, further supporting the notion that individuals with greater stress resilience are more protected against the progression of alcohol problems or the development of alcohol disorders over time (Lyvers et al., 2020; Sanchez et al., 2022; Sheerin et al., 2021). Unexpectedly, the effect of stress resilience appeared to diminish by the last wave of our study. Given that stress resilience was treated as a static trait measured at the beginning of the study, this change suggests its dynamic nature, as posited by researchers like Kalisch et al. (2017). The authors argue that resilience can evolve depending on how individuals adapt to stressors at various points in their lives (Kalisch et al., 2017). Consequently, future research should delve deeper into how stress resilience interacts with specific stressors—such as study delays, financial challenges, and changes in relationships—to more accurately gauge its influence on problem drinking trajectories among students nearing graduation. This line of investigation will help to uncover the nuanced ways in which stress resilience impacts drinking behaviour over time.

Lastly, similar to what was observed for alcohol consumption, perceived social support did not significantly predict problem drinking trajectories. This lack of impact could be attributed to several factors. First, the efficacy of social support may be time-specific, potentially playing a crucial role during the initial stages of the pandemic and the first COVID-19 lockdown when students were more isolated. Second, the student sample was already engaged in HED at the start of the study. This pre-existing pattern of consumption could have overshadowed any mitigative effects that social support might typically provide. The students may have been influenced by their immediate social circles, who exhibited similar drinking behaviours. Future research should investigate whether the timing of social support, relative to the onset of heavy drinking behaviours, affects its efficacy as a protective factor against problem drinking. This approach will help determine if earlier or differently timed interventions could more effectively prevent or mitigate problem drinking.

Gender and Drinking Frequency Before COVID-19 as Robust Factors in Alcohol Consumption and Alcohol Problems

Consistent with broader patterns observed in our study, gender and drinking frequency before COVID-19 were identified as robust predictors of alcohol consumption and alcohol problem trajectories. Specifically, men consumed more alcohol than women and encountered more alcohol-related problems (Ashenhurst et al., 2015; Haardörfer et al., 2021; Jackson et al., 2001), which aligns with the sustained prevalence of heavy episodic drinking noted particularly among men. These findings raise important questions about the persistence of high drinking rates among men throughout the study, suggesting a need to explore how gender socialization contributes to these behaviours and to develop targeted prevention strategies. Specifically, it indicates that male students might benefit from selective, evidence-based prevention efforts tailored to address the unique influences and expectations that sustain their hazardous drinking.

Our study found that students who were frequent drinkers before the onset of the COVID-19 pandemic had higher alcohol consumption initially and continued to experience problems throughout the study. Although there was an initial reduction in drinking problems due to lockdown measures, a ‘rebound effect' was observed as restrictions eased. Over time, even the most frequent drinkers began to show signs of moderating their drinking problems, yet they maintained relatively higher levels compared to those with lower pre-pandemic frequency. This suggests a potential, albeit delayed, shift towards maturing out, supporting our hypothesis about the temporary impact of pandemic conditions on drinking behaviours. The pandemic may have delayed, but not entirely prevented, the natural process of maturing out of hazardous drinking problems among university students.

Acknowledgment of Limitations

Despite the valuable insights provided by our study, it is crucial to acknowledge its limitations for a comprehensive understanding of the scope and applicability of our findings. Firstly, hazardous drinking is deeply influenced by cultural norms and practices. For example, institutions in the Netherlands (i.e., Trimbos Instituut, 2022) have observed a decrease in hazardous drinking beginning around the age of 25, often coinciding with the transition to work-life. This shift in lifestyle may not mirror the experiences of young graduates in other countries. Additionally, the COVID-19 restrictions specific to the Netherlands during our study may limit the replicability of our results in areas with different pandemic management. These considerations are crucial for appropriately generalizing our findings beyond the Dutch context.

While our sample size was adequate for exploring high-risk behaviours among university students, a larger sample could have provided a broader exploration of our findings. For instance, it would have allowed for a more detailed investigation of the quadratic relationships observed, potentially shaped by COVID-19 restrictions. Additionally, incorporating mixture growth trajectories into our models could further elucidate how predictors impact different clusters of hazardous drinking. Moreover, our reliance on self-reported measures introduces the potential for bias, despite the use of well-established questionnaires and validated instruments.

Implications for Targeted Interventions in University Populations

To build on our findings, it is crucial to concentrate on students who exhibit at-risk drinking behaviours, especially those with higher problem drinking scores. These individuals are particularly vulnerable to hazardous drinking during periods of increased stress and transitions and heightened risk of social isolation. Implementing targeted screenings with tools like the AUDIT can facilitate early interventions. Specifically focusing on problem drinking scores will help identify students at greater risk of drinking in response to stress and loneliness.

Our findings suggest that interventions focusing on stress management would be particularly effective for students with a history of HED, especially as they approach the critical transition into professional life. Additionally, programmes aimed at enhancing building connections and social confidence should be prioritized to mitigate immediate risks, especially among students who experience loneliness. Although this issue was only cross-sectionally associated with problem drinking, actively addressing it can significantly reduce the impact. Furthermore, there should be a separate focus on male students who are consistently at higher risk for heavy consumption and related problems. This heightened risk could relate to cultural and gender role expectations that influence hazardous drinking. By enhancing social networks beyond merely ‘drinking buddies' and potentially reducing reliance on alcohol as a stress relief strategy, interventions not only support individual students but also contribute to a healthier campus environment and help prevent unhealthier transitions into professional life.

Conclusions

Our pre-registered 5-wave longitudinal study has provided valuable and innovative insights into hazardous drinking trajectories among university students with HED in the last years of university and beyond, amid the COVID-19 pandemic in the Netherlands. Specifically, our research suggests that hazardous drinking is higher among risk groups such as students experiencing greater loneliness, male students, and more frequent drinkers prior to the COVID-19 outbreak. Notably, men are generally considered an at-risk group for higher consumption and problems, a pattern that holds true independently of other psychological factors. Concerning hazardous drinking trajectories, prior drinking frequency and trait stress resilience are important predictors of problem drinking over time. A surprising finding of the present study was that social support, independent of its source, was not associated with hazardous drinking. Furthermore, our results underscore the importance of treating alcohol consumption and related problems as distinct issues. This distinction is crucial for developing interventions that focus on addressing predictors of problems arising from drinking, rather than just the amount consumed.

Supplemental Material

Supplemental Material - From Risk to Resilience? Hazardous Drinking Trajectories in and Beyond the Last Years of University Life

Supplemental Material for From Risk to Resilience? Hazardous Drinking Trajectories in and Beyond the Last Years of University Life by Milagros Rubio, Antonius H. N. Cillessen, Maartje Luijten, Jacqueline M. Vink, and Maaike Verhagen in Emerging Adulthood.

Supplemental Material

Supplemental Material - From Risk to Resilience? Hazardous Drinking Trajectories in and Beyond the Last Years of University Life

Footnotes

Authors Note

The data that support the findings of this study are available on request from co-author, Jacqueline Vink (jacqueline.vink@ru.nl).

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is supported by ‘t Trekpaert foundation and Behavioural Science Institute.

Ethical Statement

ORCID iD

Milagros Rubio

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Milagros Rubio is currently a Ph.D. candidate at Radboud University’s Behavioural Science Institute, within the Developmental Psychopathology research group. The author holds a Master’s in Clinical and Cognitive Neuroscience, specializing in Neuropsychology, from Maastricht University (2019), after completing a Psychology degree in Argentina. Her Ph.D. project focuses on hazardous drinking among university students approaching graduation, emphasizing psycho-social risk and resilience factors.

Antonius H. N. Cillessen is Professor and Chair of developmental psychology in the Behavioural Science Institute at Radboud University in Nijmegen, The Netherlands. He previously held positions at Duke University and the University of Connecticut. His research interests include social status, peer group relations, aggression and antisocial behavior, and research methods for developmental psychology.

Maartje Luijten is Associate Professor in the Behavioural Science Institute at Radboud University in Nijmegen, The Netherlands. She conducted her Ph.D. research at the Erasmus University Rotterdam and was a visiting research fellow at Melbourne University and Cambridge University. Her research interests include substance use, well-being, mental health, and resilience of young adults.

Jacqueline M. Vink is Professor and Chair of developmental psychopathology in the Behavioural Science Institute at Radboud University in Nijmegen, The Netherlands. She conducted her Ph.D. research in the Biological Psychology department at the VU University in Amsterdam, the Netherlands. She is interested in genetic and epidemiological research in the area of substance use, addiction, food and mental health, with a focus on the developmental perspective in adolescents and young adults.

Maaike Verhagen is Associate Professor in the Behavioural Science Institute at Radboud University in Nijmegen, The Netherlands. She conducted her Ph.D. research at the Radboud University Medical Center, the Netherlands. Her research interests include loneliness, social health, mental health and emotion regulation processes in adolescents.

References

Akerlind

Hörnquist

J. O.

(1992). Loneliness and alcohol abuse: A review of evidences of an interplay. Social Science & Medicine, 34(4), 405–414. https://doi.org/10.1016/0277-9536(92)90300-f

Arria

A. M.

Caldeira

K. M.

Allen

H. K.

Vincent

K. B.

Bugbee

B. A.

O’Grady

K. E.

(2016). Drinking like an adult? Trajectories of alcohol use patterns before and after college graduation. Alcoholism: Clinical and Experimental Research, 40(3), 583–590. https://doi.org/10.1111/acer.12973

Ashenhurst

J. R.

Harden

K. P.

Corbin

W. R.

Fromme

(2015). Trajectories of binge drinking and personality change across emerging adulthood. Psychology of Addictive Behaviors, 29(4), 978–991. https://doi.org/10.1037/adb0000116

Babor

T. F.

Higgins-Biddle

J. C.

Saunders

J. B.

Monteiro

M. G.

(2001). AUDIT: The alcohol use disorders identification test: Guidelines for use in primary health care. World Health Organization.

Bewick

B. M.

Mulhern

Barkham

Trusler

Hill

A. J.

Stiles

W. B.

(2008). Changes in undergraduate student alcohol consumption as they progress through university. BMC Public Health, 8(1), 163. https://doi.org/10.1186/1471-2458-8-163

Bollen

K. A.

Curran

P. J.

(2006). Latent curve models: A structural equation perspective. Wiley.

Bush

Kivlahan

D. R.

McDonell

M. B.

Fihn

S. D.

Bradley

K. A.

& Ambulatory Care Quality Improvement Project (ACQUIP) . (1998). The AUDIT alcohol consumption questions (AUDIT-C): An effective brief screening test for problem drinking. Ambulatory care quality improvement project (ACQUIP). Alcohol use disorders identification test. Archives of Internal Medicine, 158(16), 1789–1795. https://doi.org/10.1001/archinte.158.16.1789

Cusack

S. E.

Wright

A. W.

Amstadter

A. B.

(2023). Resilience and alcohol use in adulthood in the United States: A scoping review. Preventive Medicine, 168, Article 107442. https://doi.org/10.1016/j.ypmed.2023.107442

De Clerck

(2009). Het psychosociaal welzijn bij Congolese ex-kindsoldaten: De rol van sociale steun. Ghent University Library.

10.

Distel

M. A.

Rebollo-Mesa

Abdellaoui

Derom

C. A.

Willemsen

Cacioppo

J. T.

Boomsma

D. I.

(2010). Familial resemblance for loneliness. Behavior Genetics, 40(4), 480–494. https://doi.org/10.1007/s10519-010-9341-5

11.

Egerton

G. A.

Read

J. P.

(2019). Relationship role transitions and problem alcohol use in emerging adulthood. Emerging Adulthood, 7(4), 291–303. https://doi.org/10.1177/2167696818772950

12.

Enders

C. K.

(2010). Applied missing data analysis (pp. xv, 377). Guilford Press.

13.

Gotham

H. J.

Sher

K. J.

Wood

P. K.

(1997). Predicting stability and change in frequency of intoxication from the college years to beyond: Individual-difference and role transition variables. Journal of Abnormal Psychology, 106(4), 619–629. https://doi.org/10.1037/0021-843X.106.4.619

14.

Groh

D. R.

Jason

L. A.

Davis

M. I.

Olson

B. D.

Ferrari

J. R.

(2007). Friends, family, and alcohol abuse: An examination of general and alcohol-specific social support. American Journal on Addictions, 16(1), 49–55. https://doi.org/10.1080/10550490601080084

15.

Haardörfer

Windle

Fairman

R. T.

Berg

C. J.

(2021). Longitudinal changes in alcohol use and binge-drinking among young-adult college students: Analyses of predictors across system levels. Addictive Behaviors, 112(3), Article 106619. https://doi.org/10.1016/j.addbeh.2020.106619

16.

Herchenroeder

Post

S. M.

Stock

M. L.

Yeung

E. W.

(2022). Loneliness and alcohol-related problems among college students who report binge drinking behavior: The moderating role of food and alcohol disturbance. International Journal of Environmental Research and Public Health, 19(21), Article 13954. https://doi.org/10.3390/ijerph192113954

17.

Herrero-Montes

Alonso-Blanco

Paz-Zulueta

Pellico-López

Ruiz-Azcona

Sarabia-Cobo

Boixadera-Planas

Parás-Bravo

(2022). Excessive alcohol consumption and binge drinking in college students. PeerJ, 10, Article e13368. https://doi.org/10.7717/peerj.13368

18.

Hingson

Zha

Smyth

(2017). Magnitude and trends in heavy episodic drinking, alcohol-impaired driving, and alcohol-related mortality and overdose hospitalizations among emerging adults of college ages 18-24 in the United States, 1998-2014. Journal of Studies on Alcohol and Drugs, 78(4), 540–548. https://doi.org/10.15288/jsad.2017.78.540

19.

Holt-Lunstad

(2017). The potential public health relevance of social isolation and loneliness: Prevalence, epidemiology, and risk factors. Public Policy & Aging Report, 27(4), 127–130. https://doi.org/10.1093/ppar/prx030

20.

Hughes

M. E.

Waite

L. J.

Hawkley

L. C.

Cacioppo

J. T.

(2004). A short scale for measuring loneliness in large surveys. Research on Aging, 26(6), 655–672. https://doi.org/10.1177/0164027504268574

21.

Jackson

K. M.

(2008). Heavy episodic drinking: Determining the predictive utility of five or more drinks. Psychology of Addictive Behaviors: Journal of the Society of Psychologists in Addictive Behaviors, 22(1), 68–77. https://doi.org/10.1037/0893-164X.22.1.68

22.

Jackson

K. M.

Sher

K. J.

Gotham

H. J.

Wood

P. K.

(2001). Transitioning into and out of large-effect drinking in young adulthood. Journal of Abnormal Psychology, 110(3), 378–391. https://doi.org/10.1037/0021-843X.110.3.378

23.

Jason

Stevens

Ferrari

Thompson

Legler

(2012). Social networks among residents in recovery homes. Advances in Psychology Study. 1(3), 4–12. https://www.semanticscholar.org/paper/Social-Networks-among-Residents-in-Recovery-Homes.-Jason-Stevens/fe966e14621f80cb2535a97d6ce3016fcf7becf7

24.

Kalisch

Baker

D. G.

Basten

Boks

M. P.

Bonanno

G. A.

Brummelman

Chmitorz

Fernàndez

Fiebach

C. J.

Galatzer-Levy

Geuze

Groppa

Helmreich

Hendler

Hermans

E. J.

Jovanovic

Kubiak

Lieb

Lutz

Kleim

(2017). The resilience framework as a strategy to combat stress-related disorders. Nature Human Behaviour, 1(11), 784–790. https://doi.org/10.1038/s41562-017-0200-8

25.

Lechner

W. V.

Laurene

K. R.

Patel

Anderson

Grega

Kenne

D. R.

(2020). Changes in alcohol use as a function of psychological distress and social support following COVID-19 related University closings. Addictive Behaviors, 110, Article 106527. https://doi.org/10.1016/j.addbeh.2020.106527

26.

Lee

M. R.

Chassin

Villalta

I. K.

(2013). Maturing out of alcohol involvement: Transitions in latent drinking statuses from late adolescence to adulthood. Development and Psychopathology, 25(4 Pt 1), 1137–1153. https://doi.org/10.1017/S0954579413000424

27.

Lee

M. R.

Sher

K. J.

(2018). “Maturing Out” of binge and problem drinking. Alcohol Research: Current Reviews, 39(1), 31–42.

28.

Leontjevas

de Beek

W. O.

Lataster

Jacobs

(2014). Resilience to affective disorders: A comparative validation of two resilience scales. Journal of Affective Disorders, 168, 262–268. https://doi.org/10.1016/j.jad.2014.07.010

29.

Limesurvey GmbH . (2021). LimeSurvey: An open source survey tool (Version 5.2.3). LimeSurvey GmbH. https://www.limesurvey.org/

30.

Lyvers

Holloway

Needham

Thorberg

F. A.

(2020). Resilience, alexithymia, and university stress in relation to anxiety and problematic alcohol use among female university students. Australian Journal of Psychology, 72(1), 59–67. https://doi.org/10.1111/ajpy.12258

31.

Maciejewski

D. F.

van Lier

P. A. C.

Branje

S. J. T.

Meeus

W. H. J.

Koot

H. M.

(2015). A 5-year longitudinal study on mood variability across adolescence using daily diaries. Child Development, 86(6), 1908–1921. https://doi.org/10.1111/cdev.12420

32.

McArdle

J. J.

Epstein

(1987). Latent growth curves within developmental structural equation models. Child Development, 58(1), 110–133. https://doi.org/10.2307/1130295

33.

Meredith

Tisak

(1990). Latent curve analysis. Psychometrika, 55(1), 107–122. https://doi.org/10.1007/BF02294746

34.

Myntti

W. W.

Armstrong

S. B.

(2022). Bouncing back from COVID-19: The role of resilience in depression, substance use, and loneliness in college students amidst the pandemic. J Am Coll Health, 1–10. https://doi.org/10.1080/07448481.2022.2135377. 36328785.

35.

Neighbors

Lee

C. M.

Lewis

M. A.

Fossos

Larimer

M. E.

(2007). Are social norms the best predictor of outcomes among heavy-drinking college students? Journal of Studies on Alcohol and Drugs, 68(4), 556–565. https://doi.org/10.15288/jsad.2007.68.556

36.

NIAAA . (2023). Harmful and underage college drinking. National Institute on Alcohol Abuse and Alcoholism (NIAAA). https://www.niaaa.nih.gov/publications/brochures-and-fact-sheets/college-drinking

37.

Nugawela

M. D.

Langley

Szatkowski

Lewis

(2016). Measuring alcohol consumption in population surveys: A review of international guidelines and comparison with surveys in England. Alcohol and Alcoholism, 51(1), 84–92. https://doi.org/10.1093/alcalc/agv073

38.

Perlman

Peplau

(1981). Toward a social psychology of loneliness Personal relationships 3. Personal Relationships in Disorder, 3, 31–43.

39.

Petker

T. D.

Halladay

Vedelago

Ferro

M. A.

Tucker

Goldman

M. S.

Murphy

J. G.

MacKillop

(2019). Predictors of aging out of heavy episodic drinking in emerging adults: A systematic review protocol. Systematic Reviews, 8(1), 230. https://doi.org/10.1186/s13643-019-1139-9

40.

Pocuca

London-Nadeau

Geoffroy

M.-C.

Chadi

Séguin

J. R.

Parent

Boivin

Tremblay

R. E.

Côté

S. M.

Castellanos-Ryan

(2022). Changes in emerging adults’ alcohol and cannabis use from before to during the COVID-19 pandemic: Evidence from a prospective birth cohort. Psychology of Addictive Behaviors: Journal of the Society of Psychologists in Addictive Behaviors, 36(7), 786–797. https://doi.org/10.1037/adb0000826

41.

R Core Team . (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/

42.

Richardson

Elliott

Roberts

(2017). Relationship between loneliness and mental health in students. Journal of Public Mental Health, 16(2), 48–54. https://doi.org/10.1108/JPMH-03-2016-0013

43.

Rijksinstituut voor Volksgezondheid en Milieu . (2022). Tijdlijn van coronamaatregelen 2022. RIVM. https://www.rivm.nl/gedragsonderzoek/tijdlijn-maatregelen-covid-2022

44.

Rosseel

(2012). lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48(2), 1–36. https://doi.org/10.18637/jss.v048.i02

45.

Rubio

van Hooijdonk

Luijten

Kappe

Cillessen

A. H. N.

Verhagen

Vink

J. M.

(2023). University students’ (binge) drinking during COVID-19 lockdowns: An investigation of depression, social context, resilience, and changes in alcohol use. Soc Sci Med, 326, 115925. https://doi.org/10.1016/j.socscimed.2023.115925

46.

Russell

Peplau

L. A.

Cutrona

C. E.

(1980). The revised UCLA Loneliness Scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39(3), 472–480. https://doi.org/10.1037//0022-3514.39.3.472

47.

Sanchez

Gainza Perez

Field

C. A.

(2022). The role of resilience in alcohol use, drinking motives, and alcohol-related consequences among Hispanic college students. The American Journal of Drug and Alcohol Abuse, 48(1), 100–109. https://doi.org/10.1080/00952990.2021.1996584

48.

Sanchez-Roige

Palmer

A. A.

Fontanillas

Elson

S. L.

Adams

M. J.

Howard

D. M.

Edenberg

H. J.

Davies

Crist

R. C.

Deary

I. J.

McIntosh

A. M.

Clarke

T. K.

(2019). Genome-wide association study meta-analysis of the alcohol use disorders identification test (AUDIT) in two population-based cohorts. American Journal of Psychiatry, 176(2), 107–118. https://doi.org/10.1176/appi.ajp.2018.18040369

49.

Saunders

J. B.

Aasland

O. G.

Babor

T. F.

de la Fuente

J. R.

Grant

(1993). Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption--II. Addiction, 88(6), 791–804. https://doi.org/10.1111/j.1360-0443.1993.tb02093.x

50.

Schippers

G. M.

Broekman

T. G.

(2010). Audit: Alcohol use disorders identification test. Nederlandse Vertaling.

51.

Segrin

McNelis

Pavlich

C. A.

(2018). Indirect effects of loneliness on substance use through stress. Health Communication, 33(5), 513–518. https://doi.org/10.1080/10410236.2016.1278507

52.

Sheerin

C. M.

Bountress

K. E.

Hicks

T. A.

Lind

M. J.

Aggen

S. H.

Kendler

K. S.

Amstadter

A. B.

(2021). Longitudinal examination of the impact of resilience and stressful life events on alcohol use disorder outcomes. Substance Use & Misuse, 56(9), 1346–1351. https://doi.org/10.1080/10826084.2021.1922454

53.

Smith

B. W.

Dalen

Wiggins

Tooley

Christopher

Bernard

(2008). The brief resilience scale: Assessing the ability to bounce back. International Journal of Behavioral Medicine, 15(3), 194–200. https://doi.org/10.1080/10705500802222972

54.

Tinajero

Cadaveira

Rodríguez

M. S.

Páramo

M. F.

(2019). Perceived social support from significant others among binge drinking and polyconsuming Spanish university students. International Journal of Environmental Research and Public Health, 16(22), 4506. https://doi.org/10.3390/ijerph16224506

55.

Trimbos Institute . (2017). Zwaar drinken. Trimbos-Instituut. https://www.trimbos.nl/begrippenlijst/zwaar-drinken/

56.

Trimbos Instituut . (2022). Cijfers alcoholgebruik volwassenen—Expertisecentrum alcohol. Trimbos-instituut. https://www.trimbos.nl/kennis/alcohol/alcohol-in-cijfers/cijfers-alcoholgebruik-volwassenen/

57.

van Gils

Dom

Dierckx

Van Alphen

S. P. J.

Franck

(2022). Resilience, depression and anxiety and hazardous alcohol use behaviour among community dwelling older adults. Aging & Mental Health, 26(1), 56–64. https://doi.org/10.1080/13607863.2020.1870211

58.

van Hooijdonk

K. J. M.

Rubio

Simons

S. S. H.

van Noorden

T. H. J.

Luijten

Geurts

S. A. E.

Vink

J. M.

(2022). Student-, Study- and COVID-19-Related Predictors of Students’ Smoking, Binge Drinking and Cannabis Use before and during the Initial COVID-19 Lockdown in The Netherlands. Int J Environ Res Public Health, 19(2). https://doi.org/10.3390/ijerph19020812

59.

Veer

I. M.

Riepenhausen

Zerban

Wackerhagen

Puhlmann

L. M. C.

Engen

Köber

Bögemann

S. A.

Weermeijer

Uściłko

Mor

Marciniak

M. A.

Askelund

A. D.

Al-Kamel

Ayash

Barsuola

Bartkute-Norkuniene

Battaglia

Bobko

Kalisch

(2021). Psycho-social factors associated with mental resilience in the Corona lockdown. Translational Psychiatry, 11(1), 67. https://doi.org/10.1038/s41398-020-01150-4

60.

Verhoog

Dopmeijer

J. M.

de Jonge

J. M.

van der Heijde

C. M.

Vonk

Bovens

R. H. L. M.

de Boer

M. R.

Hoekstra

Kunst

A. E.

Wiers

R. W.

Kuipers

M. A. G.

(2020). The use of the alcohol use disorders identification test—consumption as an indicator of hazardous alcohol use among university students. European Addiction Research, 26(1), 1–9. https://doi.org/10.1159/000503342

61.

Vink

(2023). Na corona drinken studenten eerder meer dan minder. Sociale Vraagstukken. https://www.socialevraagstukken.nl/na-corona-drinken-studenten-eerder-meer-dan-minder/

62.

Wakabayashi

Sugiyama

Takada

Kinjo

Iso

Tabuchi

(2022). Loneliness and increased hazardous alcohol use: Data from a nationwide internet survey with 1-year follow-up. International Journal of Environmental Research and Public Health, 19(19), Article 12086. https://doi.org/10.3390/ijerph191912086

63.

Wechsler

Nelson

T. F.

(2001). Binge drinking and the American college students: What’s five drinks? Psychology of Addictive Behaviors, 15(4), 287–291. https://doi.org/10.1037/0893-164X.15.4.287

64.

Werner

A. M.

Tibubos

A. N.

Mülder

L. M.

Reichel

J. L.

Schäfer

Heller

Pfirrmann

Edelmann

Dietz

Rigotti

Beutel

M. E

(2021). The impact of lockdown stress and loneliness during the COVID-19 pandemic on mental health among university students in Germany. Sci Rep, 11(1), 22637. https://doi.org/10.1038/s41598-021-02024-5

65.

White

Hingson

(2014). The burden of alcohol use: Excessive alcohol consumption and related consequences among college students. Alcohol Research: Current Reviews, 35(2), 201–218.

66.

Wood

C. I

Sealy

D.-A

Moss

Zigbuo-Wenzler

McFadden

Landi

Brace

A. M

(2024). Mental health impacts of the COVID-19 pandemic on college students. J Am Coll Health, 72(2), 463–468. https://doi.org/10.1080/07448481.2022.2040515

67.

Zimet

G. D.

Dahlem

N. W.

Zimet

S. G.

Farley

G. K.

(1988). The multidimensional scale of perceived social support. Journal of Personality Assessment, 52(1), 30–41. https://doi.org/10.1207/s15327752jpa5201_2

68.

Zysset

Volken

Amendola

von Wyl

Dratva

(2022). Change in Alcohol Consumption and Binge Drinking in University Students During the Early COVID-19 Pandemic. Front Public Health, 10, 854350. https://doi.org/10.3389/fpubh.2022.854350

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.27 MB

0.23 MB