The Sequential Indirect Effect of Negative Urgency on Drinking Consequences Through Distress Intolerance and Drinking Motives: Initial Examination in College Students Reporting Past Month Alcohol Use

Abstract

Identification of cognitive and affective vulnerabilities among college drinkers may aid in developing focused interventions that promote a reduction in the prevalence of alcohol use. Negative urgency (NU) and distress intolerance (DI) evidence concurrent, unique, and synergistic relations with drinking motives and negative consequences of alcohol use. Utilizing a sequential multiple mediation framework to investigate a comprehensive model of these variables, we examined NU as a behavioral risk factor that potentiates the development of DI, thereby contributing to drinking motives that increase the risk of problematic use in young adults. A diverse sample of undergraduate students (N = 616; M_age= 19.1, SD=1.4, range=18–25; 50.6% female; 60.6% Caucasian; recruited between September 2015 and Spring 2017) reporting past month alcohol use completed an online questionnaire battery. The results suggested that NU may contribute to negative alcohol use outcomes via its relation to DI and the motivation to drink in order to cope with negative emotional states and conform to social pressure. These findings suggest that NU may be a primary intervention target in young adults.

Problematic alcohol use is prevalent among college-aged students, with 39.0% reporting a past month binge and 12.7% reporting heavy alcohol use (Substance Abuse and Mental Health Services Administration 2013). Extant studies show early drinking onset (e.g., < 20 years old) is prospectively related to higher odds of alcohol dependence and involvement in risky behaviors compared to later drinking onset (e.g., > 20 years old; Hingson and Zha 2009). Given the relation between alcohol use and immediate negative outcomes (Hingson et al. 2009) and increased risk for long-term problems (e.g., including injury, sexual abuse, and death; Hingson et al. 2005, 2009; White and Hingson 2014), systematic prevention efforts have been put forward (Schaus et al. 2009; Weitzman et al. 2004). However, extant interventions have resulted in little change in the prevalence of problematic alcohol use in this population (Wechsler et al. 2002; White and Hingson 2014). Moreover, rather than interpret a causal relation between age of drinking onset and negative consequences, researchers have suggested that other factors, such as environment, temperament, and behavioral patterns, should be central to prevention and treatment efforts in college samples (e.g., Hingson et al. 2012; Kuntsche et al. 2016). Therefore, identification of college drinkers at the greatest risk of developing alcohol dependence, and behavioral vulnerabilities that may increase risk, may lead to defining potential treatment targets. Here, underlying psychological vulnerabilities may play a role in the development of beliefs about the utility of alcohol and, hence, lead to alcohol use with adverse consequences. Specifically, an individual’s response tendency to negative affect may contribute to beliefs that negative affect is intolerable, contributing to a model by which to understand behavioral and cognitive factors that promote persistent problematic alcohol use (Kaiser et al. 2012).

Negative urgency (NU), the behavioral tendency to act rashly under intense negative emotion, has been linked to both problematic alcohol use and beliefs that may contribute to longer-term drinking problems, such as heavy drinking (Cyders and Smith 2008). For example, cross-sectional and longitudinal studies have found that NU is associated with problematic drinking (e.g., quantity, frequency, and dependence; Curcio and George 2011; Dir et al. 2013; Kaiser et al. 2012; Kuvaas et al. 2014; Murphy and MacKillop 2012; Shishido et al. 2013) and negative alcohol consequences (e.g., unplanned sex and reckless driving; Curcio and George 2011; Karyadi and King 2011; Keough et al. 2015; Murphy and MacKillop 2012; Shishido et al. 2013; Wardell et al. 2016) among college students, with results extending to both clinical and nonclinical population with large effect sizes for both populations (Coskunpinar et al. 2013). NU also appears to be related to the development of strongly held reasons for alcohol use, such as to cope with negative emotions (i.e., coping motives) or anxiety-provoking social situations (i.e., conformity motives). For example, NU is directly associated with negative-reinforcement drinking motives (Curcio and George 2011) as well as indirectly associated with alcohol dependence through those motives (Coskunpinar and Cyders 2012; Littlefield et al. 2010), suggesting NU promotes problematic alcohol use via the promotion of specific drinking motives.

Using alcohol as a means to avoid negative emotions may impede the ability to develop adaptive coping mechanisms and contribute to the belief that one is not able to tolerate negative emotional states, i.e., distress intolerance (DI; Leyro et al. 2010). For example, individuals higher in NU may perceive that they cannot tolerate distress and may execute behaviors aimed at immediately terminating distress (McHugh and Otto 2012), which in turn strengthens negative-reinforcement alcohol use motives. In support of this, DI is positively associated with drinking motives to alleviate distress (Howell et al. 2010) and adverse alcohol-related consequences (Dennhardt and Murphy 2011; Holzhauer et al. 2017) among college students and clinical populations (Berking et al. 2011).

While concurrent, unique, or synergistic relations between NU and DI are supported (Cougle et al. 2012; Kaiser et al. 2012; Kelly et al. 2014; Peterson et al. 2014), NU, the behavioral tendency to escape distress, may promote DI, the perceived inability to tolerate distress, as proposed by the theoretical framework of a bottom-up approach (Magidson et al. 2014). A bottom-up model suggests that traits, such as DI, are malleable and may change through preceding behavior change, as opposed to a top-down approach that suggests cognitions alter behavior. In support of this framework, behavior-focused interventions have been shown to change the level of cognitive-affective traits (Chapman et al. 2014; Magidson et al. 2014). Therefore, NU may be sequentially related to the development of DI and specific drinking motives that promote greater alcohol consumption and subsequent alcohol use consequences (Greenberg et al. 2016; Marshall-Berenz et al. 2011). Moreover, given that drinking motives to reduce negative affect (i.e., coping and conformity motives) significantly contribute to both heavy drinking and adverse behavioral consequences among college-aged population (Carey and Correia 1997; Cooper et al. 1995; Merrill et al. 2014), it has become increasingly important to examine factors that promote the development of these motives. Identification of these factors may help to develop a comprehensive model of alcohol use and related consequences.

Together, college students who are higher in their behavioral tendency to respond impulsively to negative emotional states may become less tolerant of distress, via avoidance. It may then amplify the affect modulation of alcohol use (Leventhal and Zvolensky 2015), thereby promoting coping and conformity motives for use, and greater alcohol-related negative consequences. Extant studies have examined the impact of either NU or DI on alcohol use motives and consequences in college samples reporting past month (Coskunpinar and Cyders 2012), 3-month (Shishido et al. 2013), 12-month alcohol use (Curcio and George 2011), or at least one prior heavy drinking episode (Dennhardt and Murphy 2011), but have not examined their sequential effects. In order to examine our model in a sample of college drinkers, utilizing a sequential multiple mediation framework, the current investigation seeks to elucidate the impact of linear relations between NU, DI, and negative reinforcement-based drinking motives on adverse alcohol use outcomes within a college sample reporting past month alcohol use. Specifically, we hypothesize that the relation between NU and problematic alcohol use, i.e., past month/year negative consequences due to alcohol and alcohol dependence, will be explained by the sequential mediators of DI and coping/conformity motives.

Method

Participants

All participants were enrolled in an introductory psychology course and were recruited through the psychology department’s undergraduate participant pool, which included a variety of questions to help identify participants who were eligible for the current study. Study inclusion criteria were as follows: (1) having had at least one standard alcoholic beverage in the past one month (i.e., moderate alcohol drinkers; US Department of Health and Human Services 2017) and (2) between the ages of 18–25. Eligible participants interested in the current study signed up for participation via an online portal and were directed to complete informed consent and a questionnaire battery via Qualtrics software (Qualtrics, Provo, UT). A total of n = 637 were recruited. Our final n = 616 (M_age = 19.1, SD = 1.4, range = 18–25; 50.6% female) was determined after removing participants with either aberrant (n = 4)¹ or incomplete responses (n = 17). Participants received one-course research participation credit for their participation. The study procedures were approved by the Institutional Review Board at the study site.

Measures

Variables of Sample Characteristics

Alcohol History Questionnaire (AHQ; Filbey et al. 2008) is a 42-item self-report measure that assesses alcohol consumption quantity and frequency. In the current study, the AHQ was used as a descriptive measure of alcohol use history.

The Timeline Followback (TLFB; Sobell and Sobell 1992, 1995) is a well-validated method of assessing substance use (Robinson et al. 2014). In the current study, participants were asked to retrospectively estimate their daily alcohol use in the past 30 days (Sobell et al. 1996). During assessment completion, participants were provided with standard alcohol drink definitions (i.e., 12 oz. beer, 5 oz. glass of wine, and 1.5 oz. of 80-proof hard liquor; National Institute on Alcohol Abuse and Alcoholism 2003). The TLFB has demonstrated good test-retest reliability (Sobell et al. 1996).

Proposed Independent Variables (NU) and Mediators (DI, Coping, and Conformity)

The 4-item NU subscale from the short version of Short Urgency, Premeditation, Perseverance, and Sensation Seeking Impulsive Behavior was used to assess NU (SUPPS-P; Cyders et al. 2014) is a 20-item self-report measure that assesses five facets of trait impulsivity, such as how rashly individuals react in response to negative moods. Participants respond using a 4-point Likert-type scale. The SUPPS-P demonstrates adequate reliability as well as adequate convergent and discriminant validity (Cyders et al. 2014). The current study utilized the four items comprising the negative urgency (NU) subscale. Cronbach’s alpha for the NU subscale in the current sample was 0.77.

Distress Intolerance Index (DII; McHugh and Otto 2012) is a 10-item self-report measure that assesses individual differences in perceived capacity to withstand and tolerate general somatic and psychological distress. Participants rate items on a 5-point Likert-type scale. Previous research indicates the DII is strongly associated with both behavioral and self-report indices of DI (McHugh and Otto 2011). Cronbach’s alpha in the current sample was 0.92.

Drinking Motives Questionnaire (DMQ; Cooper 1994) is a 20-item self-report measure assessing social, enhancement, coping, and conformity alcohol drinking motives. Participants rate items on a 5-point Likert-type scale. The DMQ has demonstrated good validity and reliability among both adolescents (Kuntsche et al. 2008) and adults (Németh et al. 2011). The current study utilized 10 items corresponding with the conformity and coping subscales. Cronbach’s alpha in the current sample was 0.87 and 0.85 for coping and conformity drinking motives, respectively.

Criterion Variables

Brief Young Adult Alcohol Consequences Questionnaire (YAACQ; Kahler et al. 2005) is a 24-item self-report measure assessing both general consequences (e.g., blackouts) and college-age-specific consequences (e.g., schoolwork) related to alcohol use. Participants respond to items dichotomously (yes/no), based on whether or not they have experienced certain problems in the past month. The YAACQ has demonstrated strong psychometric properties including internal consistency and predictive validity (Read et al. 2007). Cronbach’s alpha in the current sample was 0.91.

Rutgers Alcohol Problem Index² (RAPI; White and Labouvie 1989) is a 23-item measure assessing negative consequences of drinking during the past year targeting an adolescent and young adult population. Each item is rated on a 5-point Likert scale. The RAPI has demonstrated good internal consistency and moderate concurrent validity (White and Labouvie 1989). Cronbach’s alpha in the current sample was 0.94.

Alcohol Use Disorders Identification Test (AUDIT; Saunders et al. 1993) is a 10-item measure assessing a range of hazardous alcohol use. Participants respond to items regarding the frequency and effects of their drinking on psychosocial functioning. Items are rated on 5-point Likert-type scales. In the current study, the level of alcohol dependence was assessed by the total score of the AUDIT. The psychometric properties of this scale are well established and standardized for use in both clinical and research settings (Babor et al. 2001). Cronbach’s alpha in the current sample was 0.79. The observed internal consistency from the current sample is comparable to those reported in previous studies among college students (Carretero et al. 2016) and community population (Rumpf et al. 2002).

Data Analysis

Zero-order correlations were first conducted to examine theoretical associations among predictor (NU), mediators (DII and coping/conformity motives), criterion variables (i.e., YACCQ, RAPI, and AUDIT total score), and covariates (age and sex). Next, a series of sequential multiple mediation analyses were conducted to examine the direct and indirect effect of NU on the criterion variables using PROCESS for SPSS 24.0 to test for both direct and sequential indirect effects (Hayes 2013). PROCESS allows for examination of two mediators, in either a causal or sequential way, by simultaneously testing the indirect effects of each mediator independently (Hayes 2013). In the current study, specific indirect effects of DI (a₁*b₁) and either coping or conformity motives (a₂*b₂) were estimated (Fig. 1). Moreover, the specific indirect effect of DI and each criterion variable (a₁*a₃*b₂) was estimated in order to test the significance of the sequential mediation effect. All models included NU as a predictor and DI as the first mediator (M1), and the criterion variables (Fig. 1). Coping motives was entered as the second sequential mediator (M2-1) in the first set of analyses, and conformity motives (M2-2) was entered as the second sequential mediator in the second set of analyses (Table 3). In each model, age and sex were included as covariates, due to previously observed relations with outcomes of interest and theoretical relevance. Completely standardized indirect effects of the predictor on criterion variables were calculated for each mediation model as an index of effect size (ES; Preacher and Hayes 2008a; Preacher and Kelley 2011). The 95-percentile bias-corrected confidence intervals (BC 95% CI) estimated with bootstrap analyses of 10,000 samples was used for all study models to determine the statistical significance (Hayes 2009; Hayes and Scharkow 2013; Preacher and Hayes 2004, 2008b).

Fig. 1.

Sequential multiple mediation model. AUDIT-TOT, alcohol use disorders identification test total score; NU, negative urgency; DII, distress intolerance index; YAACQ, young adult alcohol consequences questionnaire; RAPI, Rutgers alcohol problem index

Results

Participants

Sample characteristics are shown in Table 1. On average, the sample met for moderate alcohol problems (M = 10.5, SD = 4.8, range = 1–29) as indexed by AUDIT total score (Babor et al. 2001). According to the TLFB, the mean number of days with any alcohol use in the past month was 7.2 days (SD = 5.0). A majority of the sample reported binge drinking (i.e., 76.3%; ≥ 4 drinks, females; ≥ 5 drinks, males, on one occasion; National Institute on Alcohol Abuse and Alcoholism 2015) in the past month. The mean age of regular drinking was 16.3 years old (SD = 4.6), and the sample reported regular drinking for an average of 2.2 years (SD = 2.8).

Table 1.

Sample characteristics

Descriptive summary	Total N = 616
Race n (%)
Caucasian	373 (60.6%)
African American	26 (4.2%)
Asian	141 (22.9%)
Other	15 (2.4%)
Unknown	8 (1.3%)
More than one race	53 (8.6%)
Ethnicity n (%)
Hispanic	80 (13.0%)
Non-Hispanic	518 (84.1%)
Unknown	18 (2.9%)
College year n (%)
Freshman	352 (57.1%)
Sophomore	134 (21.8%)
Junior	69 (11.2%)
Senior	53 (8.6%)
Other	8 (1.3%)
Drinking history M (SD)
Age of onset of regular use	16.3 (4.6)
Average # drinks at one occasiona	4.8 (2.7)
Largest # drinks at one occasiona	8.3 (7.8)
Average # days/week with drinkinga	1.9 (1.2)
Years of regular use	2.2 (2.8)

M, mean; SD, Standard deviation

^aPast month use

Zero-Order Correlations

Study predictors, proposed mediators, and criterion variables were normally distributed according to West et al. (1995). As shown in Table 2, zero-order correlations revealed NU was significantly correlated with each proposed mediator (DI, coping, and conformity motives, all p’s < .01) and criterion variables (past month/year alcohol use problems and AUDIT total score, all p’s < .01). The same correlation pattern with criterion variables was observed for the proposed mediators, as well as coping and conformity motives. The proposed mediator, DI, was significantly correlated with each criterion variable (all ps < .01), with the exception of the AUDIT total score. Lastly, older individuals reported greater past month/year alcohol use consequences compared to younger individuals (all ps < .05). Females reported greater levels of DI as compared to males (p < .01) while males reported greater levels of conformity motives, past year alcohol use consequences, and alcohol dependence (all ps < .01).

Table 2.

Zero-order correlations for study variables (N = 616)

	1	2	3	4	5	6	7	8	9	M (SD)
1. Agea	1	− .16**	− .02	.00	.01	.04	.09*	.15**	.06	19.1 (1.4)
2. Sexb		1	.06	.22**	.04	− .14**	−.06	− .12**	− .14**	50.6%c
3. NU			1	.40**	.33**	.22**	.33**	.27**	.24**	10.2 (2.8)
4. DII total				1	.35**	.21**	.18**	.16**	.07	25.4 (8.7)
5. Coping					1	.48**	.34**	.35**	.36**	11.5 (5.0)
6. Conformity						1	.34**	.40**	.29**	8.8 (4.2)
7. YAACQ							1	.59**	.63**	6.4 (5.7)
8. RAPI								1	.62**	31.1 (10.2)
9. AUDIT-TOT									1	10.5 (4.8)

^*p < .05

^**p < .01, two-tailed. M, mean; SD, standard deviation; sex was coded 1 for males and 2 for females; AUDIT-TOT, alcohol use disorders identification test total score; DII, distress intolerance index (high score means high in DI); NU, negative urgency; YAACQ, young adult alcohol consequences questionnaire; RAPI, Rutgers alcohol problem index

^aMissing one participant

^bSpearman correlation coefficients

^cFemales

Sequential Multiple Mediation Analysis

Results from the planned path models examining the indirect effect of NU on each criterion variable (i.e., YAACQ, RAPI, and AUDIT total score), through DII (M1) and either coping (M2–1) or conformity motives (M2–2) can be found in Table 3. Including sex and age as covariates, both the total effect and full models examining the indirect effect of NU on criterion variables through DI and drinking motives (i.e., coping and conformity) were significant (Table 3).

Table 3.

Results of sequential multiple mediation analysis

Y	Patha	B	SE	t	CI (lower)	CI (upper)	Total effect model	Full model with mediators
	a1	1.174	.111	10.564***	.956	1.393
M2–1 = copingb
	a2	.410	.071	5.769***	.270	.550
	a3	.149	.024	6.249***	.102	.195
YAACQ	b1	.010	.027	.366	− .044	.064	R2 = .13 F (3, 612) = 31.06***	R2 = .19 F (5, 610) = 28.75***
	b2	.288	.045	6.344***	.199	.377
	c′	.500	.082	6.106***	.339	.660
	c	.679	.076	8.970***	.531	.828
	a1*b1	.012	.034		− .055	.080
	a2*b2	.118	.029		.064	.181
	a1*a3*b2	.050	.013		.027	.077
RAPI	b1	.017	.049	.348	− .079	.113	R2 = .12 F (3, 612) = 26.98***	R2 = .19 F (5, 610) = 29.10***
	b2	.584	.081	7.245***	.425	.742
	c′	.658	.145	4.527***	.372	.943
	c	1.019	.136	7.497***	.752	1.286
	a1*b1	.020	.063		− .105	.141
	a2*b2	.239	.053		.142	.349
	a1*a3*b2	.102	.026		.056	.158
AUDIT-TOT	b1	− .048	.023	− 2.048*	− .094	− .002	R2 = .08 F (3, 612) = 18.72***	R2 = .18 F (5, 610) = 26.91***
	b2	.328	.039	8.484***	.252	.404
	c′	.287	.070	4.111***	.150	.423
	c	.422	.066	6.400***	.292	.551
	a1*b1	− .056	.028		− .113	− .002
	a2*b2	.134	.029		.081	.193
	a1*a3*b2	.057	.013		.034	.086
M2-2 = conformityb
	a2	.237	.062	3.829***	.115	.358
	a3	.086	.021	4.153***	.045	.126
YAACQ	b1	.022	.027	.804	− .031	.074	R2 = .13 F(3, 612) = 31.06***	R2 = .20 F(5, 610) = 30.85***
	b2	.364	.052	7.020***	.262	.465
	c′	.531	.080	6.638***	.374	.689
	c	.679	.076	8.970***	.531	.828
	a1*b1	.025	.033		− .039	.090
	a2*b2	.086	.027		.038	.145
	a1*a3*b2	.037	.012		.016	.063
RAPI	b1	.034	.047	.730	− .058	.127	R2 = .12 F(3, 612) = 26.98***	R2 = .22 F(5, 610) = 35.16***
	b2	.808	.091	8.889***	.629	.986
	c′	.706	.140	5.028***	.430	.982
	c	1.019	.136	7.497***	.752	1.286
	a1*b1	.040	.060		− .080	.160
	a2*b2	.191	.053		.091	.298
	a1*a3*b2	.081	.027		.035	.142
AUDIT-TOT	b1	− .023	.024	− .979	− .070	.023	R2 = .08 F(3, 612) = 18.72***	R2 = .14 F(5, 610) = 19.31***
	b2	.278	.046	6.096***	.189	.368
	c′	.355	.071	5.036***	.217	.494
	c	.422	.066	6.400***	.292	.551
	a1*b1	− .027	.028		− .084	.024
	a2*b2	.066	.021		.029	.110
	a1*a3*b2	.028	.010		.012	.049

^*p < .05

^**p < .01

^***p < .001 two-tailed; sex and age were entered as covariates in all study models; M1 and M2s, proposed mediators; a₁ and a₂, pathways from the predictor to each of M1 and M2, respectively; b₁ and b₂, pathways from each of M1 and M2 to the criterion variables, respectively; a₃, pathways from M1 to M2s; a₁*b₁, and a₂*b₂, specific indirect effects through M1 and M2, respectively; a₁*a₃*b₂, specific indirect effect via both M1 and M2. The 95% CIs for indirect effects were obtained by bootstrapping with 10,000 resamples. Italic texts indicate significant sequential indirect effects. AUDIT-TOT, alcohol use disorders identification test total score; YAACQ, young adult alcohol consequences questionnaire; RAPI, Rutgers alcohol problem index

^aPath a₁ does not change across models

^bPaths a₂ and a₃ do not change across two models

With regard to the results from the three models including coping motives as the second mediator (M2-1), both the total effect (path c) and the direct effect (path c′) of NU in relation to alcohol use consequences in the past month/year and alcohol dependence, controlling for the mediators, were significant. A significant sequential indirect effect of NU via DI (M1) and coping motives (M2-1) was found for each criterion variable. Specifically, NU was predictive of greater negative consequences in both past month (YACCQ; B = .050, SE = .013, BC 95% CI .027 to .077) and year (RAPI; B = .102, SE = .026, BC 95% CI .056 to .158), as well as AUDIT total scores (B = .057, SE = .013, BC 95% CI .034 to .086). The size of the indirect effect was small to medium (ES with YACCQ = .025, SE = .006, BC 95% CI .014 to .039; ES with RAPI = .029, SE = .007, BC 95% CI .016 to .045; ES with AUDIT total = .034, SE = .008, BC 95% CI .020 to .051).

Regarding the results from the three models including conformity motives as the second mediator (M2-2), both the total effect (path c) and the direct effect (path c′) of NU in relation to alcohol use consequences in the past month/year and alcohol dependence, controlling for the mediators, were significant. A significant sequential indirect effect of NU via DI (M1) and conformity motives (M2-2) was found for each criterion variable. Specifically, NU was predictive of greater negative consequences in both past month (YACCQ; B = .037, SE = .012, BC 95% CI .016 to .063) and year (RAPI; B = .081, SE = .027, BC 95% CI .035 to .142), as well as AUDIT total scores (B = .028, SE = .010, BC 95% CI .012 to .049). The size of the indirect effect was small to medium (ES with YACCQ = .018, SE = .006, BC 95% CI .008 to .032; RAPI = .023, SE = .008, BC 95% CI .010 to .040; AUDIT total = .017, SE = .006, BC 95% CI .007 to .029).

Discussion

The current study sought to examine the extent to which the behavioral tendency to act impulsively in the context of affective distress contributes to the belief that distress is intolerable, thereby promoting specific alcohol use motives and problematic alcohol use in young adults. Our sequential multiple mediation analyses demonstrate that NU is indirectly associated with both problematic alcohol use (i.e., AUDIT total score) and adverse consequences in the past month/year (i.e., YACCQ/RAPI) through DI (M1) and coping (M2-1)/conformity motives (M2-2). These results extend previous work indicating a synergistic relation between DI and NU in relation to alcohol use motives and outcomes by suggesting that NU potentiates the development of DI, thereby contributing to coping motives that increase alcohol-related adverse consequences. This finding promotes the conceptualization of a comprehensive model whereby the tendency to immediately respond to intense emotions (NU) may impede the ability to learn skills to tolerate distress. This may subsequently contribute to the perceived inability to tolerate distress (DI) that may then promote behavior aimed at the immediate relief of distress resulting in poor long-term outcomes. This model (Fig. 1) builds upon extant work supporting the relation between drinking motives and alcohol-related negative consequences (Coskunpinar et al. 2013; Curcio and George 2011; Dir et al. 2013; Murphy and MacKillop 2012) by identifying DI as a possible mechanism through which NU is associated with poor alcohol use outcomes. Moreover, these findings are aligned with the perspective that distinct cognitive and behavioral vulnerabilities may not only act independently, but sequentially (Borges et al. 2017; Kaiser et al. 2012).

We found a small to medium effect for the mediating role of conformity motives, which has not been previously reported in the literature. Although extant research supports the association between coping motives and heavy drinking (Carey and Correia 1997; Merrill et al. 2014), and adverse long-term consequences (Kuntsche et al. 2005; Merrill et al. 2014), there is limited evidence linking conformity motives with such outcomes (Kuntsche et al. 2005). Extant research also indicates that DI may be significantly associated with coping motives to drink, but not conformity motives (Howell et al. 2010). However, in previous studies, conformity motives were only examined in the context of alcohol use and consequences (e.g., Merrill et al. 2014) without accounting for cognitive-affective vulnerabilities. Moreover, conformity motives were sometimes not even considered in study models (Curcio and George 2011). To the best of our knowledge, the current study is the first study that provides novel support for the association between NU, DI, and coping/conformity motives as well as the role of intrapersonal and interpersonal factors in problematic alcohol use and consequences in a large diverse college-aged sample.

Several study limitations warrant discussion. First, the cross-sectional and nonexperimental design of the current study limits our ability to make causal inferences. Prospective work would allow us to more carefully examine whether our proposed model holds when examined temporally (Rindfleisch et al. 2008). Second, the utilization of a single self-report index of DI, that is considered a domain-general measure of DI, is a limitation given extant research demonstrates a differential relation between subtypes of DI (e.g., frustration or pain tolerance) and domain-general DI indices (McHugh and Otto 2011). Third, the current findings may have limited generalizability to other populations given the sample (i.e., college students) chosen to examine the proposed model. Given extant findings on the relation between DI, NU, alcohol use motives, and negative consequences in other populations (e.g., Berking and Wupperman 2012; Coskunpinar et al. 2013), it is possible that the current model can be applied to the more general population to inform prevention and treatment efforts. One strength of the study lies in the diversity of the current sample compared to the sample of previous studies, which increases the generalizability of the current findings to other college-aged samples. Specifically, our sample was racially and ethnically diverse and 42.9% of our sample was upperclassmen, in contrast to extant work (e.g., > 80% freshman, Curcio and George 2011; < 5% Asian, Dir et al. 2013). A second strength of the current study is the observed effect sizes with adequately powered sample size (Fritz et al. 2015) and the utilization of bootstrapping that minimizes the contribution of measurement error in interpreting the sequential indirect effect (Shrout and Bolger 2002).

Despite these limitations, the study recruited a large diverse sample of young adults, who reported a range of alcohol consumption and alcohol-related adverse consequences, and the observed relations may be representative of pathways to alcohol use in the general young adult population. The current study demonstrated that cognitive vulnerabilities, such as the perception that distress is intolerable, and beliefs that alcohol may serve to reduce distress mediate the established relation between individual differences in behavioral tendency, NU, and alcohol-related outcomes implicated in the development of alcohol use disorders in young adults. The results demonstrate that behavioral traits may affect cognitions and beliefs to promote drinking and negative consequences in college students. These findings provide a preliminary framework for targeting the general behavioral tendency to act rashly in the context of distress that offer potential inroads to reduce drinking in this population through interventions that improve skills for regulating affect, such as distress tolerance or acceptance skills (Bornovalova et al. 2012).