Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 May 1.
Published in final edited form as: Addict Behav. 2024 Jan 11;152:107955. doi: 10.1016/j.addbeh.2024.107955

Evaluating Distress as a Moderator of the Relationship Between Drinking Identity and Hazardous Drinking During the Post-College Transition

Kristen P Lindgren a, Clayton Neighbors b, Bethany A Teachman c, Reinout W Wiers d
PMCID: PMC10959022  NIHMSID: NIHMS1963624  PMID: 38290321

Abstract

Drinking identity (the extent to which one associates the self with drinking alcohol) is a robust predictor of young adult hazardous drinking (HD; heavy drinking and alcohol-related problems), and decreases in drinking identity have been linked to the decline in HD that often occurs following college graduation. Identifying moderators is key to recognizing who is most at risk for continued HD given a drinking identity vulnerability. Using data from a longitudinal study of graduating college students from the U.S., we evaluated distress (depression, anxiety, stress symptoms) as a potential moderator. Between- and within-person components of drinking identity and distress were evaluated to consider both individual differences and variations within a person across time and changing contexts. Study hypotheses and data analysis plan were preregistered. Graduating college students who met HD criteria (N = 422) completed implicit and explicit drinking identity measures (assessed using reaction time and self-report measures, respectively), distress symptom questionnaires, and self-reported alcohol consumption and problems at four-month intervals for 2.5 years. Results supported moderation at the between-person level for alcohol consumption, with higher levels of implicit drinking identity and distress linked to greater subsequent alcohol consumption. Only between-person main effects for (explicit) identity and distress were linked to more subsequent alcohol-related problems. Though moderation findings were mixed, having a stronger drinking identity and/or greater distress was linked to continued HD risk in this sample. Individuals with these risk factors may benefit from enhanced prevention efforts to help graduates transition out of HD post-college.

Keywords: drinking identity, young adults, alcohol misuse, distress, moderator

1. Introduction

College student hazardous drinking (HD; heavy drinking and experiencing alcohol-related problems) is both risky and costly (Hingson et al., 2017; Merrill & Carey, 2016; Schulenberg et al., 2021). For many students, however, HD is temporary, and they transition or “mature out” of HD without treatment (Prince et al., 2019; Vergés et al., 2012). We have shown that drinking identity, or the extent to which one associates the self with drinking, is associated with declines in HD following college graduation (Lindgren et al., 2023). Research linking drinking identity to the transition out of HD is nascent, and identifying moderators of this relationship to understand who is especially at risk for continued HD is important. The current study evaluates clinically-relevant distress (stress, depression and anxiety symptoms) as a potential moderator; it is a secondary investigation that uses data from our longitudinal study of graduating college seniors who met HD criteria.

Drinking identity is a self-concept focused on the extent to which the self is linked to drinking. Our interest in drinking identity is informed by psychology literature dating back to the 1890s that focuses on the self as a organizing system for a person’s thoughts, attitudes, perceptions, memory, goals, motivations, and behavior (James, 1980; Markus, 1977; Strauss, 1959). Self-concepts are part of that self-system and are at the center of behavioral self-regulation, with theory suggesting that individuals make sense of situations and then behave congruently with their self-concepts (Frazier et al., 2021; Markus & Nurius, 1986; Oyserman et al., 2012). Individuals have multiple self-concepts, and self-concepts are malleable in terms of which ones are activated and how readily they are activated (Markus & Nurius, 1986; Nurra & Oyserman, 2018).

Theories of substance misuse and addiction, such as West’s (2006) PRIME theory, conceptualize identity as the source of motives, norms, and evaluation about substances, which in turn lead directly to behavior, such as drinking. Previous findings indicate that drinking identity increases during the early college years when drinking is typically initiated and escalates (Lindgren et al., 2016, 2018, 2020). In these studies, drinking identity was assessed by explicit measures, which ask directly about the extent individuals view drinking as part of who they are, and by implicit measures, which use computer-based reaction time (RT) tasks that indirectly assess how strongly drinking and the self are associated (Lindgren et al., 2013). The two measures correlate modestly, predict unique variance in HD (e.g., Lindgren et al., 2013; Lindgren, Neighbors, Teachman, et al., 2016), and in combination, provide a more comprehensive assessment of drinking identity. There is debate about the underlying constructs assessed by implicit and explicit measures (see Sherman & Klein, 2021); we retain the terms “implicit” and “explicit” simply to reflect the indirect and direct measurement of drinking identity.

Decreases in implicit and explicit drinking identity accompany the transition out of HD made by many graduating college seniors (Lindgren et al., 2023). These findings are consistent with theories emphasizing changes in self-concepts as determinants of behavior change and as important for recovery or reducing substance use (Dingle et al., 2015; Kearney & O’Sullivan, 2003; Vangeli & West, 2012). They also underscore the importance of identifying factors that moderate the relationship between drinking identity and HD during this time of transition.

Theories that relate self-concepts and drinking suggest that a given identity is most likely to influence behavior in contexts relevant to that identity (Oyserman et al., 2012; West & Brown, 2014). As an intrapersonal context that is highly relevant to HD, distress is a strong candidate when considering plausible moderators of drinking identity and HD. Further, alcohol is commonly used to reduce distress, including in college student populations (Cooper et al., 1995; Khantzian, 1997; Park & Levenson, 2002). We expect that a person who drinks to cope with distress (vs. drinking for social or celebratory reasons, for example) will be particularly likely to drink in problematic ways because they are feeling compelled to drink. As a result, chronic distress may create a context in which seeing oneself as a drinker is especially likely to predict HD. Notably, a longitudinal study of graduating Canadian college students provides some support for this idea (though they did not measure drinking identity). Frohlich et al. (2018) found that depression was associated with continued HD risk post-college. We aimed to extend this work by evaluating chronic distress (depression, anxiety, and stress symptoms) as moderators of drinking identity and HD.

This study uses data from our study of college students who met HD criteria during the last six months of college (Lindgren et al., 2023) and evaluated chronic distress (depression, anxiety, and stress symptoms) as moderators of the relationship between drinking identity and HD. The study had 8 timepoints, spanning 2.5 years. Implicit and explicit drinking identity, alcohol consumption and problems, and distress symptoms were assessed at every timepoint, providing an opportunity to test moderation prospectively and evaluate between- and within-person contributions of distress and identity. Greater distress was hypothesized to augment the positive relationship between drinking identity measures and indicators of HD. The same moderation pattern was expected at the between- and within-person levels for implicit and explicit drinking identity measures. The hypotheses and data analytic plan for the current study are registered at https://osf.io/7efnw.1

2. Method

Procedures and measures are summarized here. Complete information is available in Lindgren et al. (2023).

2.1. Participants

This study uses the complete sample from Lindgren et al. (2023). Study eligibility criteria included graduating college within the next six months, scoring ≥8 on the Alcohol Use Disorders Identification Test (AUDIT) to meet risk for HD criteria (Babor et al., 2001), being between the age of 18–25, and English fluency. Participants were 422 college seniors from a large public university in the U.S. Pacific Northwest (58.8% women, 41.2% men; no participants self-identified another gender identity or as transgender). The sample’s racial composition was 61.6% White/Caucasian, 20.9% Asian, 12.8% more than one race, 1.2% Black or African American, 1.2% American Indian/Alaska Native, 0.9% Native Hawaiian or Other Pacific Islander, 0.7% unknown, and 0.7% declined to answer. The majority (93%) identified as not Hispanic or Latino. Retention rates for follow-up assessments were 91%, 90%, 87%, 84%, 81%, 81%, 82% for Time 2-Time 8, respectively. Participants who remained in the study through Time 8 did not differ from those who did not on baseline scores of implicit or explicit drinking identity, DASS, drinks per week, or alcohol-related problems.

2.2. Measures

The drinking identity Implicit Association Test (IAT) evaluated implicit drinking identity (Lindgren et al., 2013). The IAT is a computer-based reaction time task that measures the strength of associations between concepts (i.e., drinker+me) relative to an alternative pairing (i.e., drinker+not me). In the IAT, a series of word stimuli are presented center-screen. Participants are instructed to press the corresponding key (e and i) to classify each stimulus into one of two pairs of categories as quickly and accurately as possible. They must correct errors before advancing. Scores were calculated using the D1 scoring algorithm; scores were screened for exclusion if >10% of trials were faster than 300ms (Greenwald et al., 2003). The drinking identity IAT was implemented as described in Lindgren et al. (2023). Higher scores indicate stronger associations between drinker and me (versus not me). The internal consistency at each timepoint was between r = .52 to .60. Though lower than most self-report measures (Nosek et al., 2007), they were consistent with those reported with other college samples (Lindgren et al., 2013, 2016).2

The Alcohol Self-concept Scale (Corte & Stein, 2007; Lindgren et al., 2013) evaluated explicit drinking identity. Participants rated their agreement with five statements about drinking being part of their identity (e.g., “Drinking is a part of my self-image”) using a 7-point Likert scale (range–3 [“strongly disagree”] to +3 [“strongly agree”]). A mean score was calculated; higher scores indicate stronger drinking identity. Cronbach’s alphas at each timepoint were between.92 and .94.3

Mood-related symptoms were assessed via the Depression Anxiety Stress Scale-21 (DASS-21) (Lovibond & Lovibond, 1996), which evaluates 21 symptoms of depression, anxiety, and stress over the last month and has been used with both non-clinical and clinical samples (Antony et al., 1998). Participants evaluated the extent to which each symptom applied to them using a scale ranging from 0 (Not at all) to 3 (Most of the time). While DASS subscales have often been used to separately examine depression, anxiety, and stress, a single composite score has demonstrated high reliability and validity (Zanon et al., 2021) and is frequently used as a measure of general distress (e.g., Chauhan et al., 2023; Collado-Navarro et al., 2021; Gloster et al., 2008; Moss-Pech et al., 2021). Following typical scoring procedures for the DASS-21, we first calculated scores for each of the three subscales by taking the sum of the seven items and doubling it (Gloster et al., 2008; Henry & Crawford, 2005). The DASS total score consisted of the mean of the three subscale scores. Cronbach’s alphas at each timepoint were between .81 and .87.

The Daily Drinking Questionnaire (DDQ; Collins et al., 1985) measured typical weekly alcohol consumption over the past three months. Participants were asked to report the number of alcoholic drinks they consume each day of a typical week. U.S. standard drink equivalencies were provided. A sum indicating total drinks per week was created.

The Rutgers Alcohol Problem Index (RAPI; White & Labouvie, 1989) evaluated how many times participants have experienced 23 negative consequences (e.g., “passed out or fainted suddenly”) from drinking or because of their alcohol use during the past four months using options ranging from never (0) to more than 10 times (4). The first item (“not able to do your homework or study for a test”) was modified after the second assessment (i.e., after participants graduated) to “not able to complete your job responsibilities or do your homework or study for a test.” Two items assessing driving under the influence were added. A sum of the items was created (range: 0–100); higher scores indicated more problems. Cronbach’s alphas at each timepoint were between .87 and .92.

2.3. Procedures

Study procedures were approved by the university’s institutional review board. Potential participants were recruited from a list of 18- to 25-year-old full-time seniors provided by the university registrar’s office. Invitations were sent via email. The email linked to the study webpage where volunteers completed informed consent and eligibility screening. Eligible individuals were invited to continue to the baseline assessment (T1) to complete the remaining measures. Follow-up assessments (T2–T8) occurred at four-month intervals. Assessments were web-based and hosted by Project Implicit. Accuracy check questions (e.g., “To answer this question correctly, you must answer ‘Strongly disagree’”) were included in each assessment to measure participant attentiveness. The majority (84%) of participants answered all T1 accuracy check questions correctly; 91% or more answered them all correctly at T2-T8. Participants were not excluded for missing accuracy check questions but their data were reviewed for potential exclusion and none were. Participants received compensation for completing assessments ($25 for T1-T4, $30 for T5-T8). Bonus compensation included (1) being entered into a drawing to win one of four $25 electronic gift cards after completing each assessment; (2) receiving a $25 bonus if they completed all of the first four assessments; and (3) receiving a $25 bonus if they completed all of the last four assessments.

2.4. Data Analysis Plan

The data analysis plan was preregistered at https://osf.io/7efnw. The changes we made to the analysis plan were in our use of the DASS-21 composite score (see Footnote 1) and approach to evaluating interactions of nonlinear models. Rather than comparing the average partial derivative representing the association between identity and drinking outcome at multiple values of distress (e.g., depression), we conducted tests of the second-order cross-partial derivative, which is more direct and parsimonious (Kim & McCabe, 2022; McCabe et al., 2021). Data were analyzed using generalized linear mixed models. We conducted one model for each outcome (alcohol consumption and problems); thus, two models total. Models were prospective, with outcomes from T2-T8 predicted from identity variables and DASS scores at T1-T7, controlling for T1 outcomes. Thus, analyses used lagged within-person identity and DASS scores. Models, thus, evaluate how past month mood and present implicit and explicit drinking identity predict changes in drinks per week over the next three months and alcohol problems over the next four months.

Drinking identity measures and DASS scores were disaggregated to independently evaluate between-person (level 2) and within-person effects (level 1). Between-person identity measure scores and DASS scores consisted of person-means of T1-T7 for each variable. The within-person scores consisted of person-mean-centered values for each timepoint from T1 to T7. Models included random intercepts and linear slopes for time.

Models were conducted hierarchically with baseline consumption (or problems), gender identity, between-person implicit identity (B-Implicit), between-person explicit identity (B-Explicit), between-person distress (B-DASS), within-person implicit identity (W-Implicit), within-person explicit identity (W-Explicit), and within-person distress (W-DASS) entered in step 1. Two-way product terms were entered in step 2 and included: (B-Implicit*B-DASS), (B-Explicit*B-DASS), (W-Implicit*W-DASS), (W-Explicit*W-DASS), (B-Implicit*W-DASS), (B-Explicit*W-DASS), (W-Implicit*B-DASS), and (W-Explicit*B-DASS).

Interactions were evaluated using approaches described by McCabe and colleagues (Kim & McCabe, 2022; McCabe et al., 2021). In log-linear models, the evaluation of whether the association between one predictor and the outcome varies as a function of another predictor (two-way interaction) cannot be determined on the basis of a single product term. In log-linear models, an interaction can be correctly evaluated by testing the second-order cross-partial derivative of the model equation with respect to the constituent variables of the potential interaction. The notation ∂f2∂w∂x indicates taking the partial derivative of f with respect to x and then taking its partial derivative with respect to w w∂∂w∂f∂x=∂f2∂w∂x. In the present research, interactions were evaluated by tests of cross-partial derivatives representing the change in associations between identity and drinking outcomes as a function of distress.

3. Results

Table 1 presents means, standard deviations, correlations, and intraclass correlations (ICCs) for study variables (gender, implicit identity, explicit identity, DASS score, drinks per week, and RAPI). ICCs for implicit and explicit identity and DASS score represent the proportion of variance accounted for by between-person differences relative to the total variance in mixed intercept-only models. ICCs for drinks per week and RAPI were based on intercept-only multilevel negative binomial models using procedures described by Leckie et al. (2020).

Table 1.

Means, Standard Deviations, Correlations and Intraclass Correlation Coefficients.

1 2 3 4 5 6
1. Gender -- −0.19*** −0.12* 0.09 −0.25*** −0.05
2. Implicit Drinking Identity -- 0.39*** −0.07 0.30*** 0.10*
3. Explicit Drinking Identity 0.09*** -- 0.14** 0.45*** 0.37***
4. DASS-Total Score 0.02 0.07*** -- 0.16** 0.51***
5. Drinks Per Week 0.09*** 0.20*** 0.07*** -- 0.60***
6. Alcohol Problems 0.07*** 0.21*** 0.25*** 0.42*** --
 Mean 0.59 0.21 −1.52 3.57 2.68 4.84
 SD 0.49 0.43 1.29 4.24 2.90 3.92
 Within-person SD 0.00 0.30 0.73 2.48 1.81 2.30
 % Variance between-person (ICC) 1.00 0.42 0.63 0.60 0.56 0.59
 N (between-person) 422 420 422 422 422 422
 N (within-person) -- 2854 2925 2925 2925 2927
Open in a new tab

Note. Between-person correlations are above the diagonal. Within-person correlations are below the diagonal. Gender was coded 0=men; 1=women. SD = standard deviation. ICC = intraclass correlation coefficient.

***

p < .001.

**

p < .01.

*

p < .05.

Table 2 presents tests of parameter estimates from a negative binomial mixed model examining changes in drinks per week as a function of implicit and explicit drinking identity (between- and within-person) and DASS score. Step 1 effects revealed that baseline drinks per week was positively associated with drinks per week at subsequent timepoints. Drinks per week decreased over time. Higher levels of implicit and explicit drinking identity were uniquely associated with increased drinks per week at the between-person level. At the within-person level, neither type of drinking identity nor DASS score was related to changes in drinks per week.

Table 2.

Drinks per Week as a Function of Implicit and Explicit Drinking Identity and the DASS

Step 1 b se ∂ ∂se Z p
Intercept 2.420 0.044
Time −0.117 0.009 −0.956 0.076 −12.54 <0.001
Week0 0.037 0.003 0.411 0.046 8.94 <0.001
Gender −0.002 0.055 −0.022 0.606 −0.04 0.971
Avg Implicit ID (BW IID) 0.306 0.091 3.395 1.020 3.33 0.001
Avg Explicit ID (BW EID) 0.136 0.028 1.502 0.318 4.72 <0.001
Avg DASS (BW DASS) 0.002 0.005 0.019 0.053 0.36 0.717
Implicit ID-Avg (WI IID) 0.018 0.033 0.202 0.369 0.55 0.584
Explicit ID-Avg (WI EID) 0.009 0.016 0.098 0.181 0.54 0.589
DASS-Avg (WI DASS) 0.002 0.003 0.025 0.035 0.71 0.479
Step 1 Random Intercept 0.012 0.002
Step 1 Random Time Slope 0.177 0.021
Step 2 b se ∂2 ∂2 se Z p
BW IID*BW DASS 0.045 0.018 0.518 0.208 2.49 0.013
BW EID*BW DASS 0.000 0.002 0.014 0.027 0.52 0.600
WI IID*WI DASS 0.014 0.011 0.155 0.118 1.31 0.189
WI EID*WI DASS 0.002 0.004 0.026 0.047 0.56 0.578
BW IID*WI DASS −0.001 0.010 −0.002 0.118 −0.01 0.989
BW EID*WI DASS 0.001 0.003 0.013 0.036 0.36 0.716
WI IID*BW DASS −0.001 0.006 −0.008 0.068 −0.11 0.910
WI EID*BW DASS 0.004 0.003 0.049 0.038 1.30 0.195
Step 2 Random Intercept 0.012 0.002
Step 2 Random Time Slope 0.172 0.021
Open in a new tab

Note. Week0 = Drinks per week at baseline. Gender was coded 0 = male, 1 = female). BW = between-person level, W.I. = within-person level. Avg = average. Implicit ID (IID) = implicit drinking identity, Explicit ID (EID) = explicit drinking identity; higher scores = stronger drinking identity. DASS = total score on the DASS-21; higher scores = greater general distress. Drinks per week = self-reported number of drinks consumed in a typical week assessed via the Daily Drinking Questionnaire (DDQ). ∂ = partial derivative also known as average marginal effect. ∂2 = 2nd order cross-partial derivative representing the interaction between the two relevant variables.

Step 2 results can be interpreted as tests of interactions in predicting changes in the log number of drinks per week. Tests of cross-partial derivatives were conducted to evaluate interactions in predicting changes in the number of drinks per week in natural units. The cross-partial derivatives was only significant in one instance: between-person implicit drinking identity and between-person DASS score. This interaction is presented in Figure 1, which shows marginal effects of between-person implicit drinking identity (BPIDI) on drinks per week at varying levels of between-person DASS scores. The average slope of each curve is the partial derivative of the model equation with respect to BPIDI at representative values of DASS scores (0–24). The range of values in the graph cover roughly 97% of the joint distribution, with all other variables fixed at their observed values. As shown in Figure 1, at increasing values of distress, the association between BPIDI and drinking increases became progressively stronger. The slopes represented by ∂BPIDI in the legend all differed significantly from each other.

Figure 1.

Figure 1

Open in a new tab

Interaction of Between-Person Implicit Drinking Identity and Between-Person DASS Predicting Drinks per Week

Tests of parameter estimates from negative binomial mixed models examining changes in alcohol-related problems (as assessed by the RAPI) as a function of implicit and explicit drinking identity (between- and within-person) and the DASS are presented in Table 3. Step 1 results indicated that baseline RAPI scores were associated with higher RAPI scores at subsequent timepoints. RAPI scores decreased over time. Higher levels of explicit, but not implicit, drinking identity were associated with increased problems at the between-person level. At the within-person level, neither explicit nor implicit drinking identity was related to changes in alcohol-related problems. The DASS was associated with increased problems at the between-person but not at the within-person level. Results in Step 2 revealed no significant product terms. Moreover, tests of cross-partial derivatives indicated no significant interactions between identity variables and DASS at the between- or within-person levels.

Table 3.

Alcohol Problems as a Function of Identity and DASS Subscales

Step 1 b se ∂ ∂se Z p
Intercept 1.678 0.061
Time −0.144 0.013 −0.560 0.089 −6.31 0.000
Alcohol Problems0 0.055 0.006 0.389 0.056 6.89 0.000
Gender 0.029 0.076 0.203 0.537 0.38 0.706
Avg Implicit ID (BW IID) 0.212 0.137 1.494 0.969 1.54 0.123
Avg Explicit ID (BW EID) 0.197 0.039 1.389 0.302 4.60 0.000
Avg DASS (BW DA) 0.038 0.009 0.270 0.066 4.09 0.000
Implicit ID-Avg (WI IID) −0.038 0.054 −0.269 0.380 −0.71 0.479
Explicit ID-Avg (WI EID) −0.020 0.024 −0.142 0.168 −0.84 0.401
DASS-Avg (WI DA) 0.008 0.004 0.059 0.030 1.94 0.052
Step 1 Random Intercept 0.026 0.004
Step 1 Random Time Slope 0.348 0.041
Step 2 b se ∂2 ∂2 se Z p
BW IID*BW DD 0.014 0.028 0.168 0.220 0.76 0.445
BW EID*BW DD −0.004 0.007 0.023 0.059 0.38 0.701
WI IID*WI DD 0.010 0.014 0.069 0.095 0.73 0.467
WI EID*WI DD −0.005 0.006 −0.036 0.040 −0.88 0.377
BW IID*WI DD 0.005 0.015 0.048 0.102 0.47 0.637
BW EID*WI DD 0.000 0.004 0.010 0.028 0.36 0.716
WI IID*BW DD −0.001 0.009 −0.015 0.066 −0.23 0.816
WI EID*BW DD 0.006 0.004 0.041 0.026 1.57 0.117
Step 2 Random Intercept 0.025 0.004
Step 2 Random Time Slope 0.349 0.041
Open in a new tab

Note. Alcohol Problems0 = Alcohol problems at baseline. Gender was coded 0 = male, 1 = female). BW = between-person level, W.I. = within-person level. Avg = average. Implicit ID (IID) = implicit drinking identity, Explicit ID (EID) = explicit drinking identity; higher scores = stronger drinking identity. DASS = total score on the DASS-21; higher scores = greater general distress. Drinks per week = self-reported number of drinks consumed in a typical week assessed via the Daily Drinking Questionnaire (DDQ). ∂ = partial derivative also known as average marginal effect. ∂2 = 2nd order cross-partial derivative representing the interaction between the two relevant variables.

4. Discussion

The post-college transition is a period where many who meet HD criteria reduce their drinking on their own. The current investigation is a secondary analysis that tested whether chronic distress (depression, anxiety, and stress symptoms) interacted with drinking identity and were associated with continued HD in a 2.5-year study of graduating college students. Consistent with the larger study’s findings (Lindgren et al., 2023), consumption and problems declined over time, but the expected moderation effect was observed only at the between-person levels for implicit identity and distress when predicting alcohol consumption. This suggests that though, overall, the sample reduced HD during the post-college transition, individuals who had both stronger (vs. weaker) implicit drinking identity and experienced more (vs. less) distress were more likely to continue drinking heavily. Should findings replicate, this may indicate a need for targeted interventions for soon-to-graduate or recently-graduated college students who are higher in drinking identity and higher in distress symptoms.

Moderation was not observed for alcohol problems. That moderation was observed for alcohol consumption and not for problems may reflect that there is a more proximal link between drinking identity and alcohol consumption than between drinking identity and alcohol-related problems. Also, other studies found stronger associations between distress-related factors (e.g., stress, depression, and emotion regulation difficulties) and alcohol-related problems versus with alcohol consumption (Acuff et al., 2018; Dvorak et al., 2014; Lee et al., 2019), and consistent with such findings, the current study found a significant positive association between between-person distress and subsequent alcohol-related problems. Between-person explicit drinking identity was also associated with more problems. Thus, while the combination of higher distress and (explicit) drinking identity does not increase risk, individuals who have either or both appear to be at higher risk for continued alcohol-related problems post-college.

The present results also support the assessment of drinking identity via multiple measures. Consistent with other studies (Lindgren et al., 2013, 2016), implicit and explicit measures were moderately correlated and predicted unique variance in future alcohol consumption over time. Of interest, explicit (but not implicit) identity uniquely predicted future problems in college graduates who had initially met HD criteria. That the implicit and explicit identity yielded different findings could be due to differences in measurement, including levels of awareness and intentional endorsement and expression.

The null moderation effects for the within-person components of identity and distress may indicate that fluctuations from one’s typical levels of identity or distress are not associated with differential prediction of shifts in drinking. Additionally, the four-month interval between assessments may be too long to detect the effects of fluctuations in distress or drinking identity on drinking outcomes. That interval was chosen for the purposes of the larger study, which also focused on shifts in one’s day-to-day social network. Four months was viewed as long enough to allow for network changes due to graduation, starting a new school or job, moving to a location, etc., to occur, but not too long that it might miss meaningful changes. To detect within-person changes in distress and identity, a shorter time scale might be more useful. For example, one could imagine an ecological momentary approach in which current affect/distress, current identification with drinking, and same and next-day drinking consequences were evaluated for a period of several days or weeks. We note, as well, that future studies might also explore associations between behavior and identity (as moderated by distress or negative affect).

While this study had many strengths, it also had limitations, including the lack of assessment of drinking motives, which limits our understanding of the extent to which participants were engaging in drinking to cope with any distress they reported experiencing. Further, the sample is largely white/Caucasian, highly-educated, and recruited from a single university in the US Pacific Northwest which limits generalizability. Different timeframes were used for the measures (e.g., consumption focused on last 3 months, problems on the last 4 months, and distress on the last 1 month. It is possible that using the same time frame for all measures would yield different results. Despite these limitations, the study was rigorous in using validated assessments, repeated measurement, and strong retention over time. Should findings replicate, graduating college students with stronger drinking identities and greater distress may benefit from targeted interventions to facilitate the transition out of hazardous drinking.

Highlights:

  • Implicit and explicit drinking identity are hazardous drinking (HD) risk factors

  • They were examined as risk factors for continued HD in the post-college transition

  • Chronic distress was assessed as a moderator

  • Moderation was supported in some instances

  • Chronic distress and implicit identity may increase risk for continued HD

Acknowledgments

Funding was provided by the National Institute on Alcohol Abuse and Alcoholism (NIAAA), Grant Number: R01 AA024732 (PI: Lindgren). Manuscript support was provided by NIAAA, Grant Number: R01AA029672 (MPIs: Lindgren & Neighbors). NIAAA played no role in the study design; the collection, analysis, and interpretation of data; the writing of this paper; or the decision to submit this paper for publication. Bethany Teachman has a significant financial interest in Project Implicit, Inc., which provided data collection services supporting this project under contract with the University of Washington.

Footnotes

Conflicts of Interest:

The authors declare they have no conflicts of interest.

Declarations of Interest:

Bethany Teachman has a significant financial interest in Project Implicit, Inc., which provided data collection services supporting this project under contract with the University of Washington. No other authors have any interests to declare.

Author CRediT statement

Kristen Lindgren: Conceptualization, Funding acquisition, Investigation, Methodology, Writing - original draft, Writing - review & editing. Clayton Neighbors: Conceptualization, Investigation, Formal analysis, Visualization, Roles/Writing - original draft, Writing - review & editing. Bethany Teachman: Conceptualization, Writing - Writing - review & editing. Reinout Wiers: Conceptualization, Writing - Writing - review & editing.

1

We originally conducted separate analyses for each type of distress symptoms. We appreciate our anonymous reviewers’ suggestions to reconsider this approach to be more parsimonious because we had the same predictions for each type. The original results are available at https://osf.io/c287q/.

2

We note the ongoing debate about many aspects of the IAT’s interpretation and validity (Kurdi et al., 2019; Schimmack, 2021).

3

Given the limited availability of psychometric information for this measure, we evaluated measurement invariance of the five items over time for the eight timepoints. Comparison of fit for a model in which all parameters were free to vary at each time point, χ2 df(40) = 231.39, and a model in which loadings and intercepts were invariant over time, χ2 df(66) = 292.90, demonstrated pattern and scalar invariance, χ2Δ df(56) = 61.51, p = .28.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Acuff SF, Soltis KE, Luciano MT, Meshesha LZ, Pedrelli P, Dennhardt AA, & Murphy JG (2018). Depressive symptoms as predictors of alcohol problem domains and reinforcement among heavy drinking college students. Psychology of Addictive Behaviors, 32(7), 792–799. 10.1037/adb0000397 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Antony MM, Bieling PJ, Cox BJ, Enns MW, & Swinson RP (1998). Psychometric properties of the 42-item and 21-item versions of the Depression Anxiety Stress Scales in clinical groups and a community sample. Psychological Assessment, 10(2), 176–181. 10.1037/1040-3590.10.2.176 [DOI] [Google Scholar]
  3. Babor TF, Higgins-Biddle JC, Saunders JB, & Monteiro MG (2001). The Alcohol-use disorders Identification Test (AUDIT): Guidelines for use in primary care (2nd ed.). World Health Organization, Department of Mental Health and Substance Dependence. [Google Scholar]
  4. Chauhan K, Donahue J, & Thompson R (2023). The predictive validity of the DSM-5 alternative model for borderline personality disorder: Associations with coping strategies, general distress, rumination, and suicidal ideation across one year. Personality and Mental Health, 17(3), 272–281. 10.1002/pmh.1580 [DOI] [PubMed] [Google Scholar]
  5. Collado-Navarro C, Navarro-Gil M, Pérez-Aranda A, López-del-Hoyo Y, Garcia-Campayo J, & Montero-Marin J (2021). Effectiveness of mindfulness-based stress reduction and attachment-based compassion therapy for the treatment of depressive, anxious, and adjustment disorders in mental health settings: A randomized controlled trial. Depression and Anxiety, 38(11), 1138–1151. 10.1002/da.23198 [DOI] [PubMed] [Google Scholar]
  6. Collins RL, Parks GA, & Marlatt GA (1985). Social determinants of alcohol consumption: The effects of social interaction and model status on the self-administration of alcohol. Journal of Consulting and Clinical Psychology, 53(2), 189–200. 10.1037/0022-006X.53.2.189 [DOI] [PubMed] [Google Scholar]
  7. Cooper ML, Frone MR, Russell M, & Mudar P (1995). Drinking to regulate positive and negative emotions: A motivational model of alcohol use. Journal of Personality and Social Psychology, 69(5), 990–1005. [DOI] [PubMed] [Google Scholar]
  8. Corte C, & Stein KF (2007). Self-Cognitions in Antisocial Alcohol Dependence and Recovery. Western Journal of Nursing Research, 29(1), 80–99. 10.1177/0193945906295480 [DOI] [PubMed] [Google Scholar]
  9. Dingle GA, Cruwys T, & Frings D (2015). Social Identities as Pathways into and out of Addiction. Frontiers in Psychology, 6. 10.3389/fpsyg.2015.01795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dvorak RD, Sargent EM, Kilwein TM, Stevenson BL, Kuvaas NJ, & Williams TJ (2014). Alcohol use and alcohol-related consequences: Associations with emotion regulation difficulties. The American Journal of Drug and Alcohol Abuse, 40(2), 125–130. 10.3109/00952990.2013.877920 [DOI] [PubMed] [Google Scholar]
  11. Frazier LD, Schwartz BL, & Metcalfe J (2021). The MAPS model of self-regulation: Integrating metacognition, agency, and possible selves. Metacognition and Learning, 1–22. 10.1007/s11409-020-09255-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Frohlich JR, Rapinda KK, O’Connor RM, & Keough MT (2018). Examining co-patterns of depression and alcohol misuse in emerging adults following university graduation. Addictive Behaviors Reports, 8, 40–45. 10.1016/j.abrep.2018.06.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gloster AT, Rhoades HM, Novy D, Klotsche J, Senior A, Kunik M, Wilson N, & Stanley MA (2008). Psychometric properties of the Depression Anxiety and Stress Scale-21 in older primary care patients. Journal of Affective Disorders, 110(3), 248–259. 10.1016/j.jad.2008.01.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Greenwald AG, Nosek BA, & Banaji MR (2003). Understanding and using the Implicit Association Test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85(2), 197–216. 10.1037/0022-3514.85.2.197 [DOI] [PubMed] [Google Scholar]
  15. Henry JD, & Crawford JR (2005). The short-form version of the Depression Anxiety Stress Scales (DASS-21): Construct validity and normative data in a large non-clinical sample. British Journal of Clinical Psychology, 44(2), 227–239. 10.1348/014466505X29657 [DOI] [PubMed] [Google Scholar]
  16. Hingson RW, Zha W, & White AM (2017). Drinking Beyond the Binge Threshold: Predictors, Consequences, and Changes in the U.S. American Journal of Preventive Medicine, 52(6), 717–727. 10.1016/j.amepre.2017.02.014 [DOI] [PubMed] [Google Scholar]
  17. James W (1980). The Principles of Psychology: Vol. 1. H. Holt & Company. [Google Scholar]
  18. Kearney MH, & O’Sullivan J (2003). Identity Shifts as Turning Points in Health Behavior Change. Western Journal of Nursing Research, 25(2), 134–152. 10.1177/0193945902250032 [DOI] [PubMed] [Google Scholar]
  19. Khantzian EJ (1997). The Self-Medication Hypothesis of Substance Use Disorders: A Reconsideration and Recent Applications. Harvard Review of Psychiatry, 4(5), 231–244. 10.3109/10673229709030550 [DOI] [PubMed] [Google Scholar]
  20. Kim DS, & McCabe CJ (2022). The partial derivative framework for substantive regression effects. Psychological Methods, 27(1), 121–141. 10.1037/met0000440 [DOI] [PubMed] [Google Scholar]
  21. Leckie G, Browne WJ, Goldstein H, Merlo J, & Austin PC (2020). Partitioning variation in multilevel models for count data. Psychological Methods, 25(6), 787–801. 10.1037/met0000265 [DOI] [PubMed] [Google Scholar]
  22. Lee BL, Jordan HR, & Madson MB (2019). The Moderating Effects of College Stress on the Relationship between Protective Behavioral Strategies and Alcohol Outcomes. Substance Use & Misuse, 54(11), 1845–1852. 10.1080/10826084.2019.1618330 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lindgren KP, Baldwin SA, Olin CC, Wiers RW, Teachman BA, Norris J, Kaysen D, & Neighbors C (2018). Evaluating Within-Person Change in Implicit Measures of Alcohol Associations: Increases in Alcohol Associations Predict Increases in Drinking Risk and Vice Versa. Alcohol and Alcoholism, 53(4), 386–393. 10.1093/alcalc/agy012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lindgren KP, Baldwin SA, Peterson KP, Ramirez JJ, Teachman BA, Kross E, Wiers RW, & Neighbors C (2023). Maturing Out: Between- and Within-Persons Changes in Social-Network Drinking, Drinking Identity, and Hazardous Drinking Following College Graduation. Clinical Psychological Science, 11(1), 23–39. 10.1177/21677026221082957 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lindgren KP, Baldwin SA, Peterson KP, Wiers RW, & Teachman BA (2020). Change in implicit alcohol associations over time: Moderation by drinking history and gender. Addictive Behaviors, 107, 106413. 10.1016/j.addbeh.2020.106413 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lindgren KP, Neighbors C, Teachman BA, Baldwin SA, Norris J, Kaysen D, Gasser ML, & Wiers RW (2016). Implicit alcohol associations, especially drinking identity, predict drinking over time. Health Psychology, 35(8), 908–918. 10.1037/hea0000396 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lindgren KP, Neighbors C, Teachman BA, Wiers RW, Westgate E, & Greenwald AG (2013). I drink therefore I am: Validating alcohol-related implicit association tests. Psychology of Addictive Behaviors, 27(1), 1–13. 10.1037/a0027640 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lovibond SH, & Lovibond PF (1996). Manual for the depression anxiety stress scales. Psychology Foundation of Australia. [Google Scholar]
  29. Markus H (1977). Self-schemata and processing information about the self. Journal of Personality and Social Psychology, 35(2), 63–78. 10.1037/0022-3514.35.2.63 [DOI] [Google Scholar]
  30. Markus H, & Nurius P (1986). Possible selves. American Psychologist, 41(9), 954–969. 10.1037/0003-066X.41.9.954 [DOI] [Google Scholar]
  31. McCabe CJ, Halvorson MA, King KM, Cao X, & Kim DS (2021). Interpreting Interaction Effects in Generalized Linear Models of Nonlinear Probabilities and Counts. Multivariate Behavioral Research, 1–27. 10.1080/00273171.2020.1868966 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Merrill JE, & Carey KB (2016). Drinking Over the Lifespan. Alcohol Research : Current Reviews, 38(1), 103–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Moss-Pech SA, Southward MW, & Cheavens JS (2021). Hope attenuates the negative impact of general psychological distress on goal progress. Journal of Clinical Psychology, 77(6), 1412–1427. 10.1002/jclp.23087 [DOI] [PubMed] [Google Scholar]
  34. Nosek BA, Greenwald AG, & Banaji MR (2007). The Implicit Association Test at Age 7: A Methodological and Conceptual Review. In Social psychology and the unconscious: The automaticity of higher mental processes. (pp. 265–292). Psychology Press. [Google Scholar]
  35. Nurra C, & Oyserman D (2018). From future self to current action: An identity-based motivation perspective. Self and Identity, 17(3), 343–364. 10.1080/15298868.2017.1375003 [DOI] [Google Scholar]
  36. Oyserman D, Elmore K, & Smith GC (2012). Self, self-concept, and identity. In Leary MR & Tangney JP (Eds.), Handbook of self and identity (2nd ed, pp. 69–104). Guilford Press. [Google Scholar]
  37. Park CL, & Levenson MR (2002). Drinking to cope among college students: Prevalence, problems and coping processes. Journal of Studies on Alcohol, 63(4), 486–497. 10.15288/jsa.2002.63.486 [DOI] [PubMed] [Google Scholar]
  38. Prince MA, Read JP, & Colder CR (2019). Trajectories of College Alcohol Involvement and Their Associations with Later Alcohol Use Disorder Symptoms. Prevention Science: The Official Journal of the Society for Prevention Research, 20(5), 741–752. 10.1007/s11121-018-0974-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Schulenberg JE, Patrick ME, Johnston LD, O’Malley PM, Bachman J, G., & Miech RA (2021). Monitoring the Future national survey results on drug use, 1975–2020: Volume II, College students and adults ages 19–60. Institute for Social Research, The University of Michigan. [Google Scholar]
  40. Sherman JW, & Klein SAW (2021). The Four Deadly Sins of Implicit Attitude Research. Frontiers in Psychology, 11, 604340. 10.3389/fpsyg.2020.604340 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Strauss AL (1959). Mirrors & masks: The search for identity. Free Press. [Google Scholar]
  42. Vangeli E, & West R (2012). Transition towards a ‘non-smoker’ identity following smoking cessation: An interpretative phenomenological analysis: Transition towards a ‘non-smoker’ identity. British Journal of Health Psychology, 17(1), 171–184. 10.1111/j.2044-8287.2011.02031.x [DOI] [PubMed] [Google Scholar]
  43. Vergés A, Jackson KM, Bucholz KK, Grant JD, Trull TJ, Wood PK, & Sher KJ (2012). Deconstructing the Age-Prevalence Curve of Alcohol Dependence: Why “Maturing Out” is only a Small Piece of the Puzzle. Journal of Abnormal Psychology, 121(2), 511–523. 10.1037/a0026027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. West R (2006). Theory of addiction. Blackwell Pub./Addiction Press. [Google Scholar]
  45. West R, & Brown J (2014). Theory of addiction (Second edition). Wiley Blackwell/Addiction Press. [Google Scholar]
  46. White HR, & Labouvie EW (1989). Towards the assessment of adolescent problem drinking. Journal of Studies on Alcohol, 50(1), 30–37. 10.15288/jsa.1989.50.30 [DOI] [PubMed] [Google Scholar]
  47. Zanon C, Brenner RE, Baptista MN, Vogel DL, Rubin M, Al-Darmaki FR, Gonçalves M, Heath PJ, Liao H-Y, Mackenzie CS, Topkaya N, Wade NG, & Zlati A (2021). Examining the Dimensionality, Reliability, and Invariance of the Depression, Anxiety, and Stress Scale–21 (DASS-21) Across Eight Countries. Assessment, 28(6), 1531–1544. 10.1177/1073191119887449 [DOI] [PubMed] [Google Scholar]

RESOURCES