Negative Affect, Harassment and Problematic Alcohol Use in Young Adults

Abstract

Background:

While research suggests that both negative affect and alcohol use are impacted by exposure to harassment (i.e., sexual harassment, generalized harassment or bullying), less is known about the effect of harassment on negative affect subsequently leading to alcohol consumption, particularly in young adults. We examined the mediating role of negative affect on the relationships between sexual and generalized harassment at school and alcohol misuse.

Methods:

Participants were 2899 incoming freshmen in fall of 2011 who completed a Web-based survey assessing demographics (T0), sexual and generalized harassment at school (T0-T2), negative affect (T3), and problems associated with drinking, binge drinking, and drinking to intoxication (T0, T4, T5). Separate hybrid path models were fitted in Mplus v.8.8 for generalized harassment and sexual harassment and each outcome.

Results:

Mediation analyses showed a small but significant indirect effect for the sexual harassment model (beta=.05, S.E.=.01, p<.001) and generalized harassment (beta=.03, S.E.=.01, p<.01), indicating that negative affect partially mediated the associations between harassment early in students’ college experience and later problems associated with drinking. No significant indirect effects were found for the binge drinking or intoxication models.

Conclusions:

High levels of negative affect associated with harassment may contribute to longer term impact on problematic use of alcohol in young adults, providing evidence that the effects of harassment on drinking may partly stem from harassment’s lingering effects on negative affective pathways.

Introduction

Heavy drinking and alcohol-related risky behaviors, such as binge drinking, are major public health concerns among college students (Hingson et al., 2017; Slutske, 2005; Substance Abuse and Mental Health Services Administration, 2011; Wechsler, 1994). Over 50% of college students currently consume alcohol and it is estimated that 33% engage in binge drinking, 8% engage in heavy drinking, and over 200,000 college students meet the diagnostic criteria for alcohol use disorder (Center for Behavioral Health Statistics and Quality & Substance Abuse and Mental Health Services Administration, 2020b, 2020a). Heavy alcohol consumption and alcohol-related risky behaviors such as binge drinking cause many alcohol related consequences such as poor academic performance (Piazza-Gardner et al., 2016), long-term health related issues such as liver disease (World Health Organization, 2018), high levels of negative mental health symptomology such as depression (Geisner et al., 2012), and increased engagement in risky behaviors such as risky sexual behavior (Desiderato & Crawford, 1995). Moreover, deaths related to alcohol (e.g., driving while intoxicated) and alcohol-overdose hospitalizations increased significantly from 1998–2014 (Hingson et al., 2017). These alcohol-related issues have remained persistent up to 2023, illustrating an ongoing problem (National Institute on Alcohol Abuse and Alcoholism, 2023). To reduce current rates of problematic alcohol use and the harms associated with risky drinking among college students, it is important to understand the etiology and maintenance of why certain drinking behaviors or patterns occur.

Harassment, either sexual or general (e.g., bullying) plays a role in college student drinking (National Institute on Alcohol Abuse and Alcoholism, 2023). Approximately 30% of college students reports of peer sexual harassment, with women reporting higher rates than men (Wood et al., 2021) and approximately 20–25% peer general harassment (Lund & Ross, 2017). However, many believe that experiences of harassment tend to be underreported (Klein & Martin, 2021). Generally, experiences of harassment have been related to increased psychological distress (Wolff et al., 2017) and to disrupted functioning in work related tasks or environments (Pina & Gannon, 2012). College students with history of harassment tend to engage in drinking to cope with past negative experiences (Cooper et al., 1995; Martinez et al., 2014). One widely studied model is the self-medication hypothesis, suggesting that individuals use substances such as alcohol following traumatic or stressful situations to regulate negative affect (Cooper et al., 1995; Khantzian, 1997). Negative affect is typically understood as emotional experiences of sadness, anger, fear, and depression and research has consistently showed strong evidence for the use of substances as a modality for regulating these negative emotions (Reese et al., 2018). It is possible that college students with history of harassment consume alcohol to regulate negative affect associated with harassment.

Yet, studies on the role of negative affect, harassment, and alcohol-related risky behaviors (e.g., binge drinking and intoxication) are lacking. One study on adolescents showed a proximal association between alcohol use and negative affect following experiences of sexual harassment (Livingston et al., 2022). However, another study on adolescents who experienced bullying showed greater levels of reported sadness, anger, and cigarette use, but not alcohol use (Livingston et al., 2019). Additionally, among college student workers who reported sexual harassment, psychological distress, depression symptoms, and problematic alcohol use showed reciprocal causal affects (Wolff et al., 2017). Although the evidence is mixed on negative affect, and on alcohol use and misuse following harassment, ultimately, the inability to effectively cope or manage negative affect following harassment is a growing concern.

More importantly, there is limited evidence that negative affect is associated with harassment early in students’ college experience and later problems associated with drinking. The current study fills gaps in existing research by examining the mediating role of negative affect on the relationships between sexual and generalized harassment at school and changes in problems associated with drinking. Based on our prior work, we hypothesize that negative affect would mediate the associations between harassment and alcohol use and misuse among college students. Findings from this study could help identify individuals who may benefit from intervention following harassment.

Methods

Human subjects

All study activities were approved by the Institutional Review Boards of the participating schools.

Participants and Procedures

Participants were recruited from 9100 incoming freshman at 8 colleges and universities from one state in the Midwestern United States (six schools provided random samples and two allowed us to sample all freshman students). Invitations to complete a web survey were sent via email or mail (depending on contact information available) at five time points: at the beginning of the students’ first year of college (T0), about 4 months later in spring of 2012 (T1), about four months later in fall of 2013 (T2), and then at 12 month intervals (T3, T4, T5). Students 18 years of age or older were eligible to participate. Students received a gift certificate in the amount of $25 for completing the T0 survey, $30 at T1-T3, and $40 at T4 and T5. The study was reviewed and approved by the Institutional Review Board for the University of Illinois Chicago (protocol 2010–02430). A total of N = 2984 respondents participated at T0. From that sample, we dropped n = 13 who did not identify an academic institution or who attended a school that was not one of the nine targeted for the study. We eliminated an additional n = 72 participants who were ages 21 to 49 years at T0. We eliminated respondents who were non-traditional college students (i.e., older than the conventional age of college entry around ages 17–19 years), as their drinking behaviors often vary from those of traditional-aged college students (e.g., Babb et al., 2012). Relative to the rest of the sample, there were too few of this type of respondent to consider including these cases as a subgroup (i.e., of the n = 72 excluded cases, at T0 n = 54 were ages 21–29, n = 13 were ages 30–39, and n = 5 were ages 40–49). We retained data for n = 682 cases for which age data were available (see explanation in McGinley et al., 2015).

The final analytic sample consisted of N = 2899 cases drawn from eight colleges/universities in Illinois. The sample was diverse, with 58.3% of respondents identifying as women, mean age = 18.40 years (SD = 0.53, range = 18 – 20), 54.0% as white, 17.0% Asian, 13.2% Latinx, 8.2% African American, and 7.6% other racial/ethnic identifications. Of the original sample at T0, 44.8% provided data on study variables at all six Waves. At T1, 70.7% of the sample was present, followed by 68.9% at T2 (77.8% participated in T1 and/or T2), 63.1% at T3, 65.6% at T4, and 62.0% at T5 (70.2% participated in T4 and/or T5).

Measures

Descriptive statistics are provided in Table 1. Unless specified as an exception, scores were calculated through averaging responses to items within each timepoint.

Table 1.

Descriptive Statistics

Variable	N	M (SD) / %	Range
Sexual Harassment T0*	2864	.23 (.37)	0 – 2
General Harassment T0	2881	.53 (.46)	0 – 2
Alcohol Problems T0*	2838	.11 (.26)	0 – 2.96
Binge Drinking T0	2839	1.01 (1.48)	0 – 7
Intoxication T0	2836	.95 (1.44)	0 – 7
Sexual Harassment T1*	2044	.13 (.26)	0 – 2
General Harassment T1	2048	.26 (.33)	0 – 2
Sexual Harassment T2*	1993	.09 (.24)	0 – 2
General Harassment T2	1996	.20 (.31)	0 – 2
Negative Affect T3	1829	1.66 (.71)	1 – 5
Alcohol Problems T4*	1899	.20 (.36)	0 – 3.70
Binge Drinking T4	1895	2.15 (1.85)	0 – 7
Intoxication T4	1891	2.14 (1.81)	0 – 7
Alcohol Problems T5*	1788	.21 (.36)	0 – 4
Binge Drinking T5	1791	2.16 (1.83)	0 – 7
Intoxication T5	1790	2.18 (1.82)	0 – 7

Note.

^*Values reported for raw variables, prior to transformation in order to reduce skew or kurtosis.

Sexual Harassment at School (T0-T2)

Sexual harassment was assessed with a modified 13-item version of the Sexual Experiences Questionnaire (SEQ; Fitzgerald, 1988), worded to make items applicable to both men and women. The SEQ measures three types of harassment experiences: gender harassment, unwanted sexual attention, and sexual coercion. Items were rated as occurring “0=never,” “1=once,” or “2=more than once” in the past 12 months prior to college (T0) or past 4 months (T1 and T2); ω_h = .58 and ω_t = .94 across waves. Preliminary analyses revealed excessive kurtosis, which we reduced through the application of a log transformation. In analyses, the three transformed T0-T2 scores all loaded positively on a latent variable of sexual harassment at school.

Generalized Harassment at School (T0-T2)

We used a modified 20-item version of the Generalized Workplace Harassment Questionnaire (GWHQ) to assess experiences of generalized harassment at school (Rospenda & Richman, 2004). The GWHQ assesses four conceptual types of generalized harassment: covert hostility (e.g., excluded from important school or social activities, meetings, or events; 3 items), verbal hostility (e.g., talked down to you; 7 items), manipulation (attempts to control the target’s behavior e.g., through threats; 5 items) and physical aggression (hit, kicked, pushed or threw things at you; 1 item). We added items to assess more passive forms of harassment, e.g., failing to respond to request for help, cyberbullying (e.g., hurtful emails, texts, or messages on social media), being made the target of practical jokes that weren’t funny, and pressure to do things that the student did not want to do. Items were rated as occurring “0=never,” “1=once,” or “2=more than once” in the past 12 months prior to college (T0) or past 4 months (T1 and T2); ω_h = .66 and ω_t = .95 across waves. In analyses, the three T0-T2 scores all loaded positively on a latent variable of generalized harassment at school.

Negative Affectivity (T3)

Negative affectivity was assessed with the 10 negative affect items from the Positive and Negative Affect Schedule (PANAS) (Watson et al., 1988). Participants rate the way they generally feel in terms on a 5-point scale (1= “very slightly or not at all” to 5= “extremely”). In analyses, this variable was modeled as an observed variable (ω_h = .77 and ω_t = .93).

Alcohol Problems (T0, T4, and T5)

Alcohol problems, or negative consequences from alcohol use in the past 12 months, were measured with the 23-item Rutgers Alcohol Problems Index (RAPI) (White & Labouvie, 1989). The RAPI assesses problems with relationships (e.g., fights or arguments with friends or family), problems at work or school (e.g., missing work, not able to do homework), and tolerance or withdrawal symptoms (e.g., needing to drink more to get the same effect, feeling sick because of cutting down on drinking). Items are scored on a scale from 0= “never” to 4= “more than 10 times”, and items are summed to form a composite scale (ω_h = .62 and ω_t = .96 across waves). At T0, participants who reported no lifetime drinking skipped this measure, and thus we set their scores on this variable to zero (i.e., lowest possible level of problems). At T4 and T5, those who reported no drinking in the last year skipped this measure, and thus for these waves we also set those scores to zero. Preliminary analyses revealed excessive kurtosis, which we reduced through the application of a log transformation. In analyses, the transformed RAPI problems at T0 was included as an observed variable, while the T4 and T5 transformed scores loaded positively on a latent variable.

Binge Drinking (T0, T4, and T5)

Binge drinking was measured with the item, “About how often in the past 12 months did you have 5 or more drinks (men)/4 or more drinks (women) of any alcoholic beverage on the same occasion?” (response scale 0= “never” to 5= “5 times per week or more”; Wilsnack et al., 1991). At T0, participants who reported no lifetime drinking skipped this question, and thus we set their scores to zero (i.e., lowest possible level of binge drinking). At T4 and T5, those who reported no drinking in the last year skipped this measure, and thus for these waves we also set those scores to zero. In analyses, binge drinking at T0 was included as an observed variable, while the T4 and T5 binge drinking scores loaded positively on a latent binge drinking variable.

Intoxication (T0, T4, and T5)

Frequency of intoxication was measured with the item “About how often in the last 12 months did you drink enough to feel drunk, that is, where drinking noticeably affected your thinking, talking, and behavior?” (response scale 0= “never” to 5= “5 times per week or more”) (Wilsnack et al., 1991). As with the preceding alcohol use measures, at T0, participants who reported no lifetime drinking skipped this question, and thus we set their scores to zero (i.e., lowest possible level of drinking to intoxication). At T4 and T5, those who reported no drinking in the last year skipped this measure, and thus for these waves we also set those scores to zero. In analyses, intoxication at T0 was included as an observed variable, while the T4 and T5 intoxication scores loaded positively on a latent intoxication variable.

Covariates (T0)

We covaried for gender (female=1) and race/ethnicity (four dummy codes corresponding to Black, Latinx, Asian, and other ethnicities, with white as the reference group).

Analysis Plan

Preliminary analyses included Pearson’s correlations. Hypotheses were tested in a series of hybrid path analyses in Mplus v.8.8 (Muthén & Muthén, 1998). Separate models were fitted for generalized harassment and sexual harassment (see Figure 1), and all models included all direct effects (i.e., the dependent alcohol use variables were simultaneously regressed upon negative affect, harassment, and the demographic and alcohol use covariates), and the bootstrapped indirect effect of the harassment variable on alcohol use via negative affect. The primary indicator of acceptable model fit was a nonsignificant chi-square fit statistic (χ²). As chi-square statistics are sensitive to sampling fluctuation, comparative fit indices (CFIs) larger than .95, root mean square errors of approximation (RMSEAs) smaller than .06, or 90% confidence intervals (CIs) that contain .06, and standardized root mean square residuals (SRMRs) smaller than .08 were also used to indicate sufficient fit (West et al., 2012).

Figure 1.

General Model

Note. Correlations between the exogenous covariates were also estimated, and are available upon request from the corresponding author.

Missing data analyses suggested that data were not completely missing at random (MCAR), Little’s MCAR χ² (1439) = 2040.71, p < .001. Those who were missing one or more waves of data had higher levels of alcohol problems (T0, T4, and T5), higher levels of binge drinking and intoxication at baseline, and were more likely to identify as Black, compared to those who were present at all study waves (ps < .05). When data are not MCAR, full information maximum likelihood (FIML) can generate unbiased estimates when the variables explaining missingness are modeled (Enders & Bandalos, 2001).

Results

Preliminary Analyses

Bivariate correlations are reported in Table 2. Female respondents reported higher levels of sexual harassment and negative affect than male respondents. There were no gender differences in reports of general harassment, and men reported higher levels of binge drinking and drinking to intoxication at all Waves. Relative to white participants, Black, Latinx, and Asian respondents tended to report lower levels of many dimensions of alcohol misuse across the study’s Waves. Latinx and Asian participants also reported lower levels of harassment relative to White students, though these associations were not consistent across all study Waves. High levels of sexual and general harassment in school at T0-T2 were associated with high levels of negative affect at T3 and with all three dimensions of alcohol misuse at T0, T4 and T5. High levels of negative affect at T3 were associated with high levels of alcohol problems only at T0, T4 and T5 (i.e., negative affect was not associated with binge drinking or with intoxication at any wave). High levels of alcohol misuse at T0 were associated with high levels at T4 and T5.

Table 2.

Correlations

Variable		1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
1	Female
2	Black	.08c
3	Latinx	.02	−.12c
4	Asian	−.09c	−.14c	−.18c
5	Other	.02	−.09c	−.11c	−.13c
6	Sexual Harassment T0	−.19c	.04	.03	−.12	.04a
7	General Harassment T0	.04	−.01	−.02	−.08c	.05	.57c
8	Alcohol Problems T0	.03	−.02	.03	−.09c	.01	.30c	.32c
9	Binge Drinking T0	−.04	−.09c	−.03	−.13c	.01	.10c	.14c	.52c
10	Intoxication T0	−.05a	−.10c	−.09c	−.15c	.00	.14c	.18c	.57c	.83c
11	Sexual Harassment T1	.12c	.05a	−.04	−.05a	−.01	.40c	.30c	.22c	.06a	.08b
12	General Harassment T1	−.02	.00	−.07b	−.02	.01	.32c	.40c	.20c	.08b	.11c	.57c
13	Sexual Harassment T2	.12c	.00	−.05a	−.03	−.03	.38c	.31c	.14c	.06a	.09c	.46c	.33c
14	General Harassment T2	.00	−.01	−.08c	.02	−.01	.29c	.39c	.15c	.09c	.12c	.38c	.51c	.60c
15	Negative Affect T3	.10c	.04	.03	.03	.04	.19c	.24c	.16c	.01	.01	.21c	.22c	.18c	.26c
16	Alcohol Problems T4	−.02	−.05a	−.02	−.10c	.02	.21c	.26c	.42c	.30c	.35c	.25c	.24c	.22c	.29c	.22c
17	Binge Drinking T4	−.07b	−.11c	−.08c	−.22c	−.01	.10c	.14c	.29c	.37c	.41c	.15c	.15c	.12c	.17c	.02	.58c
18	Intoxication T4	−.08b	−.11c	−.10c	−.23c	−.02	.11c	.14c	.29c	.37c	.44c	.15c	.15c	.12	.17c	.01	.59b	.90c
19	Alcohol Problems T5	−.05	−.04	−.02	−.10c	−.02	.19c	.27c	.36c	.28c	.33c	.19c	.24c	.23c	.30c	.20	.68c	.51c	.52c
20	Binge Drinking T5	−.06b	−.09b	−.07b	−.22c	−.03	.09c	.14c	.24c	.34c	.38c	.14c	.16c	.12c	.16c	.02	.48c	.73c	.73c	.56c
21	Intoxication T5	−.06b	−.09b	−.09c	−.22c	−.02	.11c	.15c	.25c	.33c	.40c	.15c	.17c	.12c	.15c	.02	.48c	.72c	.75c	.56c	.89c

Note.

^ap < .05,

^bp < .01,

^cp < .001.

Sexual Harassment Models.

Coefficients for all sexual harassment models are provided in Table 3.

Table 3.

Summary of Path Models: Sexual Harassment

	Dependent Variable
	RAPI Problems Model		Binge Drinking Model		Intoxication Model
Covariate	Negative Affect	RAPI Problems	Negative Affect	Binge Drinking	Negative Affect	Intoxication
Negative Affect		.04 (.01)***		−.02 (.06)		−.01 (.06)
Harassment at School	1.29 (.19)***	.31 (.04)***	1.52 (.19)***	1.66 (.38)***	1.51 (.19)***	1.30 (.34)***
Prior Alcohol Use	.30 (.18)**	.46 (.03)***	−.01 (.01)	.46 (.03)***	−.01 (.01)	.51 (.02)***
Female	.04 (.04)	−.05 (.01)***	.04 (.04)	−.35 (.08)***	.03 (.04)	−.33 (.08)***
Black	.14 (.07)*	−.06 (.02)**	.13 (.07)	−.83 (.14)***	.13 (.07)	−.80 (.14)***
Latinx	.12 (.05)*	−.03 (.01)*	.12 (.05)**	−.57 (.10)***	.12 (.05)*	−.62 (.10)***
Asian	.16 (.05)***	−.05 (.01)***	.15 (.05)**	−.97 (.10)***	.15 (.05)**	−.96 (.09)***
Other	.15 (.06)*	−.02 (.02)	.16 (.07)*	−.44 (.15)**	.16 (.07)*	−.49 (.15)**
Indirect Effect 95% C.I.		.05 (.01)*** [.024, .080]		−.03 (.09) [−.213, .151]		−.02 (.09) [−.220, .157]
R 2	.11***	.35***	.11***	.33***	.11***	.36***

Note. All coefficients are unstandardized betas (SE B).

^*p < .05,

^**p < .01,

^***p < .001.

Alcohol Problems

Model fit was acceptable, χ² (25) = 133.29, p < .001, CFI = .965, RMSEA = .039, 90% C.I. [.032, .045], SRMR = .020. Relative to all other ethnic groups, white youth reported lower levels of negative affect. High levels of sexual harassment at school and high levels of prior alcohol problems predicted high levels of negative affect. In turn, women reported lower levels of subsequent alcohol problems, as did Black, Latinx, and Asian students relative to their white peers. High levels of prior sexual harassment at school, prior alcohol problems and negative affect predicted high alcohol problems at T4/T5. There was a significant indirect effect of prior sexual harassment at school via negative affect, B = .05, SE B = .01, p < .001, 90% C.I. [.024, .080].

Binge Drinking

Model fit was acceptable, χ² (25) = 82.50, p < .001, CFI = .981, RMSEA = .028, 90% C.I. [.022, .035], SRMR = .016. White youth reported lower levels of negative affect relative to students who identified as Latinx, Asian, and other ethnicities. High levels of sexual harassment at school predicted high levels of negative affect. In turn, women reported lower levels of subsequent binge drinking, and white students reported higher levels of binge drinking relative to their peers from all other ethnic identifications. High levels of prior sexual harassment at school and prior binge drinking predicted high binge drinking at T4/T5. Negative affect did not predict binge drinking, and there was no significant indirect effect of prior harassment via negative affect.

Intoxication

Model fit was acceptable, χ² (25) = 82.76, p < .001, CFI = .983, RMSEA = .028, 90% C.I. [.022, .035], SRMR = .016. As before, White youth reported lower levels of negative affect relative to students who identified as Latinx, Asian, and other ethnicities. High levels of sexual harassment at school predicted high levels of negative affect. In turn, women reported lower levels of subsequent intoxication, and white students reported higher levels of drinking to intoxication relative to their peers from all other ethnic identifications. High levels of prior sexual harassment at school and prior intoxication predicted high drinking to intoxication at T4/T5. Negative affect did not predict intoxication, and there was no significant indirect effect of prior harassment via negative affect.

General Harassment Models

Coefficients for all general harassment models are provided in Table 4.

Table 4.

Summary of Path Models: General Harassment

	Dependent Variable
	RAPI Problems Model		Binge Drinking Model		Intoxication Model
Covariate	Negative Affect	RAPI Problems	Negative Affect	Binge Drinking	Negative Affect	Intoxication
Negative Affect		.03 (.01)**		−.08 (.06)		−.07 (.06)
Harassment at School	.92 (.10)***	.23 (.02)***	1.02 (.09)***	1.15 (.19)***	1.02 (.09)***	.97 (.18)***
Prior Alcohol Use	.20 (.12)	.44 (.04)***	−.01 (.01)	.45 (.03)***	−.02 (.01)	.50 (.02)***
Female	.12 (.03)***	−.03 (.01)**	.13 (.03)***	−.24 (.08)**	.13 (.03)***	−.24 (.07)***
Black	.17 (.07)*	−.05 (.02)*	.15 (.07)*	−.79 (.14)***	.15 (.07)*	−.77 (.14)***
Latinx	.17 (.05)**	−.01 (.01)	.16 (.05)*	−.52 (.10)***	.16 (.05)**	−.58 (.10)***
Asian	.14 (.05)**	−.05 (.01)***	.12 (.05)*	−.99 (.09)***	.12 (.05)**	−.97 (.09)***
Other	.15 (.07)*	−.02 (.02)	.15 (.07)*	−.46 (.16)**	.15 (.07)*	−.50 (.15)***
Indirect Effect 95% C.I.		.03 (.01)** [.009, .042]		−.08 (.07) [−.210, .050]		−.07 (.06) [−.197, .054]
R 2	.16***	.38***	.15***	.34***	.15***	.37***

Note. All coefficients are unstandardized betas (SE B).

^*p < .05,

^**p < .01,

^***p < .001.

Alcohol Problems

Model fit was acceptable, χ² (25) = 145.40, p < .001, CFI = .962, RMSEA = .041, 90% C.I. [.034, .047], SRMR = .021. Women reported higher levels of negative affect. As before, relative to all other ethnic groups, white youth reported lower levels of negative affect. High levels of general harassment at school predicted high levels of negative affect. In turn, women reported lower levels of subsequent alcohol problems, as did Black and Asian students relative to their white peers. High levels of prior general harassment at school, prior alcohol problems and negative affect predicted high alcohol problems at T4/T5. There was a significant indirect effect of prior general harassment at school via negative affect, B = .03, SE B = .01, p < .01, 90% C.I. [.009, .042].

Binge Drinking

Model fit was acceptable, χ² (25) = 68.20, p < .001, CFI = .986, RMSEA = .024, 90% C.I. [.018, .031], SRMR = .015. Women reported higher levels of negative affect. White youth reported lower levels of negative affect relative to students from all other ethnicities. High levels of general harassment at school predicted high levels of negative affect. In turn, women reported lower levels of subsequent binge drinking, and white students reported higher levels of binge drinking relative to their peers from all other ethnic identifications. High levels of prior general harassment at school and prior binge drinking predicted high binge drinking at T4/T5. Negative affect did not predict binge drinking, and there was no significant indirect effect of prior harassment via negative affect.

Intoxication

Model fit was acceptable, χ² (25) = 72.85, p < .001, CFI = .986, RMSEA = .026, 90% C.I. [.019, .033], SRMR = .016. Women reported higher levels of negative affect. As before, White youth reported lower levels of negative affect relative to students from all other ethnicities. High levels of general harassment at school predicted high levels of negative affect. In turn, women reported lower levels of subsequent intoxication, and white students reported higher levels of drinking to intoxication relative to their peers from all other ethnic identifications. High levels of prior general harassment at school and prior intoxication predicted high drinking to intoxication at T4/T5. Negative affect did not predict intoxication, and there was no significant indirect effect of prior harassment via negative affect.

Discussion

The current study attempted to examine the mediating role of negative affect on the relationships between sexual and generalized harassment at school and changes in problems associated with drinking. After covarying for race/ethnicity and gender, overall results indicated that following experiences of both sexual and general harassment, college students reported greater changes in alcohol related problems at later waves of the study (i.e., higher reported levels of alcohol problems, binge drinking, and drinking to intoxication). Additionally, results showed increased endorsement of negative affect following experiences of general and sexual harassment. However, while negative affect did not mediate the relationship between harassment and changes in binge drinking or drinking to intoxication, it significantly mediated the relationship between changes in alcohol problems and harassment, showing an indirect relationship and providing support for our hypothesis. Consistent with previous studies (Cooper et al., 1995; Martinez et al., 2014), our findings suggest that college students with history of harassment may engage in drinking to cope with negative affect associated with harassment.

For both general and sexual harassment, negative affect was higher among racial and ethnic minority participants compared to their white peers, however, alcohol related problems at T4 and T5 of the study were higher among white participants than racial and ethnic minority participants. It is possible that racially and ethnically diverse college students may be experiencing negative affect at higher rates than their white peers due to additional forms of harassment, such as racism (Brondolo et al., 2008). However, our results are consistent with previous findings showing greater rates of alcohol use and misuse among young adult white college students than among Hispanic or Black peers (Krieger et al., 2018). There is evidence that this reflects on-campus drinking culture primarily being “white space” where BIPOC students may not feel comfortable, and concerns of Black students specifically that drinking may lead to harsher consequences from the university or university police (Peralta, 2005). It’s also possible that racially diverse college students may have developed more positive coping strategies and greater stress resilience after cumulative exposure to racial discrimination and other life stressors, reducing their risk of alcohol and substance use, or may be protected by cultural factors such as extended kinship networks and greater spirituality (McLaughlin et al., 2008). Additionally, female participants were less likely than males to report drinking related problems, which is consistent with previous study citing males as more likely to engage in risky drinking behaviors than females (Nolen-Hoeksema, 2004). To fully understand the observed differences based on gender and race/ethnicity and inform tailored theoretical interventions, future research should directly study motives for drinking following experiences of sexual, general, and racial harassment, as well as differences in subsequent consequences following an event of harassment.

Negative affect (Cooper et al., 1995) and experiences of harassment (Martinez et al., 2014) have long been associated with risk for increased drinking and alcohol related problems. However, to our knowledge, this is the first study to examine negative affect as a mediator between experiences of sexual or general harassment and changes in alcohol use and misuse among college students. Additionally, these results are strengthened by the longitudinal study design, which allowed us to observe changes in alcohol use and related problems, following events of harassment. These findings lend support for a call to action for improved and tailored interventions for specific types of harassment on college campuses. Experiences of harassment (general or sexual) have lifelong consequences, such as increased psychological distress, increased engagement in risky behavior, and increased risk for the development of a substance use disorder (Klein & Martin, 2021). Additionally, these consequences could be exacerbated by the experience of young adulthood and the navigation of college life, which is known to be accompanied by many life transitions (Arnett, 2000).

Limitations

Our study should be interpreted with consideration for certain limitations.

First, this study utilized self-report data and therefore, it is subject to faulty recall and accuracy of participant memory. Furthermore, while this study is longitudinal in design, results are correlational and not causal in nature. Finally, subjective experiences during each wave of the study may have influenced participants’ reports of negative affect. Thus, it is entirely possible a different external stressor (e.g., difficulty in classes or interpersonal difficulties) may be driving negative affect scores for subsets of respondents, rather than one or more events of sexual or general harassment. These limitations notwithstanding, this study has multiple strengths. First, this study is one of the first large-scale studies of experiences of negative affect, harassment (general or sexual) and subsequent drinking behavior in a college sample. This study utilized multiple schools and had measures of key variables of interest across six time points. Additionally, while there is a growing body of literature on the effects of harassment on subsequent substance use, this study adds to the literature by identifying one mechanism through which sexual and generalized harassment experiences may increase likelihood of problematic alcohol use – through increased negative affect.

Conclusions

High levels of negative affect subsequent to harassment experiences may contribute to longer term impact on problematic use of alcohol in young adults, providing evidence that the effects of harassment on drinking may partly stem from harassment’s lingering effects on negative affective pathways. Given support for our hypothesis that negative affect mediates the relationship between harassment experiences and changes in alcohol use (i.e., alcohol related problems), future research should continue to employ longitudinal methods and collect data that can further explain possible additional mediators linking negative affect to alcohol misuse, or other processes linking harassment to negative affect, and more specific forms of harassment, such as racism. Although the present study did not consider moderation, understanding which groups are more likely to develop alcohol related problems following harassment may help reduce the development of alcohol use disorders among college students who experience such maltreatment in school contexts. Ultimately, understanding contributors to problematic alcohol use among college students addresses a major public health concern; it is imperative to understand who may be at greater risk (e.g., individuals able to positively cope with negative affect or not) and associated mechanisms through which problematic alcohol use develops.