Emotion regulation strategies moderate the impact of negative affect induction on alcohol craving in college drinkers: an experimental paradigm

Abstract

Objectives:

Observational studies suggest emotion regulation (ER) as a potential treatment target for problematic college drinking. The primary aim of this laboratory study was to determine whether trait ER strategies would moderate the impact of negative affect induction on alcohol craving in college drinkers.

Methods:

Participants were randomly assigned to a neutral (n = 74) or a negative affect induction (n = 76) and reported their craving after the affect inductions.

Results:

Greater use of drinking to cope and less use of cognitive reappraisal predicted greater alcohol craving after the negative affect induction, but not after the neutral condition. In contrast, emotion suppression did not predict alcohol craving in either condition.

Conclusion:

Our results highlight the role of ER tendencies — particularly the benefits of cognitive reappraisal — on alcohol craving when experiencing emotional distress. Therefore, ER strategies may be an important target for college drinkers to prevent and reduce problematic drinking.

Introduction

Alcohol use and binge drinking are common in college students. A 2019 national survey revealed that 52.5% of college students endorsed alcohol use and 33.0% report binge drinking episodes in the past month.¹ The National Institute on Alcohol Abuse and Alcoholism (NIAAA) defines binge drinking as consuming 4(female)/5(male) standard drinks within 2 hours.² Problematic college drinking leads to academic problems as well as physical and mental health problems such as missing class, poor grades, alcohol-related unintentional death, physical or sexual assaults, and depression and anxiety disorders.^3–6 To avoid these negative consequences, clinicians offering health services to college students need to routinely screen for problematic drinking and provide treatments to those who screened positive. To date, effective treatments targeting problematic college drinking are available to reduce alcohol use, with small to medium effect sizes.^7–10 To enhance treatment effectiveness, treatment targets for problematic college drinking need to be further investigated.

Emotion regulation (ER) theory describes the possible link between negative emotions and problematic drinking.¹¹ Alcohol, as a mood altering substance, can motivate people to drink to induce positive emotions (positive reinforcement) or reduce negative emotions (negative reinforcement).¹¹ In line with ER theory, studies on college drinkers have revealed that negative emotions are positively related to alcohol-related problems,^{12, 13} and that stress- or negative affect leads to more drinking^14–16 and alcohol craving.¹⁷ Therefore, adaptive ER strategies may lessen craving in college drinkers undergoing emotionally disturbing situations, and subsequently this may reduce and prevent problematic drinking. In fact, cumulative evidence from cross-sectional observational studies suggests that difficulty regulating negative emotions is associated with more problematic drinking in college students.^18–21 Yet, these cross-sectional studies have not examined whether ER moderates the impact of negative emotion-evoking situations on alcohol craving and problematic drinking.

It should be noted that not all ER strategies are adaptive. Among college students, for example, drinking to cope with life stressors is associated with more heavy drinking and more depressive symptoms even with the intent to reduce negative emotions.²² Drinking is a common coping strategy among college students. About 40.7% to 53.4% of college students report alcohol use to cope with stressors.^{22, 23} In undergraduates, more use of cognitive reappraisal is related to less alcohol-related problematic behaviors, but suppression of emotion is unrelated to alcohol-related problematic behaviors (e.g., arguing or fighting under the influence of alcohol).²⁴ Reappraisal refers to changes in appraisal or interpretation of a situation to reduce negative emotional impacts (e.g., thinking about a challenging situation in a way that helps you stay calm), and emotion suppression refers to inhibition of behavioral expression of a particular emotion (e.g., regulating emotions by not expressing them).²⁴ Beyond the context of alcohol use, greater use of reappraisal is associated with better psychosocial health, whereas greater use of suppression is associated with worse psychosocial health.²⁵ In a laboratory study that compared the effectiveness of reappraisal, suppression, and acceptance strategies on stress-induced craving for nicotine in college smokers, cognitive reappraisal has been shown to be superior in reducing negative emotions and nicotine craving.²⁶

To further elucidate the role of ER strategies in college drinking, the current study examined ER strategies as a moderator for the impact of a negative emotion-evoking condition on alcohol craving in college students. We examined craving following negative affect induction as an outcome variable because laboratory studies have demonstrated that stronger craving leads to more alcohol use in college students²⁷ and adults.^{28, 29} In the context of alcohol use disorder, craving is a strong desire for alcohol in its absence and it has been conceptualized to play a key role in binge drinking, especially during the later stage of alcohol use disorder to reduce anxiety and stress.³⁰ As we recruited college students drinking any amount at any frequency, our study was unique in investigating the role of ER strategy in alcohol craving among typical college drinkers.

Our aim was to examine whether trait ER strategies would moderate the effect of negative affect induction on alcohol craving. Specifically, we examined the following three ER strategies: drinking to cope, cognitive reappraisal, and emotion suppression. We hypothesized that a greater tendency to use drinking to cope and emotion suppression as maladaptive strategies would increase the impact of negative affect induction on alcohol craving and that a greater tendency to use cognitive reappraisal as an adaptive strategy would reduce the impact of negative affect induction on alcohol craving.

METHOD

Texas A&M Institutional Review Board approved the study procedures and informed consent was obtained from all participants. Undergraduates who were at least 18 years old, drank alcohol at any frequency or amount, and either sex were invited to participate. There were no exclusion criteria.

Procedures

This study was advertised through the university’s online recruitment system. Participants read the brief introduction about the study and the eligibility criteria (i.e., at least 18 years and currently drinking alcohol), and then, they voluntarily signed up for the study or cancelled it without penalty. On the day of laboratory testing, an experimenter explained the study in person during the informed consent process. Then, all study procedures were executed via a computer (see the timeline in Figure 1). After completing the baseline questionnaires, participants were randomized to either a neutral or negative affect condition,³¹ and completed post-questionnaires. All participants were debriefed in-person by an experimenter and they received course credit. The whole experiment took on average 32.3 minutes (SD = 11.5).

Figure 1.

Experimental Timeline (30 minutes)

ERQ: Emotion Regulation Questionnaire; COPE-Alcohol - 2 alcohol items in Brief COPE; DASS: Depression Anxiety Stress Scale; PASC: Penn Alcohol Craving Scale; AUDIT: Alcohol Use Disorders Identification Test; ACQ-SF: Alcohol Craving Questionnaire-Short Form-Revised; DDQ: Daily Drinking Questionnaire.

The Velten Mood Induction

The Velten Mood Induction protocol was used to induce negative affect.³¹ This protocol begins with a standardized script on the computer screen: You will now read a series of sentences and each will appear one at a time. After reading each sentence, you can click the next button to continue. Then, participants in the negative affect condition read the 60 self-referent depressed statements (e.g., I have too many bad things in my life.). Participants in the control condition read the 60 neutral-fact statements (e.g., Saturn is sometimes in conjunction, beyond the sun from the earth, and not visible.). The mood induction procedure was conducted in a private room using a computer without additional instructions from the experimenter. The Velten protocol has been widely used to alter mood temporarily in a laboratory setting by reading and processing the set of mood-altering statements and has been used in adults with drinking problems.³² More relevant to the current study, the Velten protocol has been shown to increase negative mood and alcohol craving in college students.³³

Measures

Measures assessing emotion regulation strategies

The Emotion Regulation Questionnaire (ERQ) was administered to assess ER strategies.³⁴ The ERQ consists of the 6-item cognitive reappraisal (α = .79) and the 4-item expression suppression (α = .73). Each item is rated on a 7-point Likert scale. Higher summed scores on each scale indicate higher use of the strategy.³⁴ Of the four substance use items of Brief COPE measure (α = .90),³⁵ we averaged the two alcohol specific items to assess levels of drinking to cope.²²

Measures assessing psychological health status

The 21-item Depression Anxiety Stress Scale (DASS) was administered to assess participants’ depression (α = .94), anxiety (α = .87), and stress (α = .91) symptoms for the past week.^{36, 37} Each item was rated on a 4-point Likert scale. The total summed scores of each symptom ranged from 0 to 42, with higher scores on each scale indicating greater symptoms of depression, anxiety, and stress. The respective cutoff scores for mild, moderate, and severe symptoms are 10, 14, and 21 for depression, 8, 10, and 15 for anxiety, and 15, 19, and 26 for stress.³⁶

Measures assessing drinking behaviors

The 5-item Penn Alcohol Craving Scale (PACS) was administered to assess craving for the past week (α = .92).³⁸ Each item was rated on a 7-point Likert scale and total summed scores ranged from 0 to 30, with higher total scores indicating greater alcohol craving for the past week. The PACS total scores were entered as a covariate to examine the effect of negative affect on alcohol craving at the time of the experiment in adjusting for pre-existing individual differences in alcohol craving. The Daily Drinking Questionnaire (DDQ) was administered to assess drinking patterns for the past 30 days.³⁹ On the DDQ, participants were asked to report the number of drinks for the typical drinking week, and hours spent drinking. Based on the responses on the DDQ and the NIAAA definition,² we estimated the prevalence of binge drinking in our sample. Participants were also asked to report their last drinking episode, which was also entered as a covariate to adjust for its potential impact on craving at the time of experiment.⁴⁰ The 10-item Alcohol Use Disorders Identification Test (AUDIT) was administered to assess problematic alcohol use (α = .91).⁴¹ Each item was rated on a 0–4 scale score, with higher total summed scores indicating more problematic patterns of alcohol consumption.

Measures assessing emotional states and alcohol-related states at the time of experiment

As a part of the Velten Mood Induction protocol,³¹ the following four items were administered after the affect induction. Participants were asked to rate the four items, ‘How happy are you?’ ‘How sad are you?’ ‘How depressed are you?’ and ‘How bored are you?’ on a 0–100 Visual Analogue Scale (VAS). The higher scores on each scale indicate higher levels of happiness, sadness, depression, and boredom at the time of questionnaire. The 12-item Alcohol Craving Questionnaire-Short Form-Revised (ACQ-SF)⁴² was administered after the mood induction to assess levels of craving at the time of completion (α = .85).⁴³ Each item was rated on a 7-Likert Scale and total summed scores were averaged to calculate individual score for alcohol craving.

Sample Size and Statistical Analysis

Over two semesters (Spring and Fall of 2016), the current study recruited 153 undergraduates and analyzed 150 participant’s data, which were sufficient to detect a moderate sized effect for regression analysis with 80% power and p value of .05.⁴⁴ Three participant’s data were excluded for analysis because one participant did not complete 25% of the questionnaires and two participants noted as not drinking and not applicable on the DDQ. There was no missing data. First, appropriate chi-square or independent t tests were conducted to examine whether participants in the two randomized conditions were significantly different in demographic characteristics, psychological variables, and drinking behaviors. Second, a MANOVA was conducted to examine whether the emotional states (happiness, sadness, depression, and boredom) would be different after affect induction. Next, three hierarchical multiple linear regression analyses were conducted to examine whether ER strategies would moderate the impact of negative affect induction on alcohol craving after adjusting for pre-existing differences in alcohol craving and sex (female coded as 0, male coded as 1). The neutral condition was coded as zero and negative affect induction condition was coded as 1. All the other variables in the regression analysis were z transformed. The Variance Inflation Factor (VIF) was examined for multicollinearity. All VIF values were below 10 (i.e., < 2.3), suggesting no significant violation of multicollinearity.⁴⁵ Then, we examined whether the condition by ER interaction term would be a significant predictor for acute alcohol craving. When we found a significant condition by ER interaction term, we conducted a post-hoc simple linear regression to further examine a specific ER strategy as a predictor for alcohol craving in each condition. All regression models were significant at p value < .05. To reduce Type 1 error for multiple analyses, p values were adjusted using FDR⁴⁶ (‘car’ package⁴⁷). Finally, Pearson correlation coefficients were calculated to examine the bivariate relationships among drinking-related variables, ER strategies, and psychological factors.

RESULTS

Participants

Table 1 shows the baseline demographic characteristics and drinking patterns of our sample. Our participants were young adults (M = 19.0, SD = 1.3, 18–28 years), more male (65.3%), and White/Caucasian (65.3%). All participants endorsed their sex as male or female as they were asked to select between the two. We want to note that our sample’s reported sexual identity might be different if our study used a non-binary classification. Our sample included 36.7% binge drinkers and participant’s last drinking episode occurred on average 12.9 days ago (SD = 21.6). These demographic characteristics and drinking patterns were not significantly different between the two conditions (ps > .194).

Table 1.

Demographics

			Condition
		Whole Sample	Neutral	Negative Affect
		n = 150	n = 74	n =76
		n (%)	n (%)	n (%)	χ2
Sex	Female	52 (34.7)	22 (29.7)	30 (39.5)	1.57
	Male	98 (65.3)	52 (70.3)	46 (60.5)
Race/Ethnicity	White/Caucasian	98 (65.1)	47 (63.5)	51 (67.1)	0.73
	The others	52 (34.9)	27 (36.5)	25 (32.9)
	(Hispanic)	20 (13.3)	9 (12.2)	11 (14.5)
	(African American)	8 (5.3)	6 (8.1)	2 (2.6)
	(Asian)	12 (8.0)	6 (8.1)	6 (7.9)
	(Other or bi-racial)	12 (8.0)	6 (4.0)	6 (7.9)
Drinking patterns	Binge drinkers	55 (36.7)	28 (37.8)	27 (35.5)	0.09
		M (SD)	M (SD)	M (SD)	t
Age		19.0 (1.3)	18.9 (1.3)	19.1 (1.3)	−0.94
Number of standard drinks (DDQ)
On typical drinking day		3.8 (3.4)	4.2 (3.7)	3.5 (3.2)	1.15
Days after the last drinking		12.9 (21.6)	10.6 (14.1)	15.2 (26.8)	1.70

Note: None of the χ² and independent t tests were significant at p value at .05; DDQ: Daily Drinking Questionnaire.

Table 2 summarizes our samples’ baseline characteristics of ER strategies, psychological variables, and drinking-related variables. All characteristics were not significantly different between the two conditions except PASC scores. On the PASC, the participants who were randomized to a neutral condition reported significantly higher levels of alcohol craving for the past week than the participants who were randomized to a negative affect condition, p = .041. However, AUDIT scores were not significantly different between the two conditions, p = .675.

Table 2.

Baseline characteristics of emotion regulation strategies, psychological variables, and drinking-related variables.

			Condition
		Whole Sample	Neutral	Negative Affect
		n = 150	n = 74	n =76
		M (SD)	M (SD)	M (SD)	t
COPE-	Alcohol	1.4 (0.5)	1.4 (0.6)	1.3 (0.5)	0.57
ERQ	Reappraisal	26.5 (7.2)	27.0 (7.0)	26.0 (7.4)	0.91
	Suppression	11.4 (5.8)	11.1 (5.6)	11.7 (6.0)	−0.58
DASS	Depression	9.8 (9.4)	10.5 (9.2)	9.2 (9.6)	0.83
	Anxiety	10.1 (8.7)	10.8 (8.0)	9.4 (9.3)	0.96
	Stress	15.0 (8.9)	15.8 (8.7)	14.2 (9.2)	1.05
AUDIT		6.9 (4.5)	7.2 (4.6)	6.7 (4.4)	0.67
PASC*		5.5 (4.4)	6.6 (4.2)	4.6 (4.4)	2.84

Note: None of the independent t tests in Table 2 were significant at p value at .05 except that PASC scores* were significantly different (FDR adjusted p value = .041).

COPE-Alcohol - 2 alcohol items in Brief COPE; ERQ: Emotion Regulation Questionnaire; DASS: Depression Anxiety Stress Scale; AUDIT: Alcohol Use Disorders Identification Test; PASC: Penn Alcohol Craving Scale.

Manipulation check

A MANOVA was conducted to compare the four emotional states after the affect induction. The result suggested that there was a significant group difference in emotional states, Wilks’ Λ = 0.89, p = .002. Post-hoc univariate analyses in Table 3 showed significant differences in sadness (p = .024), depression (p = .043), and boredom (p = .012), but no significant difference in happiness (p = .980). Specifically, participants in the negative affect induction condition reported higher levels of sadness and depression whereas participants in the neutral condition reported higher levels of boredom.

Table 3.

Emotional states after the Velten mood induction procedures.

	Condition
	Neutral	Negative Affect
	n = 74	n =76
(0–100 VAS)	M (SD)	M (SD)	F	p	Cohen’s d
Happiness	52.6 (23.1)	52.8 (26.0)	0.01	.980	0.00
Sadness	18.8 (20.0)	28.5 (26.3)	6.44	.024	0.42
Depression	13.2 (18.2)	21.4 (27.1)	4.68	.043	0.36
Boredom	53.1 (28.9)	38.5 (30.9)	8.90	.012	0.49

FDR adjusted p values were noted

Bivariate relationships among drinking-related variables, ER strategies, and psychological factors

Pearson correlation coefficients were computed to examine the bivariate relationships of ER strategies to drinking-related variables and psychological factors (Table 4). Higher COPE-alcohol scores were related to higher ACQ-SF, PASC, and AUDIT scores (r = 0.44 ~ 0.58, ps < .001), suggesting that drinking to cope was associated with more craving and more problematic drinking. In contrast, the two ERQ scores were unrelated to ACQ-SF, PASC, and AUDIT scores (ps > .395), suggesting that non-alcohol specific coping strategies did not have a direct relationship with alcohol craving and problematic drinking. DASS-depression, anxiety, and stress scores were positively correlated with COPE-alcohol and ERQ-suppression scores (rs = .18 ~ .31, ps < .031) and negatively correlated with ERQ- reappraisal scores (rs = −.19 ~ −.17, ps < .044). Our results suggest that more use of drinking to cope and emotion suppression was associated with worse psychological health status whereas more use of reappraisal was associated with better psychological health status.

Table 4.

Pearson correlation coefficients with drinking-related variables, emotion regulation strategies, and psychological factors. (n = 150)

		Alcohol-related variables		Emotion regulation strategies			Psychological factors
		b.	c.	d.	e.	f.	j	k.	l.
	a. ACQ-SF	.54***	.44***	.44***	−.07	.02	.22**	.18*	.15
	b. PASC		.55***	.58***	−.01	−.03	.17*	.19*	.15
	c. AUDIT			.52***	−.06	−.05	.09	−.02	−.06
	d. COPE-Alcohol				−.12	−.01	.31***	.19*	.20*
ERQ	e. Reappraisal					−.01	−.19*	−.17*	−.17*
	f. Suppression						.31*	.18*	.21**
DASS	j. Depression							.63***	.69***
	k. Anxiety								.68***
	l. Stress

Note:

^*p < .05

^**p < .01

^***p < .001.

ACQ-SF: Alcohol Craving Questionnaire-Short Form-Revised; PASC: Penn Alcohol Craving Scale; AUDIT: Alcohol Use Disorders Identification Test; COPE-Alcohol - 2 alcohol items in Brief COPE; ERQ: Emotion Regulation Questionnaire; DASS: Depression Anxiety Stress Scale.

ER skills as a moderator for the impact of negative affect induction on alcohol craving

Three hierarchical multiple linear regressions were conducted to determine whether drinking to cope, cognitive reappraisal, and emotion suppression would moderate the impact of negative affect induction on alcohol craving. First, sex, days after the last drinking episode, PASC, DASS-Depression, and DASS-Anxiety scores were entered in Level 1 (Table 5). Sex, days after the last drinking, depression, and anxiety scores were not significant predictors for ACQ-SF scores (p’s > .072), but higher PASC scores predicted greater ACQ-SF scores (β = 0.48, p < .001), suggesting that higher levels of pre-existing craving predicted craving at the time of experiment. When adjusting sex, days after the last drinking episode, and pre-existing craving, depression, and anxiety levels, the condition by COPE-Alcohol interaction term significantly predicted ACQ-SF scores (β = 0.78, p < .001, Model 1). When running a linear regression for the neutral and negative affect condition separately, COPE-Alcohol scores significantly predicted ACQ-SF scores after the negative affect induction (β = 0.67, p < .001), but not after a neutral condition (p = .139). The condition by ERQ-Reappraisal interaction term also significantly predicted ACQ-SF scores (β = −0.79, p = .013, Model 2). When running a linear regression for the neutral and negative affect condition separately, ERQ-Reappraisal scores significantly predicted ACQ-SF scores after the negative affect induction (β = −0.25, p = .029), but not after the neutral condition (p = .124). In contrast, ERQ-Suppression and its interaction with condition were not significant predictors (ps > .193, Model 3). Taken together, these results revealed that drinking to cope and cognitive reappraisal, but not emotion suppression, moderated the impact of negative affect induction on alcohol craving.

Table 5.

Three hierarchal multiple linear regressions to predict ACQ-SF scores

		β	95% CI		Adjusted p	∆R 2
Level 1	Sex (1 = Male)	0.16	0.01	0.30	.072	.33
	PASC scores	0.48	0.34	0.63	< .001
	Days after the last drinking episode	−0.04	−0.19	0.10	.613
	DASS-Depression score	0.13	−0.05	0.31	.271
	DASS-Anxiety score	0.07	−0.12	0.25	.530
Level 2
Model 1	Condition (1 = Negative Emotion)	−0.63	−0.99	−0.27	.003	.09
	COPE-Alcohol scores	−0.09	−0.29	0.11	.406
	Condition x COPE-Alcohol	0.78	0.41	1.15	< .001
Model 2	Condition (1 = Negative Emotion)	0.83	0.32	1.36	.007	.05
	ERQ-Reappraisal scores	0.19	−0.01	0.39	.132
	Condition x ERQ-Reappraisal	−0.79	−1.32	−0.26	.011
Model 3	Condition (1 = Negative Emotion)	0.25	−0.06	0.55	.193	.02
	ERQ-Suppression scores	0.05	−0.16	0.25	.692
	Condition x ERQ-Suppression	−0.20	−0.54	0.15	.368

PASC: Penn Alcohol Craving Scale; DASS: Depression Anxiety scores; COPE-Alcohol - 2 alcohol items in Brief COPE; ERQ: Emotion Regulation Questionnaire. FDR adjusted p values were noted

Model 1 in Figure 2 depicts the interaction effect. A steeper positive slope was observed in the negative affect condition whereas an almost flat line was observed in the neutral condition. These results indicated that more use of alcohol to cope predicted greater impact of negative affect induction on alcohol craving.

Figure 2.

The interaction between condition by COPE-Alcohol scores predicted ACQ-SF scores (Model 1) and the interaction between condition by ERQ-Reappraisal scores predicted ACQ-SF scores (Model 2). ACQ-SF: Alcohol Craving Questionnaire-Short Form-Revised; ERQ: Emotion Regulation Questionnaire; COPE-Alcohol - 2 alcohol items in Brief COPE.

The opposite interaction effect was observed for ERQ-Reappraisal scores (Figure 2, Model 2). A negative slope was observed in the negative affect induction. These results revealed that more use of cognitive reappraisal predicted less impact of negative affect induction on alcohol craving.

Discussion

With the long-term goal of reducing problematic drinking in college students, we examined whether use of adaptive versus maladaptive ER strategies (drinking to cope, cognitive reappraisal, emotion suppression) would predict alcohol craving at the time of negative affect induction. Prior to affect induction, we found that the tendency to use drinking to cope was positively associated with alcohol craving for the past week, problematic drinking, and emotional distress. In contrast, the tendency to use more cognitive reappraisal and less emotion suppression was associated with less emotional distress, but not associated with alcohol craving and problematic drinking. After affect induction, drinking to cope and cognitive reappraisal strategies moderated the impact of negative affect induction on alcohol craving. Specifically, more use of drinking to cope and less use of cognitive reappraisal predicted greater alcohol craving following negative affect induction in college drinkers, whereas neither ER strategy predicted craving following the neutral affect induction. In contrast, emotion suppression was not associated with alcohol craving in either condition. Taken together, the current results highlight the role of ER tendencies — particularly the benefits of cognitive reappraisal — on alcohol craving at the time of experiencing emotional distress.

Drinking to Cope and Alcohol Craving

The positive relationship between alcohol craving and problematic drinking at baseline in the current study is consistent with prior laboratory work showing associations between alcohol craving and problematic drinking in college students.^{27, 48} Moreover, the positive relationship between drinking to cope and both problematic drinking and negative psychological health status is consistent with prior work showing relationships between drinking to cope and more alcohol consumption,⁴⁹ episodes of heavy drinking,⁵⁰ and depressive symptoms.²² The current study extends the literature by demonstrating that college drinkers with a tendency to drink to cope experienced more craving when undergoing a negative emotion-evoking condition, but not when undergoing a neutral condition. These findings suggest while the tendency to drink to cope relates to overall levels of craving and drinking behaviors, drinking as a coping strategy also plays a role for episodic college drinkers when they are experiencing challenging emotions. Building on the Motivational Model of Alcohol Use,⁵¹ the current study’s results suggest that during negative emotional experiences, college drinkers experience greater alcohol cravings that are compounded by greater drinking to cope tendencies. Heightened craving may then lead to alcohol use to relieve negative emotions in the short-term, but may contribute to more problematic drinking behaviors long-term.

Emotion Regulation and Alcohol Craving

We found that neither cognitive appraisal nor suppression tendencies were associated with alcohol craving or drinking behaviors at baseline. Our current findings contrast with past cross-sectional and observational research documenting that difficulty regulating negative emotions is associated with more problematic college drinking.^18–21 However, the lack of relationship between cognitive appraisal tendencies and problematic drinking before affect induction in the current study is consistent with a recent study that too shows no significant relationship between cognitive appraisal tendencies and problematic drinking among older adolescents.⁵² This study has further compared the role of cognitive reappraisal in problematic drinking between two adolescent groups: immigrant vs native, and has found the protective role of cognitive reappraisal only in immigrant adolescents. Although psychological symptoms are not measured in this recent study,⁵² depression symptoms and daily stressors have been reported to be higher in immigrant adolescents than native adolescents.^{53, 54} Our study also showed that the tendency to use cognitive reappraisal to cope predicted less craving in those who underwent negative affect induction. Therefore, cognitive appraisal may have a protective role against craving and problematic drinking only in those undergoing emotionally challenging, stressful situations. Additionally, the current findings correspond with Gross’ Process Model of Emotion Regulation,⁵⁵ which suggests ER strategies, such as cognitive reappraisal, reduce and modulate emotional responses to negative cues. In the neutral condition, there was no negative cue for ER tendencies to modulate. Thus, ER tendencies may play a significant role in alcohol craving only when undergoing emotionally challenging situations. Our findings, in combination with previous findings, suggest that ER tendencies, drinking to cope,^{19, 20} psychosocial maturity,¹⁸ childhood adversity,²¹ and drinking refusal self-efficacy²¹ influence alcohol craving and drinking behaviors in college drinkers.

We did not find a significant role of emotion suppression in alcohol craving. This outcome is consistent with a prior laboratory study reporting that cognitive reappraisal, but not emotion suppression, moderated the effect of a mood manipulation task on smoking behavior among a depressed sample of 46 adult smokers.⁵⁶ Also consistent with our findings, a cross-sectional study of 134 college students reported that more problematic behaviors (arguing or fighting) while under the influence of alcohol were related to less use of cognitive reappraisal, but not related to emotion suppression.²⁴ Based on these two studies and our findings, there may be no significant role of emotion suppression in alcohol craving, alcohol-related problematic behaviors, and smoking. Alternatively, emotion suppression may play a role in craving only for those who use emotion suppression in high frequency. For instance, a cross-sectional study in 1,857 college students reported that among students drinking alcohol before a main event (pregaming), emotion suppression was significantly associated with alcohol problems only for students whose ER-suppression score of 14 or higher.⁵⁷ Hence, a future study should examine whether there would be a non-linear relationship between emotion suppression and alcohol craving using a large sample of college drinkers.

The current study is the first to demonstrate the moderating role of cognitive reappraisal tendencies on alcohol craving at the time of negative affect induction. This finding is consistent with a laboratory study demonstrating the favorable impact of explicit (instructed) cognitive reappraisal on cue-induced smoking craving.²⁶ The impact of the interaction between trait cognitive reappraisal and state mood on craving is also consistent with a recent study that observed among individuals with Alcohol Use Disorder, explicit cognitive reappraisal reduced negative emotions and cue-induced craving.⁵⁸ Taken together, craving responses, whether elicited from negative experiences or alcohol cues, are inversely related to cognitive reappraisal.

Clinical Implications

The present findings suggest potential intervention targets for alcohol use among episodic college drinkers. Because cognitive reappraisal has been found to moderate the impact of negative emotions on alcohol craving, health care providers at university behavioral health centers may need to routinely assess and target ER strategies to increase reappraisal in episodic college drinkers to reduce alcohol craving, especially during emotionally challenging times. Recent work suggests that brief cognitive therapy, an intervention characterized by cognitive reappraisal or restructuring, is a suitable way for reducing alcohol use in students showing risky drinking behaviors⁵⁹ and may prevent future development of problematic drinking or alcohol use disorder. Lastly, ER strategies may be an important treatment target not only for college drinkers, but for diverse groups of young adult drinkers who may experience disproportionate and uniquely challenging situations. For example, one laboratory study has demonstrated that negative mood induced by photographs with sexual stigma content increases alcohol craving among community-dwelling young adult heavy drinkers who self-identified as sexual minority.⁶⁰ Considering the relevance of cognitive reappraisal tendencies to alcohol craving in the current study, future study should examine whether learning cognitive reappraisal techniques moderate and ultimately reduce the impact of negative mood on alcohol craving and alcohol use in young adult drinkers from diverse sociodemographic backgrounds.

Limitations and Future Studies

The current study has limitations and strengths for consideration. First, this study examined cognitive reappraisal tendencies, limiting any interpretations of cognitive reappraisal use during the negative emotion-manipulation condition. Future experimental research manipulating explicit cognitive reappraisal among college students with episodic drinking could assess the feasibility of reducing craving in negative affect induction circumstances. Additionally, the current study focused on the relationship between self-reported ER, craving, and alcohol use, along with the impact of ER and coping tendencies on self-reported craving response to induced negative emotion. An integral next step is to examine whether ER and coping tendencies contribute to the development of problematic drinking. Replication and extension of the current results to reduce not only craving reports but likewise drinking behavior would support the practical and theoretical applicability of these data. Additionally, higher levels of negative affect and craving in the negative mood induction condition were likely due to the mood induction because we used randomization and a validated mood induction protocol. Affective states in each mood induction condition were all in the expected direction. Yet, the current study did not include baseline measures of mood and craving. Therefore, future studies assessing the changes in mood and craving pre-post induction are warranted to determine the effect of negative mood on alcohol craving and the moderating role of emotion regulation strategies among college drinkers. Acknowledging this limitation, we ran regression analyses adjusting for pre-existing levels of negative affect and craving and our analyses revealed that cognitive reappraisal was a significant moderator of the impact of negative mood induction on alcohol craving. Finally, the generalizability of our findings to female and racial/ethnic minority students may be limited because our sample included more male and white students. Despite these limitations, the current study has several strengths. First, we did not collect any personally identifiable information and all experimental procedures were done using a private computer without in-person interaction to minimize social desirability bias in the report of alcohol use and craving. The experimenters available to answer participants’ questions and to solve technical issues were blinded to participants’ drinking patterns and mood induction condition. Lastly, we experimentally examined the contribution of ER tendency on alcohol craving during an emotional challenge by randomly assigning college drinkers to a neutral or negative affect induction condition.

Conclusions

Overall, the current study provides novel empirical evidence that ER strategies moderate the impact of induced negative emotions on alcohol craving in college students who engage in episodic drinking. The current results underscore the relevance of drinking to cope and cognitive reappraisal on craving in episodic drinkers during negative emotional states. Considering the high prevalence rate of binge drinking in college students and their increased risk of developing future alcohol disorders, a better understanding of emotion regulation’s impact on craving responses offers insights into intervention targets for this population.