Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 May 7.
Published in final edited form as: Alcohol Clin Exp Res (Hoboken). 2023 Oct 4;47(11):2184–2196. doi: 10.1111/acer.15195

Examining the validity of the drinking motives questionnaire for identifying reward and relief drinking subgroups in college students

H M Hebden 1,2, V R Votaw 1,2, L Herchenroeder 3, K Witkiewitz 1,2, E W Yeung 3
PMCID: PMC11075668  NIHMSID: NIHMS1987711  PMID: 38226750

Abstract

Background:

Heavy alcohol use in college is associated with a risk of developing alcohol use disorder. Characterizing variability in individual risk factors for alcohol use could help mitigate risk by informing personalized approaches to prevention. This study examined the validity of a brief measure for identifying reward/relief drinking phenotypes in non-treatment-seeking young adults.

Methods:

College students (n = 454) who reported binge drinking completed the Drinking Motives Questionnaire-Revised-Short Form (DMQ-R-SF). Confirmatory factor and latent profile analyses (CFA; LPA) of the DMQ-R-SF were performed to assess structural validity and identify reward/relief drinking subgroups. We compared models measuring reward drinking with the DMQ-R-SF enhancement motives (reward-enhancement) subscale to models measuring reward drinking with enhancement and social motives (reward-enhancement/social). Across models, relief drinking was measured with coping motives. We examined associations between reward/relief drinking subgroups and alcohol and personality variables concurrently and prospectively at a 6-week follow-up.

Results:

A two-factor reward and relief structure of the DMQ-R-SF was supported. Three latent profiles were identified (low reward/low relief: n = 133, high reward/low relief: n = 249; high reward/high relief: n = 72). Both CFA and LPA models that utilized reward-enhancement/social items indicated a better fit than reward-enhancement items alone. At baseline, individuals in the high-reward/high-relief profile demonstrated the poorest alcohol use outcomes and higher negative affect. Those in the high-reward/low-relief profile demonstrated greater alcohol use severity than those in the low-reward/low-relief profile. Prospectively, individuals classified in the high-reward/low-relief subgroup reported greater binge drinking frequency and those in the high-reward/high-relief profile reported greater alcohol consequences.

Conclusions:

The DMQ-R-SF is a valid measure for identifying reward and relief drinking subgroups in college students with binge drinking and could have utility for precision prevention efforts that target individuals in the high-reward/low-relief and high-reward/high-relief subgroups.

Keywords: alcohol, precision medicine, prevention, relief drinking, reward drinking

INTRODUCTION

Heavy alcohol use in young adults has short- and long-term consequences, including assaults, injuries, and academic problems, and is associated with alcohol use disorder (AUD) symptoms up to 10 years later (Carter et al., 2010; Jennison, 2004; Prince et al., 2019; Slutske, 2005). Targeted prevention efforts that address drinking motives, alcohol expectancies, and perceived norms have been shown to mitigate short-term consequences and prevent the onset of more severe problems (Cronce & Larimer, 2011; Reid & Carey, 2015). Characterizing variability in these factors in young adults can help inform precision medicine approaches by matching individuals to interventions that might work best based on their individual characteristics (Boness & Witkiewitz, 2022; Litten et al., 2015; Martens et al., 2008; Roos et al., 2021).

Reward and relief drinking phenotypes, hypothesized in the three-pathway psychobiological model of craving (Verheul et al., 1999), characterize individual differences in desire to drink elicited by different motivations for alcohol use. Reward drinking is classified as alcohol use driven by stimulation, enhancement, and social reward motives (Verheul et al., 1999). These processes are hypothesized to reflect the positive reinforcement aspect of the addiction cycle (also known as the binge-intoxication stage; Koob & Volkow, 2010, 2016), which involves repeated sensitization to reward cues via dysregulation in dopaminergic and opioidergic neural circuitry (Koob & Volkow, 2010, 2016; Robinson & Berridge, 1993, 2001). Relief drinking, on the other hand, is characterized by drinking for stress reduction and withdrawal coping motives (Verheul et al., 1999). The addiction cycle reflects relief drinking when alcohol use shifts to primarily negative reinforcement (known as the withdrawal-negative affect stage; Koob & Volkow, 2010, 2016; Verheul et al., 1999).

As described above, reward and relief drinking phenotypes overlap with stages of the Koob and Volkow (2010, 2016) alcohol addiction cycle and corresponding neurobiological dysfunction (Verheul et al., 1999). However, these phenotypes also align with the incentive motivational model of alcohol use (Verheul et al., 1999), which posits that alcohol use is a goal-driven behavior to obtain expected effects from alcohol that are characterized across two domains: approach or avoidance tendencies and internal or external motivation (Cox & Klinger, 2011; Votaw & Witkiewitz, 2021) Combinations of these two domains result in four motives for alcohol use: coping (internal, avoid), enhancement (internal, approach), conformity (external, avoid), and social (external, approach; Cox & Klinger, 2011; Votaw & Witkiewitz, 2021). Personality traits such as reward responsiveness, anxiety sensitivity, affinity for social reward, and subjective alcohol response, among others, likely also contribute to reward and relief drinking (Boness et al., 2021; Verheul et al., 1999).

Given that individuals with heavy alcohol use or AUD likely demonstrate some level of both reward and relief drinking (Verheul et al., 1999), recent work has leveraged latent variable mixture models to identify subgroups based on varying levels of reward and relief drinking, primarily in clinical samples of adults. Typically, four latent, phenotypic subgroups have been identified, including low reward/low relief, high reward/high relief, high reward/low relief, and high relief/low reward (Glöckner-Rist et al., 2013; Mann et al., 2018; Roos et al., 2017; Votaw, Mann, et al., 2022; Witkiewitz et al., 2019). Pharmacological precision medicine treatments for reward and relief drinking phenotypes have shown utility in improving alcohol treatment outcomes. Naltrexone, an opioid antagonist, has shown particular efficacy for those with high-reward/low-relief drinking (Mann et al., 2018; Witkiewitz et al., 2019), while acamprosate, a glutamatergic down-regulator, has shown efficacy in high-relief/low-reward drinkers (Roos et al., 2021).

Although most studies examining reward/relief drinking phenotypes and pharmacotherapy response have been conducted in adult samples with AUD, one recent study examined the reward drinker-naltrexone response hypothesis in young adults with frequent heavy drinking (Roos et al., 2021). Studies conducted with adult clinical samples used measures of craving and heavy drinking in various social contexts to assess reward and relief drinking (e.g., Inventory of Drinking Situations, Alcohol Abstinence Self-Efficacy Scale; Annis et al., 1982; DiClemente et al., 1994; Glöckner-Rist et al., 2013; Mann et al., 2018; Roos et al., 2017; Witkiewitz et al., 2019), while Roos et al. (2021) leveraged the enhancement and coping subscales of the Drinking Motives Questionnaire-Revised (DMQ-R) to assess reward and relief drinking, respectively. In this study, Roos et al. (2021) showed that the DMQ-R, which has been widely validated and examined in young adult samples (Kuntsche et al., 2005), effectively identified three phenotypic subgroups: high reward/high relief, high reward/low relief, and low reward/low relief. The lack of a high-relief/low-reward subgroup is consistent with findings that enhancement motives are more commonly reported in young adults than coping motives (Kuntsche et al., 2005). Furthermore, young adults with high-reward/low-relief and high-reward/high-relief subgroups derived benefits through treatment with naltrexone (Roos et al., 2021).

Current study

Assessing reward and relief drinking in non-clinical settings necessitates brief measures with good psychometric properties. The overarching aim of the present study was to examine the validity of reward and relief drinking phenotypes using the Drinking Motives Questionnaire-Revised-Short Form (DMQ-R-SF; Kuntsche & Kuntsche, 2009) in young adult college students. We first assessed the structural validity of the DMQ-R-SF, expecting that a two-factor reward and relief structure would provide a good fit to the data. Second, we examined latent reward and relief subgroups in exploratory latent profile analyses. For these first two aims, we compared models that measured reward drinking with only the DMQ-R-SF enhancement motives subscale to those that measured reward drinking with enhancement and social motives; in all models, relief drinking was measured with the coping motives subscale. Although prior work in young adults has only used the enhancement motives subscale to measure reward drinking (Roos et al., 2021), the three-pathway psychobiological model of alcohol craving suggests that reward drinking is composed of both incentive salience and social dimensions (Verheul et al., 1999). Furthermore, DMQ enhancement and social motives subscales are highly correlated (Kuntsche & Kuntsche, 2009) and prior studies of reward/relief drinking phenotypes in adults samples have used measures that capture drinking and craving in response to social situations (Mann et al., 2018; Roos et al., 2017; Votaw, Mann, et al., 2022; Witkiewitz et al., 2019).

After determining if models with or without social motives items were optimal for measuring the latent reward factor and identifying reward/relief drinking subgroups, we assessed the concurrent validity of the DMQ-R-SF-identified reward/relief subgroups for alcohol use and personality measures. We expected that subgroups characterized by high-reward drinking would be associated with sensation seeking, positive urgency, experiences of pleasure, alcohol quantity, and binge drinking frequency, while subgroups characterized by high-relief drinking would be associated with negative urgency, loneliness, rejection sensitivity, and alcohol consequences. Lastly, we assessed the predictive validity of the reward/relief subgroups identified for alcohol use outcomes 6 weeks following the initial assessment. We hypothesized that the high reward/high relief and high reward/low relief would have greater alcohol quantity, binge drinking frequency, and consequences than the low-reward/low-relief subgroup. The three-pathway psychobiological model of alcohol craving and prior studies on drinking motives informed these hypotheses (Cooper et al., 2015; Kuntsche et al., 2005; Verheul et al., 1999).

MATERIALS AND METHODS

Participants and design

The present study was a secondary analysis of data collected from 461 college students recruited from a university in the Northeast United States to participate in a three-wave longitudinal study of binge drinking among college students. Recruitment was conducted using SONA and participants were compensated with course credit. In addition, participants who participated in all three waves of data collection were entered into a lottery to win one of several $50 gift cards. Individuals who elected to participate were informed that that data would remain confidential, and that participation was voluntary at each timepoint. Eligible participants were 18 years of age or older and reported binge drinking in the 30 days before the baseline assessment. Participants completed online surveys approximately every 3 weeks over an academic semester with participant start dates spanning 1–2 months at the beginning of the semester. All procedures were approved by the Institutional Review Board at George Washington University. The present study utilized wave 1 (i.e., baseline; n = 454, given seven participants were missing data on all reward and relief items at baseline) and wave 3 (i.e., follow-up; n = 214 with any wave 3 data on variables included in follow-up analyses) data, which was collected approximately 6 weeks after the baseline assessment.

Full demographic information for participants who completed the baseline and follow-up assessments is presented in Table 1.

TABLE 1.

Descriptive statistics for those who completed the baseline (wave 1; n = 454) and those who completed the follow-up assessment (wave 3; n = 214).

Completed baseline (n = 454)
 Completed follow-up (n = 214)
n Mean (SD) or % n Mean (SD) or %
Age 453 19.61 (1.54) 213 19.58 (1.45)
Sex assigned at birth
 Female 354 78.0% 174 81.3%
 Male 100 22.0% 40 18.7%
Racial identity
 White 304 67.0% 143 66.8%
 Asian 58 12.8% 31 14.5%
 Multiracial 40 8.9% 19 8.9%
 Black or African American 23 5.1% 8 3.7%
 Another race 16 3.5% 7 3.3%
 Middle Eastern or North African 10 2.2% 2 0.9%
 American Indian or Alaska Native 1 0.2% 0 0.0%
 Native Hawaiian or Pacific Islander 1 0.2% 1 0.5%
 Missing 1 0.2% 3 1.4%
Hispanic ethnicity
 No 395 87.0% 184 86.0%
 Yes 59 10.0% 30 14.0%
Year in school
 First 160 35.2% 69 32.2%
 Second 111 24.4% 61 28.5%
 Third 104 22.9% 46 21.5%
 Fourth 79 17.4% 38 17.8%
Greek life member
 No 359 79.1% 165 77.1%
 Yes 95 20.9% 49 22.9%
Open in a new tab

Note: n = complete sample size for each measure.

Abbreviation: SD, standard deviation.

Compared to those who only completed the baseline assessment, those who also completed the follow-up did not differ on any demographic factors. Concerning reward and relief drinking measures (including identified latent classes) and concurrent and predictive validity measures (described below), those who completed the follow-up assessment had statistically significantly lower loneliness (completed baseline mean (SD) = 1.12 (0.72); completed follow-up mean (SD) = 0.97 (0.73), t(df) = 2.16 (448), p = 0.031) and sensation-seeking (completed baseline mean (SD) = 2.76 (0.69); completed follow-up mean (SD) = 2.63 (0.71), t(df) = 2.05 (449), p = 0.041) scores than those who only completed the baseline assessment, with no other differences identified.

Measures

Reward and relief drinking measures

The Drinking Motives Questionnaire-Revised-Short Form (DMQ-R-SF; Kuntsche & Kuntsche, 2009) was used to measure reward and relief drinking at the baseline assessment. The DMQ-R-SF contains four subscales assessing coping, enhancement, conformity, and social motives for alcohol use, with three items per subscale. Items assess drinking motives across contexts with response options ranging from 0 (almost never) to 4 (almost always) and responses are summed within subscale. Consistent with a prior analysis that used the DMQ-R to assess reward and relief drinking in young adults (Roos et al., 2021), we used items from the enhancement motives subscale to assess reward drinking (i.e., reward-enhancement; Cronbach’s a coefficient = 0.699) and the coping motives subscale to assess relief drinking (a = 0.839). Consistent with theory that reward drinking includes a social motives dimension (Verheul et al., 1999), we also examined the structural validity of a second two-factor model, which contained items from the both the enhancement and social motives subscales as one reward factor (i.e., reward-enhancement/social; a = 0.847) and items from the coping motives subscale as a relief factor (a = 0.839). For item-level descriptions and descriptive statistics, see Table 2.

TABLE 2.

Drinking motives questionnaire-revised-short form reward and relief items and descriptive statistics.

Item number Item description n Item mean (SD)
Reward-Enhancement 1 Because you like the feeling 454 3.47 (1.22)
Reward-Enhancement 2 To get high 453 2.10 (1.36)
Reward-Enhancement 3 Because it is fun 454 3.87 (1.14)
Reward-Social 4 Because it helps you enjoy a party 454 3.60 (1.19)
Reward-Social 5 Because it makes social gatherings more fun 454 3.66 (1.19)
Reward-Social 6 Because it improves parties and celebrations 454 3.55 (1.25)
Relief 1 To forget your worries 454 1.94 (1.15)
Relief 2 Because it helps when you feel depressed or nervous 453 1.90 (1.18)
Relief 3 To cheer you up when you are in a bad mood 454 2.14 (1.21)
Open in a new tab

Concurrent validity measures

All concurrent validity measures were administered at baseline. The Short UPPS-P Impulsive Behavior Scale (S-UPPS-P; Cyders et al., 2014) was used to assess domains of impulsivity hypothesized to be associated with reward and relief drinking by asking participants to rate their agreement with items on a scale of 1 (agree strongly) to 4 (disagree strongly), which are reverse scored so that higher scores indicate greater impulsivity and averaged within subscale. Selected subscales from the SUPPS-P reflecting domains of impulsivity used in our analyses included negative urgency (e.g., When I am upset, I often act without thinking; a = 0.803), positive urgency (e.g., I tend to lose control when I am in a great mood; a = 0.884), and sensation-seeking (e.g., I quite enjoy taking risks; a = 0.695).

The Temporal Experience of Pleasure scale (TEP; Gard et al., 2006) contains 18 questions assessing anticipatory (e.g., When something exciting is coming up in my life, I really look forward to it) and consummatory (e.g., I really enjoy the feeling of a good yawn) dimensions of pleasure on a scale of 1 (very true for me) to 6 (very false for me), which are summed with lower scores indicating greater anticipatory and consummatory pleasure. In the present sample, the TEP displayed good internal consistency reliability (a = 0.888).

The UCLA Loneliness Scale (Russell et al., 1978) consists of 20 items assessing loneliness (e.g., I am no longer close to anyone) with response options on a scale of 0 (never) to 3 (often), which are averaged. This scale displayed good internal consistency reliability in our sample (a = 0.963).

The Adult-Rejection Sensitivity Questionnaire (A-RSQ; Berenson et al., 2009; Downey et al., 2006; Downey & Feldman, 1996) contains nine items describing an interpersonal interaction (e.g., You ask your supervisor for help with a problem you have been having at work). Each items contains a pair of responses assessing anxiety or fear of rejection in that situation, with response options from 1 (very unconcerned) to 6 (very concerned). The second paired item asks participants to rate how likelihood of rejection on a scale of 1 (very unlikely) to 6 (very likely). The two items that comprise a pair are multiplied, and these multiplied scores are then averaged. The A-RSQ displayed acceptable internal consistency in our sample (a = 0.818).

Alcohol use measures for concurrent and predictive validity

Alcohol use outcomes for predictive validity were assessed at baseline and wave 3, and we examined average weekly drinking quantity, average weekly binge drinking frequency, and alcohol-related consequences. Participants completed the Daily Drinking Questionnaire (DDQ; Collins et al., 1985) which asks participants to report times that they typically consume alcohol during a given week using 4-h time blocks and report the number of standard drinks consumed in each block. Visual representations of standard drinks were provided. The average weekly drinking quantity was calculated by summing the total number of drinks consumed in a typical week over the prior 21 days. Binge drinking frequency was calculated by summing instances where participants reported consuming 5+ (male) or 4+ (female) drinks within a period of 2 h or less over the prior 21 days.

The Brief Young Adult Alcohol Consequences Questionnaire (BYAACQ; Kahler et al., 2005) was administered to assess alcohol-related consequences. Participants endorsed yes/no to 24 alcohol-related consequences in the prior 21 days (e.g., When drinking, I have said or done embarrassing things), and these items were summed (baseline a = 0.844; follow-up a = 0.876).

Statistical analyses

Examining the factor structure of the DMQ in assessing reward and relief drinking

To assess the structural validity of the DMQ-R-SF for measuring reward and relief drinking, we conducted and compared a series of five confirmatory factor analyses (CFAs). In models assessing reward drinking measured with only the DMQ-R-SF enhancement motives subscale (i.e., reward-enhancement) and relief drinking, we compared a one-factor model (i.e., all reward-enhancement and relief items loading onto one factor) and a two-factor model (i.e., a reward-enhancement factor and relief factor). In models that also included social motives as a measure of reward drinking (i.e., reward-enhancement/social), we conducted a one-factor model (i.e., all reward-enhancement/social and relief items loading onto one-factor), a two-factor model (i.e., a reward-enhancement/social factor and relief factor), and a three-factor model (i.e., a reward-enhancement factor, reward-social factor, and relief factor).

Model fit was evaluated using Comparative Fit Index (CFI; acceptable: >0.90, good: ≥0.95), Tucker-Lewis Index (TLI: acceptable: >0.90, good: ≥0.95), root mean square error of approximation (RMSEA; good: <0.05, acceptable: 0.05–0.08), and standardized root mean square residual SRMR (<0.08, acceptable; Hu & Bentler, 1998).

Identifying reward and relief drinking phenotypes

We conducted two sets of latent profile analyses (LPAs) to identify phenotypic subgroups with (1) reward-enhancement and relief items and (2) reward-enhancement/social and relief items from the DMQ-R-SF as indicators. For both sets of LPAs, we started with a two-profile solution and proceeded until the best-fitting model was identified. Model selection was based on Akaike’s Information Criteria (AIC), Bayes Information Criteria (BIC), and sample-size adjusted BIC (aBIC) with lower AIC, BIC, and aBIC values indicating better fitting models. We also used the Lo–Mendell Rubin adjusted likelihood ratio test (LRT), in which a statistically-significant LRT indicates a model with k profiles significantly improves over a model with k − 1 profiles, and theoretical interpretability (Nylund et al., 2007). Lastly, we assessed classification precision using model entropy.

Concurrent and predictive validity of reward and relief drinking phenotypes

We proceeded to concurrent and predictive validity analyses after selecting the set of items (e.g., reward-enhancement and relief versus reward-enhancement/social and relief) that displayed superior structural validity, produced theoretically interpretable latent profiles, and converged most clearly upon a profile solution. Concurrent validity was assessed using an automatic three-step BCH procedure (Bolck et al., 2004), examining differences in profile membership across sex, age, and concurrent validity measures described above. The BCH procedure estimates differences in concurrent validity variables across profile membership by saving poster probabilities of class membership for each participant and robustly examining their relation to each profile (Nylund-Gibson et al., 2019).

To examine the predictive validity of baseline reward/relief profile membership, measured with the DMQ-R-SF, for alcohol use outcomes, we used linear regression analyses. First, we saved participants’ most likely latent class membership from the final selected LPA. Then, separate linear regression analyses were estimated for each outcome at wave 3: average weekly alcohol quantity, average weekly binge drinking frequency, and alcohol-related consequences. All models examined the main effects of reward/relief profile membership, controlling for the baseline level of sex and age. Of note, for each outcome, we conducted two separate regression analyses; one examined the low-reward/low-relief subgroup as the reference group, and one examined the high-reward/high-relief subgroup as the reference to compare all subgroups.

Statistical packages and missing data

All analyses were conducted using MPlus Version 8.8 (Muthén & Muthén, 2022). For CFA models, we utilized the diagonally weighted mean least squares (WLSMV) estimator, which reduces bias compared to the robust maximum likelihood (MLR) estimator when observed variables are ordinal or categorical, as was in our data (Li, 2016). Missing data were handled using pairwise deletion under the WLSMV estimator, allowing participants to be included in analyses if they responded to any observed variable. For the LPA and BCH procedure, maximum likelihood estimation was used to account for missing data in the indicator variables and outcomes. Given the high rate of attrition at wave 3, we used multiple imputation to account for missing data for the linear regression analyses examining the predictive validity of the reward/relief drinking profiles. Accordingly, the total sample size was 454 for all analyses.

RESULTS

Examining the factor structure of the DMQ in assessing reward and relief drinking

Model fit statistics for the five CFA models evaluating the structural validity of the DMQ-R-SF items for measuring reward and relief drinking are presented in Table 3. Factor loadings for all five models are presented in Table S1. Beginning with models that included reward-enhancement and relief items, a one-factor CFA provided a poor fit to the data and a two-factor CFA provided an acceptable fit to the data. Specifically, for the two-factor model, CFI indicated good fit, TLI and SRMR indicated acceptable fit, and the chi-square and RMSEA estimates indicated poor fit. Standardized factors loadings fell within a range of 0.818 to 0.890 for the relief factor and 0.572 to 0.956 for the reward factor. Reward-Enhancement Item 2 (i.e., To get high) had the lowest factor loading at 0.572. The reward and relief latent factors were moderately correlated (r = 0.365, p < 0.001).

TABLE 3.

Model fit statistics for confirmatory factor analyses evaluating the structural validity of the DMQ-R-SF items for measuring reward and relief drinking.

Model X2 (df) p-value CFI TLI RMSEA (90% CI) SRMR
Reward-enhancement and relief items
 One-factor 471.782 (9) <0.001 0.817 0.694 0.337 (0.311, 0.363) 0.147
 Two-factor 91.735 (8) <0.001 0.967 0.938 0.152 (0.125, 0.181) 0.058
Reward-enhancement/Social and relief items
 One-factor 960.495 (27) <0.001 0.846 0.794 0.276 (0.261, 0.291) 0.148
 Two-factor 262.443 (26) <0.001 0.961 0.946 0.142 (0.126, 0.157) 0.061
 Three-factor 156.160 (24) <0.001 0.978 0.967 0.110 (0.094, 0.127) 0.053
Open in a new tab

Abbreviations: CFI, Comparative Fit Index; CI, confidence interval; df, degrees of freedom; RMSEA, root mean square error of approximation; SRMR, standardized root mean square residual; TLI, Tucker-Lewis Index.

Regarding CFA models that included reward-enhancement/social and relief items, all fit indices indicated that a one-factor model fit the data poorly. A two-factor reward-enhancement/social and relief model demonstrated improved model fit. Specifically, CFI, TLI, and SRMR indicated good fit, while the chi-square estimate and RMSEA indicated poor fit to the data. Standardized factor loadings for this model ranged from 0.484 to 0.906 for the reward factor and 0.812 to 0.831 for the relief factor, with Reward-Enhancement Item 2 again demonstrating the lowest factor loading. Reward and relief factors were moderately correlated (r = 0.365, p < 0.001). We also examined model fit of a three-factor CFA model, consisting of reward-enhancement, reward-social, and relief factors. As with the two-factor model, this three-factor model provided good fit to the data as indicated by the CFI, TLI, and SRMR, and poor fit according to the chi-square estimate and the RMSEA. Factor loadings ranged from 0.532 to 0.942 for the reward-enhancement factor (with Reward-Enhancement Item 2 demonstrating the lowest factor loading), 0.773 to 0.931 for the reward-social factor, and 0.812 to 0.884 for the relief factor. The relief factor was moderately correlated with both the reward-enhancement (r = 0.352, p < 0.001) and reward-social (r = 0.341, p < 0.001) factors, while the reward-enhancement and reward-social factors were strongly correlated (r = 0.779, p < 0.001).

Identifying reward and relief drinking phenotypes

Two sets of LPAs were conducted. The first included reward-enhancement and relief items as indicators, and the second included reward-enhancement/social and relief items as indicators. Model fit statistics and entropy for the two- through five-class LPA solutions for both sets of analyses are presented in Table 4. Furthermore, conditional item means and standard errors by latent class membership for the two- through five-profile solutions for both sets of LPAs are presented in Table S2.

TABLE 4.

Indicators of model fit for the 2- through 5-class latent profile analysis solutions with Drinking Motives Questionnaire-Revised-Short Form reward and relief drinking items as indicators.

Profiles AIC BIC aBIC LRT p-value Entropy
Reward-enhancement and relief
 2 8126.341 8204.585 8144.285 <0.001 0.917
 3 7835.662 7942.733 7860.218 <0.001 0.891
 4 7733.055 7868.952 7764.221 0.277 0.912
 5 7550.225 7714.949 7588.002 0.049 0.910
Reward-enhancement/Social and relief
 2 12,156.418 12,271.724 12,182.862 <0.001 0.903
 3 11,688.093 11,844.581 11,723.982 0.002 0.908
 4 11,481.403 11,679.072 11,526.736 0.192 0.932
 5 11,300.425 11,539.275 11,355.203 0.094 0.913
Open in a new tab

Note: The final selected model to proceed with construct validity evaluations is bolded.

Abbreviations: AIC, Akiake’s Information Criteria; BIC, Bayesian information criteria; aBIC, sample-size adjusted BIC; LRT, Lo-Mendell Rubin adjusted likelihood ratio test.

For LPAs that included reward-enhancement and relief drinking items as indicators, AIC, BIC, and aBIC all decreased across all estimated models. The LRT for the three-profile model was significant compared to the two-profile model (p < 0.001). This solution was theoretically interpretable, with profiles that we titled low-reward/low-relief drinking (Profile 1; n = 124, 27.31%), high-reward/low-relief drinking (Profile 2; n = 254, 55.95%), and high-reward/high-relief drinking (Profile 3; n = 76, 16.74%); conditional item means by latent profile are presented in Table S2. The LRT estimate for the four-profile model compared to the three-profile model was not statistically significant (p = 0.277), indicating the three-profile solution provided a better fit to the data than the four-profile solution. However, the five-profile solution was significant compared to the four-profile solution (p = 0.049), indicating that five-profile was an improvement over four-profile. The five-profile solution was not theoretically interpretable (see Table S2) and contained two small class sizes (Profile 4, n = 42, 9.25%; Profile 5, n = 38, 8.37%).

For LPAs that included reward-enhancement/social and relief drinking items as indicators, AIC, BIC, and aBIC decreased across all estimated models. The LRT for the three-profile model was significant compared to the two-profile model (p < 0.001). The LRT estimate for the four-profile model was not statistically significant compared to the three-profile model (p = 0.192), and the estimate for the five-profile solution was nonsignificant compared to the four-profile solution (p = 0.094), indicating the three-profile solution provided the best fit to the data. The three-profile solution was also theoretically interpretable, with profiles that we titled low-reward/low-relief drinking (Profile 1; n = 133, 29.3%), high-reward/low-relief drinking (Profile 2; n = 249, 54.8%), and high-reward/high-relief drinking (Profile 3; n = 72, 15.6%). Conditional item means by latent class membership for this three-profile solution are presented in Figure 1. Given the indicators of model fit for the LPA models that included reward-enhancement/social items more clearly converged on one final profile solution than those that only included reward-enhancement items (i.e., multiple statistically-significant LRT p-values for the models that only included reward-enhancement items), we proceeded with the three-profile solution presented in Figure 1 (i.e., including reward-enhancement/social items) for our concurrent and predictive validity analyses.

FIGURE 1.

FIGURE 1

Open in a new tab

Conditional means on drinking motives questionnaire-revised-short form (DMQ-R-SF) reward and relief drinking items by latent profile membership.

Concurrent validity of reward and relief drinking phenotypes

Results of the BCH analyses (see Table 5 for full results) indicated generally good concurrent validity. Sex and S-UPPS-P sensation seeking did not differ by latent profile membership. Age was significantly different across profile membership, such that those in the low-reward/low-relief profile were older compared to the high-reward/low-relief profile and the high-reward/high-relief profile. The high-reward/high-relief profile was associated with significantly higher scores on baseline alcohol quantity, binge drinking frequency, alcohol-related consequences, S-UPPS-P negative urgency, S-UPPS-P positive urgency, TEP scores (indicating lower experiences of pleasure), and UCLA Loneliness Scale scores, and A-RSQ scores compared to the low-reward/low-relief profile. Furthermore, the high-reward/high-relief profile evidenced greater alcohol-related consequences, S-UPPS-P negative urgency, S-UPPS-P positive urgency, TEP scores (indicating lower experiences of pleasure), UCLA Loneliness Scale scores, and A-RSQ scores than the high-reward/low-relief profile. Membership in the high-reward/low-relief profile was associated with significantly higher baseline alcohol quantity, binge drinking frequency, alcohol-related consequences, S-UPPS-P negative urgency, and UCLA Loneliness Scale scores compared to the low-reward/low-relief group.

TABLE 5.

Comparisons among reward/relief drinking subgroups on concurrent validity measures using the BCH approach.

Concurrent validity variable Profile 1 (low reward/low relief) Profile 2 (high reward/low relief) Profile 3 (high reward/high relief) Overall test p-value Profile 1 vs. 2 p-value Profile 1 vs. 3 p-value Profile 2 vs. 3 p-value
DMQ-R-SF reward-enhancement 6.564 (0.223) 10.395 (0.143) 11.485 (0.337) <0.001 <0.001 <0.001 0.003
DMQ-R-SF reward-enhancement/Social 13.102 (0.309) 22.968 (0.209) 24.078 (0.543) <0.001 <0.001 <0.001 0.061
DMQ-R-SF relief 4.576 (0.177) 4.939 (0.112) 12.022 (0.254) <0.001 0.092 <0.001 <0.001
Male (%) 0.274 (0.040) 0.208 (0.027) 0.164 (0.046) 0.187 0.187 0.075 0.42
Age 20.019 (0.177) 19.419 (0.084) 19.499 (0.174) 0.011 <0.003 0.037 0.685
S-UPPS-P negative urgency 1.791 (0.067) 1.971 (0.048) 2.575 (0.104) <0.001 0.033 <0.001 <0.001
S-UPPS-P sensation seeking 2.714 (0.067) 2.709 (0.045) 2.634 (0.088) 0.727 0.947 0.469 0.457
S-UPPS-P positive urgency 1.622 (0.068) 1.698 (0.044) 2.111 (0.112) 0.001 0.363 <0.001 0.001
Temporal Experience of Pleasure Scale 39.673 (1.555) 38.559 (1.081) 47.105 (2.364) 0.005 0.567 0.009 0.001
UCLA Loneliness Scale 0.812 (0.057) 0.988 (0.046) 1.719 (0.084) <0.001 0.019 <0.001 <0.001
Adult-Rejection Sensitivity Questionnaire 8.081 (0.370) 8.824 (0.307) 12.344 (0.665) <0.001 0.134 <0.001 <0.001
DDQ alcohol quantity 11.243 (0.805) 13.321 (0.554) 15.734 (1.314) 0.01 0.039 <0.004 0.096
DDQ binge frequency 3.648 (0.219) 4.439 (0.147) 4.801 (0.323) 0.003 0.004 0.003 0.315
BYAACQ alcohol consequences 4.649 (0.407) 5.853 (0.243) 8.695 (0.609) <0.001 0.013 <0.001 <0.001
Open in a new tab

Abbreviations: BYAACQ, Brief Young Adult Alcohol Consequences Questionnaire; DDQ, Daily Drinking Questionnaire; DMQ-R-SF, Drinking Motives Questionnaire-Revised-Short Form; S-UPPS-P, Short UPPS-P Impulsive Behavior Scale.

Predictive validity of reward and relief drinking phenotypes

Linear regression models examining the predictive validity of the reward/relief drinking subgroups for alcohol outcomes at wave 3 are presented in Table 6. Profile membership was not significantly associated with quantity at wave 3 across profile membership (all ps > 0.05). However, the high-reward/low-relief subgroup reported greater binge drinking frequency at wave 3 compared to the low-reward/low-relief subgroup (b(SE) = 1.215 (0.355), p = 0.001). There were no differences in wave 3 binge drinking frequency between the low-reward/low-relief subgroup and the high-reward/high-relief subgroup or between the high-reward/low-relief subgroup and the high-reward/high-relief subgroup (all ps >0.05). Additionally, membership in the high-reward/high-relief subgroup (b(SE) = 3.745 (0.870), p < 0.001) and the high-reward/low-relief subgroup (b(SE) = 1.800 (0.637), p = 0.005) predicted greater alcohol consequences at wave 3 compared to the low-reward/low-relief subgroup. Membership in the high-reward/low-relief subgroup predicted significantly lower consequences at follow-up compared to the high-reward/high-relief subgroup (b(SE) = −1.945 (0.758), p = 0.010).

TABLE 6.

Linear regression analyses examining the effects of reward/relief drinking profile membership at baseline (wave 1) on quantity, binge frequency, and alcohol consequences at follow-up (wave 3).

Alcohol quantity (DDQ)
 Binge frequency (DDQ)
 Alcohol consequences (BYAACQ)
Variable b (SE) P b (SE) P b (SE) P
Analyses with the low-reward/low-relief group as the reference group
 Intercept 11.150 (8.811) 0.206 6.076 (2.127) <0.01 6.966 (3.694) 0.059
 Low reward/Low relief Reference Reference Reference
 High reward/Low relief 2.591 (1.425) 0.069 1.215 (0.355) 0.001 1.800 (0.637) 0.005
 High reward/High relief 0.229 (1.883) 0.903 0.647 (0.440) 0.142 3.745 (0.870) <0.001
 Age −0.095 (0.445) 0.831 −0.163 (0.105) 0.121 −0.196 (0.190) 0.304
 Male sex 4.136 (1.738) 0.017 0.588 (0.393) 0.135 −0.121 (0.700) 0.862
Analyses with the high-reward/high-relief group as the reference group
 Intercept 11.379 (8.840) 0.198 6.723 (2.077) 0.001 10.711 (3.680) 0.004
 Low reward/Low relief −0.229 (1.883) 0.903 −0.647 (0.440) 0.142 −3.745 (0.870) <0.001
 High reward/Low relief 2.362 (1.817) 0.194 0.568 (0.415) 0.171 −1.945 (0.758) 0.010
 High reward/high relief Reference Reference Reference
 Age −0.095 (0.445) 0.831 −0.163 (0.105) 0.121 −0.196 (0.190) 0.304
Male sex 4.136 (1.738) 0.017 0.588 (0.393) 0.135 −0.121 (0.700) 0.862
Open in a new tab

Abbreviations: b = unstandardized regression estimate, BYAACQ, Brief Young Adult Alcohol Consequences Questionnaire; DDQ, Daily Drinking Questionnaire; SE, standard error.

Linear regression analyses including baseline outcome variables as covariates are presented in Table S3. Notably, when controlling for baseline alcohol quantity, membership in the high-reward/high-relief subgroup predicted lower average weekly alcohol quantity at wave 3 compared to the low-reward/low-relief subgroup (b(SE) = −3.367 (1.589), p = 0.034). However, when omitting baseline alcohol quantity, this association was nonsignificant (b(SE) = 0.229 (1.883), p = 0.069), indicating suppression effects of the baseline alcohol quantity variable.

DISCUSSION

Characterizing reward and relief drinking phenotypes in young adults has the potential to inform targeted prevention efforts and brief interventions to mitigate consequences associated with heavy drinking (Cronce & Larimer, 2011; Slutske, 2005). Prior studies have examined the validity of reward/relief drinking phenotypes in adults (Mann et al., 2018; Roos et al., 2017; Witkiewitz et al., 2019) and young adults enrolled in clinical trials of pharmacotherapies for AUD (Roos et al., 2021). We extended this prior work by examining the structural validity of a brief measure, the DMQ-R-SF, to assess reward/relief drinking in non-treatment-seeking college students with recent binge drinking. We also evaluated the ability of this measure to identify latent reward/relief drinking profiles and the concurrent and predictive validity of these profiles for personality and alcohol use constructs. Notably, we examined the extent to which including reward-social motives in addition to reward-enhancement motives improved the structural validity of the DMQ-R-SF for assessing reward and relief drinking and identification of phenotypic subgroups.

We first aimed to examine the structural validity of the DMQ-R-SF for measuring reward and relief drinking, including a six-item version with reward-enhancement and relief motives, consistent with previous work by Roos et al. (2021), and a nine-item version that included reward-social motives. Across both versions of the DMQ-R-SF evaluated, results of the CFA indicated a two-factor model representing reward and relief drinking, provided a moderately good fit to the data, though two indicators of model fit (i.e., RMSEA and the chi-square estimate) indicated poor fit to the data. In addition, the TLI estimate for the six-item version of the DMQ-R-SF evaluating only reward-enhancement motives indicated acceptable fit as opposed to good fit. Notably, model fit was similar across the two-factor reward-enhancement/social and relief model and three-factor reward-enhancement, reward-social, and relief model, and the correlation between the reward-enhancement and reward-social factors was high (r = 0.779). These findings align with premises of the three-pathway psychobiological model of alcohol craving that reward drinking is composed of both incentive salience and social dimensions (Verheul et al., 1999).

As both the DMQ-R-SF reward-enhancement and the reward-enhancement/social models provided generally good fit to the data, we proceeded with latent profile analyses of both measures. The results of these latent profile analyses indicated that utilizing items from the reward-enhancement/social and relief subscales enabled clearer selection of a three-profile solution than using only the reward-enhancement subscale. We identified three distinct motivational profiles utilizing the reward-enhancement/social and relief items in our sample of young adults with recent binge drinking: low reward/low relief, high reward/low relief, and high reward/high relief. These three profiles are consistent with those identified in a prior study of treatment-seeking young adults with heavy alcohol use (Roos et al., 2021), and the absence of a high-relief/low-reward drinking profile is consistent with work indicating that coping motives are relatively uncommon in college students (Kuntsche et al., 2005). This finding also aligns with theoretical premises that reward and relief drinking processes reflect the addiction cycle, with relief drinking emerging at later stages of alcohol use severity (Koob & Volkow, 2016; Verheul et al., 1999).

Although not a primary aim of the present study, we also attempted to examine the ability of the Reward and Relief Inventory of Drinking Situations (RR-IDS; Votaw, Mann, et al., 2022), which assesses frequency of heavy drinking in various situations, to identify reward/relief drinking phenotypes in the present sample and compare the validity of phenotypes identified with the RR-IDS versus the DMQ-R-SF. Although the two-factor structure of the RR-IDS provided a good fit to the data in the present sample (Χ2[34] = 95.020, p < 0.001; CFI = 0.995; TLI = 0.993; RMSEA = 0.064 [90% CI: 0.049, 0.079], p = 0.063; SRMR = 0.038), distinct reward/relief drinking profiles were not identified, with participants only classified within overall low or overall high profiles. This is, somewhat unsurprising given the RR-IDS was developed with adults with AUD (Annis & Graham, 1995; Votaw, Mann, et al., 2022), while the DMQ-R-SF was developed with young adult and adolescent sample (Cooper, 1994; Kuntsche & Kuntsche, 2009). Taken together these findings indicate that the DMQ-R-SF reward-enhancement/social and relief model is likely a better measure compared to the IDS to identify distinct reward/relief drinking phenotypes and inform precision interventions in young adults with heavy alcohol use.

The results of concurrent and predictive validity analyses based on our identified three-profile solution of the DMQ-R-SF reward-enhancement/social and relief subscales were largely consistent with our hypotheses and the three-pathway psychobiological model of alcohol craving (Verheul et al., 1999), with some exceptions. Results indicating that the high-reward/high-relief profile was characterized by greater negative urgency, loneliness, rejection sensitivity, and alcohol consequences than the other two profiles reflect premises that relief drinking is motivated by greater experience of and reactivity to negative affect (Verheul et al., 1999) and findings that coping motives are consistently associated with greater alcohol-related consequences (Cooper et al., 2016). Also consistent with hypotheses, those in the high-reward/low-relief subgroup demonstrated greater alcohol quantity, binge drinking frequency, and consequences than those in the low-reward/low-relief drinking profile.

Unexpectedly, the high-reward/high-relief profile had greater levels of positive urgency than the high-reward/low-relief profiles. Some AUD frameworks have considered this construct a cognitive control deficit, as opposed to a reward deficit, that contributes to the etiology and maintenance of alcohol use (Boness & Witkiewitz, 2022). Accordingly, those with high relief-motivated alcohol use might have greater positive urgency than those with high reward-motivated drinking, given the strong associations between relief drinking and cognitive control deficits (Grodin et al., 2019; Votaw, Boness, et al., 2022). Although we did not expect to identify a high-relief/low-reward profile in the present study, the lack of this profile precludes strong conclusions about the unique role of relief drinking processes in explaining construct validity findings. It was also somewhat surprising that experiences of pleasure (measured by the TEP) were not elevated in the subgroups characterized by high-reward drinking, and were lower among the high-reward/high-relief subgroup compared to the other two subgroups (see Table 5). Reward drinking is hypothesized to be driven, in part, by reward responsiveness, a construct closely related to experiences of pleasure (Verheul et al., 1999). However, the measure of experiences of pleasure used in the present study was general, and future investigations should also assess if differences between these two profiles exist in alcohol-related reward (Grodin et al., 2019).

Results of the predictive validity analyses suggested that those in the high-reward/low-relief profile demonstrated greater binge drinking frequency and alcohol-related consequences 6 weeks following the initial assessment than the low-reward/low-relief profile. Membership in the high-reward/high-relief profile prospectively predicted greater alcohol-related consequences than membership in the other two profiles. These findings align with prior work indicating enhancement motives are associated with binge drinking, while coping motives are associated with alcohol-related consequences (Cooper et al., 2016). These two profiles, identified with the DMQ-R-SF, might benefit from interventions preventing the onset of more severe alcohol problems. Brief interventions targeted to distinct motives (e.g., providing alternative coping strategies for those high in coping motives) have shown some promise, though this work has largely focused on coping motives instead of enhancement motives (Blevins & Stephens, 2016; Reid & Carey, 2015). Given the greater rates of enhancement and social motives in young adult heavy drinkers (Kuntsche et al., 2005) and frequent and heavy alcohol use and binge drinking among those with high-reward/low-relief drinking in the present sample, interventions targeting enhancement and social motives (reward drinking) are also warranted. For example, those who drink primarily for reward might benefit from brief interventions to increase engagement in reinforcing alternatives to substance use (Murphy et al., 2019). Future research should examine longitudinal trajectories of alcohol use in unique reward/relief drinking profiles to determine optimal timing and targets for prevention efforts to mitigate alcohol-related consequences. It may also be useful for future work to compare the clinical utility of the subgroups identified with reward-enhancement/social motives versus reward-enhancement motives, given prior work examining the reward/relief phenotype for precision medicine in young adults has only used the enhancement motives subscale (Roos et al., 2021).

Limitations

The results of this manuscript should be interpreted in the context of several limitations. First, only 47% of participants who completed the baseline assessment also completed the wave 3 assessment, a rate that parallels other studies that recruited college students using SONA (Booker et al., 2022; Nelsen et al., 2023). Although we used multiple imputation to account for missing data at wave 3 and those who completed the follow-up did not significantly differ on demographic or alcohol-related variables from those who completed only the baseline assessment, it is possible that this high attrition rate impacted predictive validity results. Notably, we had a relatively short follow-up period of 6 weeks. Future work should utilize a longer timeframe in assessing predictive validity of reward/relief profiles in young adults and the stability of these profiles. In addition, the majority of participants in the present sample were female and White. Future work is needed to determine the measurement invariance of DMQ-R-SF reward and relief drinking subscales across various demographic factors to ensure this measure can be used to inform targeted prevention efforts for racially, ethnically, and gender-diverse young adults with heavy alcohol use. In addition, this longitudinal study was not designed to assess the validity of reward/relief drinking phenotypes, and therefore this secondary data analysis was limited to measures included in the study to meet the primary aims. Future work examining additional construct validity variables (e.g., reinforcement from substance-free vs substance-related activities, anxiety and depression symptoms) would help further clarify the validity of reward/relief drinking phenotypes and mechanisms by which these processes develop.

CONCLUSION

In conclusion, the results of the present study provide initial evidence for the validity of the DMQ-R-SF as a measure of reward/relief drinking phenotypes in non-treatment-seeking seeking young adults with recent binge drinking. Further, the inclusion of the reward-social subscale in operationalizing reward drinking improved subgroup delineation over the DMQ-R-SF reward-enhancement and relief subscales. Given the concurrent and prospective associations between the high-reward/low-relief and high-reward/high-relief profiles and poorer alcohol use outcomes, future work might examine precision prevention efforts to reduce the harms of alcohol use for those in these two subgroups. It is important to note that statistical methods used in this manuscript are not easily implementable within medical and University platforms. Developing observed cutoff scores for DMQ-R-SF-identified reward/relief drinking profiles in young adults (e.g., Votaw, Mann, et al., 2022) might help facilitate such research. Furthermore, future studies are needed to: (1) replicate the present findings and continue validating reward/relief drinking profiles identified with the DMQ-R-SF reward-enhancement/social and relief factors using additional concurrent validity measures, (2) examine the measurement invariance of these reward and relief factors in diverse samples of young adults, (3) examine alcohol use trajectory in reward/relief drinking profiles longitudinally to inform the optimal time to deliver prevention and intervention efforts in those with distinct motivational pathways, and (4) compare the clinical utility of the reward/relief phenotypes identified with reward-enhancement/social items versus reward/enhancement items alone.

Supplementary Material

Supplementary

FUNDING INFORMATION

Efforts on this manuscript were supported by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) of the National Institutes of Health, award number F31 AA029266 (PI: Votaw). The content is the sole responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

SUPPORTING INFORMATION

Additional supporting information can be found online in the Supporting Information section at the end of this article.

DATA AVAILABILITY STATEMENT

Research data are not shared.

REFERENCES

  1. Annis HM & Graham JM (1995) Profile types on the inventory of drinking situations: implications for relapse prevention counseling. Psychology of Addictive Behaviors, 9, 176–182. [Google Scholar]
  2. Annis HM,Graham JM & Davis CS(1982) Inventory of drinking situations. Ontario, Canada: Addiction Research Foundation. [Google Scholar]
  3. Berenson KR, Gyurak A, Ayduk Ö, Downey G, Garner MJ, Mogg K et al. (2009) Rejection sensitivity and disruption of attention by social threat cues. Journal of Research in Personality, 43, 1064–1072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Blevins CE & Stephens RS (2016) The impact of motives-related feedback on drinking to cope among college students. Addictive Behaviors, 58, 68–73. [DOI] [PubMed] [Google Scholar]
  5. Bolck A, Croon M & Hagenaars J (2004) Estimating latent structure models with categorical variables: one-step versus three-step estimators. Political Analysis, 12, 3–27. [Google Scholar]
  6. Boness CL, Watts AL, Moeller KN & Sher KJ (2021) The etiologic, theory-based, ontogenetic hierarchical framework of alcohol use disorder: a translational systematic review of reviews. Psychological Bulletin, 147, 1075–1123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Boness CL & Witkiewitz K (2022) Precision medicine in alcohol use disorder: mapping etiologic and maintenance mechanisms to mechanisms of behavior change to improve patient outcomes. Experimental and Clinical Psychopharmacology, 31, 769–779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Booker JA, Hernandez E, Talley KE & Dunsmore JC (2022) Connecting with others: dispositional and situational relatedness during the college transition. Journal of Social and Personal Relationships, 39, 198–220. [Google Scholar]
  9. Carter AC, Brandon KO & Goldman MS (2010) The college and noncollege experience: a review of the factors that influence drinking behavior in young adulthood. Journal of Studies on Alcohol and Drugs, 71, 742–750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Collins RL, Parks GA & Marlatt GA (1985) Social determinants of alcohol consumption: the effects of social interaction and model status on the self-administration of alcohol. Journal of Consulting and Clinical Psychology, 53, 189–200. [DOI] [PubMed] [Google Scholar]
  11. Cooper ML (1994) Motivations for alcohol use among adolescents: development and validation of a four-factor model. Psychological Assessment, 6, 117–128. [Google Scholar]
  12. Cooper ML, Kuntsche E, Levitt A, Barber LL & Wolf S (2015) Motivational models of substance use. Oxford Handbooks Online. [Google Scholar]
  13. Cooper ML, Kuntsche E, Levitt A, Barber LL & Wolf S (2016) Motivational models of substance use: a review of theory and research on motives for using alcohol, marijuana, and tobacco. In: The Oxford handbook of substance use and substance use disorders, vol. 1, Oxford library of psychology. New York: Oxford University Press, pp. 375–421. [Google Scholar]
  14. Cox WM & Klinger E (2011) A motivational model of alcohol use: determinants of use and change. In: Handbook of motivational counseling: goal-based approaches to assessment and intervention with addiction and other problems, 2nd edition. Hoboken: Wiley Blackwell, pp. 131–158. [Google Scholar]
  15. Cronce JM & Larimer ME (2011) Individual-focused approaches to the prevention of college student drinking. Alcohol Research & Health, 34, 210–221. [PMC free article] [PubMed] [Google Scholar]
  16. Cyders MA, Littlefield AK, Coffey S & Karyadi KA (2014) Examination of a short English version of the UPPS-P Impulsive Behavior Scale. Addictive Behaviors, 39, 1372–1376. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. DiClemente CC, Carbonari JP, Montgomery RPG & Hughes SO (1994) The alcohol abstinence self-efficacy scale. Journal of Studies on Alcohol, 55, 141–148. [DOI] [PubMed] [Google Scholar]
  18. Downey G, Berenson KR & Kang J (2006) Adult Rejection Sensitivity Questionnaire. [Google Scholar]
  19. Downey G & Feldman SI (1996) Implications of rejection sensitivity for intimate relationships. Journal of Personality and Social Psychology, 70, 1327–1343. [DOI] [PubMed] [Google Scholar]
  20. Gard DE, Gard MG, Kring AM & John OP (2006) Anticipatory and consummatory components of the experience of pleasure: a scale development study. Journal of Research in Personality, 40, 1086–1102. [Google Scholar]
  21. Glöckner-Rist A, Lémenager T, Mann K & PREDICT Study Research Group. (2013) Reward and relief craving tendencies in patients with alcohol use disorders: results from the PREDICT study. Addictive Behaviors, 38, 1532–1540. [DOI] [PubMed] [Google Scholar]
  22. Grodin EN, Bujarski S, Venegas A, Baskerville W-A, Nieto SJ, Jentsch JD et al. (2019) Reward, relief and habit drinking: initial validation of a brief assessment tool. Alcohol and Alcoholism, 54, 574–583. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hu L & Bentler PM (1998) Fit indices in covariance structure modeling: sensitivity to underparameterized model misspecification. Psychological Methods, 3, 424–453. [Google Scholar]
  24. Jennison KM (2004) The short-term effects and unintended long-term consequences of binge drinking in college: a 10-year follow-up study. The American Journal of Drug and Alcohol Abuse, 30, 659–684. [DOI] [PubMed] [Google Scholar]
  25. Kahler CW, Strong DR & Read JP (2005) Toward efficient and comprehensive measurement of the alcohol problems continuum in college students: the brief young adult alcohol consequences questionnaire. Alcoholism, Clinical and Experimental Research, 29, 1180–1189. [DOI] [PubMed] [Google Scholar]
  26. Koob GF & Volkow ND (2010) Neurocircuitry of addiction. Neuropsychopharmacology, 35, 217–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Koob GF & Volkow ND (2016) Neurobiology of addiction: a neurocircuitry analysis. Lancet Psychiatry, 3, 760–773. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kuntsche E, Knibbe R, Gmel G & Engels R (2005) Why do young people drink? A review of drinking motives. Clinical Psychology Review, 25, 841–861. [DOI] [PubMed] [Google Scholar]
  29. Kuntsche E & Kuntsche S (2009) Development and validation of the drinking motive questionnaire revised short form (DMQ–R SF). Journal of Clinical Child and Adolescent Psychology, 38, 899–908. [DOI] [PubMed] [Google Scholar]
  30. Li C-H (2016) Confirmatory factor analysis with ordinal data: comparing robust maximum likelihood and diagonally weighted least squares. Behavior Research Methods, 48, 936–949. [DOI] [PubMed] [Google Scholar]
  31. Litten RZ, Ryan ML, Falk DE, Reilly M, Fertig JB & Koob GF (2015) Heterogeneity of alcohol use disorder: understanding mechanisms to advance personalized treatment. Alcoholism, Clinical and Experimental Research, 39, 579–584. [DOI] [PubMed] [Google Scholar]
  32. Mann K, Roos CR, Hoffmann S, Nakovics H, Leménager T, Heinz A et al. (2018) Precision medicine in alcohol dependence: a controlled trial testing pharmacotherapy response among reward and relief drinking phenotypes. Neuropsychopharmacology, 43, 891–899. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Martens MP, Neighbors C, Lewis MA, Lee CM, Oster-Aaland L & Larimer ME (2008) The roles of negative affect and coping motives in the relationship between alcohol use and alcohol-related problems among college students. Journal of Studies on Alcohol and Drugs, 69, 412–419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Murphy JG, Dennhardt AA, Martens MP, Borsari B, Witkiewitz K & Meshesha LZ (2019) A randomized clinical trial evaluating the efficacy of a brief alcohol intervention supplemented with a substance-free activity session or relaxation training. Journal of Consulting and Clinical Psychology, 87, 657–669. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Muthén B & Muthén L (2022) MPlus user’s guide. In: Handbook of item response theory. Los Angeles: Muthén & Muthén. [Google Scholar]
  36. Nelsen SK, Kayaalp A & Page KJ (2023) Perfectionism, substance use, and mental health in college students: a longitudinal analysis. Journal of American College Health, 71, 257–265. [DOI] [PubMed] [Google Scholar]
  37. Nylund KL, Asparouhov T & Muthén BO (2007) Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14, 535–569. [Google Scholar]
  38. Nylund-Gibson K, Grimm RP & Masyn KE (2019) Prediction from latent classes: a demonstration of different approaches to include distal outcomes in mixture models. Structural Equation Modeling: A Multidisciplinary Journal, 26, 967–985. [Google Scholar]
  39. Prince MA, Read JP & Colder CR (2019) Trajectories of college alcohol involvement and their associations with later alcohol use disorder symptoms. Prevention Science, 20, 741–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Reid AE & Carey KB (2015) Interventions to reduce college student drinking: state of the evidence for mechanisms of behavior change. Clinical Psychology Review, 40, 213–224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Robinson TE & Berridge KC (1993) The neural basis of drug craving: an incentive-sensitization theory of addiction. Brain Research. Brain Research Reviews, 18, 247–291. [DOI] [PubMed] [Google Scholar]
  42. Robinson TE & Berridge KC (2001) Incentive-sensitization and addiction. Addiction, 96, 103–114. [DOI] [PubMed] [Google Scholar]
  43. Roos CR, Bold KW, Witkiewitz K, Leeman RF, DeMartini KS, Fucito LM et al. (2021) Reward drinking and naltrexone treatment response among young adult heavy drinkers. Addiction, 116, 2360–2371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Roos CR, Mann K & Witkiewitz K (2017) Reward and relief dimensions of temptation to drink: construct validity and role in predicting differential benefit from acamprosate and naltrexone. Addiction Biology, 22, 1528–1539. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Russell D, Peplau LA, & Ferguson ML (1978) Developing a Measure of Loneliness. Journal of Personality Assessment, 42, 290–294. [DOI] [PubMed] [Google Scholar]
  46. Slutske WS (2005) Alcohol use disorders among US college students and their non–college-attending peers. Archives of General Psychiatry, 62, 321–327. [DOI] [PubMed] [Google Scholar]
  47. Verheul R, van den Brink W & Geerlings P (1999) A three-pathway psychobiological model of craving for alcohol. Alcohol and Alcoholism, 34, 197–222. [DOI] [PubMed] [Google Scholar]
  48. Votaw VR, Boness CL, Stein E, Watts A, Sher K & Witkiewitz K (2022) Examining the validity of the addictions Neuroclinical assessment domains in a crowdsourced sample of adults with current alcohol use. Experimental and Clinical Psychopharmacology. (published online ahead of print 25 May 2023) Available from: 10.1037/pha0000648 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Votaw VR, Mann K, Kranzler HR, Roos CR, Nakovics H & Witkiewitz K (2022) Examining a brief measure and observed cutoff scores to identify reward and relief drinking profiles: psychometric properties and pharmacotherapy response. Drug and Alcohol Dependence, 232, 109257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Votaw VR & Witkiewitz K (2021) Motives for substance use in daily life: a systematic review of studies using ecological momentary assessment. Clinical Psychological Science, 9, 535–562. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Witkiewitz K, Roos CR, Mann K & Kranzler HR (2019) Advancing precision medicine for alcohol use disorder: replication and extension of reward drinking as a predictor of naltrexone response. Alcoholism, Clinical and Experimental Research, 43, 2395–2405. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary
NIHMS1987711-supplement-Supplementary.docx (28.9KB, docx)

Data Availability Statement

Research data are not shared.

RESOURCES