Drinking context and alcohol’s harm from others among men and women in the 2010 US National Alcohol Survey

Abstract

Background

Little is known about how drinking in different contexts is associated with harms from someone else’s drinking, including marital problems, financial problems, and assault. We examined how drinking in four different contexts was associated with alcohol’s harm from others (AHFO).

Methods

We utilized the landline sample of the 2010 US National Alcohol Survey (n = 5,885) to examine associations between drinking context and AHFO using weighted binary logistic regression.

Results

For women, drinking when friends dropped over was positively associated with assault and financial troubles due to someone else’s drinking. Drinking when friends dropped over was negatively associated with assault for men. For men, drinking at a bar, party, or during a quiet evening at home were each significantly associated with more assault by someone who had been drinking. Bar drinking among women was significantly associated with more marital problems, whereas drinking at a party at someone else’s home was associated with significantly less marital problems.

Conclusions

Context-specific drinking has differential associations with specific types of harms from someone else’s drinking for men and women. Additional research on drinking context, relationship to the harmer, and violence experienced by men and women is needed.

INTRODUCTION

Environmental context can contribute to greater drinking (Holyfield et al., 1995). Specific contexts such as bars or parties where heavy drinking and other risky behaviors are framed as normative can encourage such behaviors to occur (McAlaney et al., 2011; Ward, 2011). Indeed, research has found that bar environments can influence social drink refusal self-efficacy, descriptive drinking norms, and consumption (Lau-Barraco & Dunn, 2009; Monk & Heim, 2013; Neighbors et al., 2004). Heavy episodic drinking and frequent bar-going are associated with assault perpetration and victimization in bars among college students (Wells et al., 2011). However, only a few studies have examined the impact of drinking context on alcohol outcomes using general population samples, and such studies tend to focus on alcohol consumption, not on harms from someone else’s drinking. Thus, the purpose of this exploratory study is to examine how drinking context is associated with alcohol’s harm from others using a United States (U.S.) national sample of adult men and women.

Drinking has historically been conceptualized as an individualistic problem, not as an issue which affects those other than the drinker, although this topic is a growing public health concern (Laslett et al., 2013; Room et al., 2010). Because alcohol can have a widespread impact on the health of others than the drinker the World Health Organization (WHO) has recognized alcohol’s harm to others as an important component of global strategies to reduce the harmful use of alcohol (WHO, 2012). Despite the need for research on alcohol’s harm from others, to date studies utilizing general population samples and that examine the role of drinking context in exposure to harms have been limited.

Early work in the U.S. examined alcohol’s harm from others and drinking contexts using a general population sample from Berkeley, California (Fillmore, 1985). Alcohol-related violence victimization was more common in the home among women, and more common in bars or in public among men. Furthermore, frequency of bar going was strongly associated with multiple victimizations and was more a stronger risk factor for harms among women (Fillmore, 1985). Prior research utilizing the 1984, 1995, and 2005 U.S. National Alcohol Surveys to examine drinking context and alcohol problems among Black, Hispanic and White men and women and found that those who drank mostly at bars and other public venues were at a higher risk of arguments, fights, and drunk driving than those who drank less in these settings (Nyaronga et al., 2009). In a multi-national study, individual-level bar drinking frequency was associated with past year harms from someone else’s drinking (Roberts et al., 2013). Event-based analysis of bar-going adult women has documented that sexual victimization often occurred physically outside of bars, with perpetrators who were intoxicated (Parks, 2000). Notably, country-level genderedness of bar drinking (the proportion days of drinking in bars in a country in the past year by men) modified the associations between bar going and assault, family problems, and work harm, with the likelihood of harms associated with bar drinking increasing at a slower rate in countries with predominately male bar drinking (Roberts et al., 2013). Relatedly, studies have demonstrated that women experience violent and sexual victimization in bar settings (Parks, 2000). In nations such as the U.S. where women and men often frequent bars and other alcohol-oriented settings together, other drinkers’ harms against women in such settings may be more prominent.

Alcohol’s harm from others in the context of drinking in settings other than bars remains understudied, and thus is a focus of the current U.S. study. Karriker-Jaffe and Greenfield (2014) recently found that neighborhood context can change the patterning of alcohol-related harms from others for women and men, which indicates that environmental context is an important aspect of social exposures to harms in the U.S. Therefore, we examine how drinking volume in different contexts is associated with a range of harms from someone else’s drinking (family/marital problems, financial problems, and assaults) for women and men using the 2010 National Alcohol Survey. We extend prior research by examining how drinking in bar, home, and party settings is associated with this range of harms, with an explicit focus on gender differences.

We include these different settings in order to examine drinking contexts which involve social drinking, because the presence of other drinkers increases the risk for alcohol-related harm. For contrast, we also include drinking during a quiet evening at home in the analysis to assess a more private drinking context that might not involve many other drinkers. Drinking in this private context could be assessing moderate consumption with a non-drinking or moderate drinking partner, which might reduce the likelihood of alcohol-related harm from another drinker. Drinking in a private setting may be related to family/martial harms that may stem from intimate partners or family members who are problem drinkers. Therefore, we hypothesized:

• H1Drinking at bars and parties will be associated with more assault.
• H2Drinking when friends dropped over will be associated with more assault and family/financial problems.

METHODS

Sample

We utilized the landline sample of the 2010 U.S. National Alcohol Survey (NAS) to examine the associations between context-specific drinking and three externalities from someone else’s drinking in the last 12 months. The survey involved computer-assisted telephone interviews with a randomly-selected sample of U.S. adults (over age 18), with targeted oversamples of Black and Hispanic respondents and residents from sparsely-populated U.S. states. Reverse directory look-up facilitated advance mailings to about 50% of the respondents. Non-response was minimized by using multiple, largely-unlimited callbacks and refusal conversion attempts. Bilingual interviewers conducted interviews in Spanish when needed or requested. Within households, one eligible respondent was chosen in households using a Kish randomization scheme. The Institutional Review Board of the Public Health Institute, Oakland, California approved the survey. More details on the NAS are documented elsewhere (Kerr et al., 2013).

The 2010 NAS achieved a cooperation rate of 49.9%. Although response rates are lower than those often seen in face-to-face surveys, they are typical for random-digit dial telephone surveys in the U.S., and do not necessarily produce biased population estimates because many refusals (hang-ups) often occur before the study topic has been mentioned (Groves, 2006; Keeter et al., 2006). The analytic sample includes 5,885 landline telephone interviews (51% women, 49% men). All analyses were stratified by gender.

Measures

Alcohol’s harm from others

Three items (Greenfield et al., 2009) assessing alcohol’s harm from others items were drawn from the 1989 Canadian Alcohol and Other Drug Survey (Eliany et al., 1992). The primary outcomes included: family problems or marital difficulties due to someone else’s drinking (henceforth referred to as marital problems); financial trouble from someone else’s drinking; and being pushed, hit or assaulted by someone who had been drinking. Here we use harm exposures during the prior 12 months, since the time frame for context-specific drinking also is the prior 12 months (see below). Due to the small percentage of men reporting family/marital or financial problems, these harms were combined for men in the analysis, hereafter referred to as “marital/financial problems” (see Table 1). Marital and financial problems were significantly and positively correlated among men (Pearson r = .282; p = <.0001).

Table 1.

Sample Characteristics for the 2010 National Alcohol Survey

	Women	Men
	N (%)	N (%)	P-value
Alcohol’s harm to others
Family/Marital Problems	180 (5.3)	84 (2.6)	<.001
Financial Harm	95 (2.8)	32 (1)	<.001
Assault	83 (2.4)	168 (5.2)	<.001
Context-specific drinking, Mean (SD)
Bar	0.3 (1.4)	1.6 (11.9)	<.001
Party	0.3 (1.6)	1.9 (19.0)	<.001
Quiet evening at home	0.8 (3.9)	3.6 (26.1)	<.001
Friends dropping over	0.3 (3.2)	1.4 (9.9)	<.001
Age, Mean (SD)	47.51 (17.94)	44.90 (17.29)	<.001
Race/ethnicity			NS
White	2415 (68.6)	2270 (68.1)
Black	414 (11.7)	358 (10.7)
Hispanic	429 (12.2)	467 (14.0)
Other	263 (7.5)	239 (7.2)
Married/living with someone	1831 (63.8)	1718 (64.5)	NS
Education			NS
College graduate	921 (26.2)	856 (25.9)
Less than HS	511 (14.6)	512 (15.5)
HS Grad	1053 (30.0)	1026 (31.0)
Some College	1023 (29.2)	914 (27.6)
Employed	1684 (48.0)	2077 (62.3)	<.001
Unemployed	701 (20.0)	743 (22.3)
Retired/homemaker	1121 (32.0)	512 (15.4)
Income			<.001
≤10k–$20,000	851 (24.2)	721 (11.3)
$20,001–$60,000	1015 (28.8)	1080 (32.4)
>$60,000	1001 (28.4)	1157 (34.7)
Overall drinking volume, Mean (SD)	96.5 (279.2)	328.8 (730.9)	<.001
Family history of alcohol problems	1113 (32.5)	1013 (30.9)	<.05
Impulsivity, Mean (SD)	4.4 (3.8)	6.8 (5.0)	<.001

Note. P-values based on independent sample t-tests (continuous variables) and chi-square tests (dichotomous and categorical variables) and are based on weighted distributions

Context-specific drinking was measured for several settings using quantity-frequency items based on drinking in the prior 12 months. The four settings included drinking in bars, at a party at someone’s home, when friends drop over, and during a quiet evening at home. We computed average volume in the metric of number of drinks per week. As shown in Table 1, women had significantly fewer drinks peer week over the past year than did men in all settings.

Socio-demographic controls

Controls were respondent’s age, measured as a continuous variable (mean 46.2, SD = 17.7); mutually-exclusive indicator variables for non-Hispanic Black/African American, Hispanic, and Other race/ethnicity (referent = White); single marital status, which included widowed, divorced and never married (referent = married/living with someone); less than high school education (referent = at least a high school diploma); unemployment, which included those working less than 35 hours per week as well as those not working because in school, retired, homemakers, temporarily out of the workforce due to illness, and disabled (referent = full time employed); and a categorical variable indicating household income (with a category for missing income) 1 = ≤$20,000; 2 = $20,001–60,000; and the referent > $60,000. Due to the possibility that alcohol’s harms from others may be associated with the respondent’s own alcohol use (Aneshensel & Huba, 1983;Caetano, 1987), we controlled for respondents’ overall alcohol consumption, assessed by 12-month volume of drinking (drinks/year) based on a graduated quantity–frequency approach (Greenfield, 2000;Rehm et al., 1999) which asks about the frequency of drinking at 6 quantity levels ranging from 1 drink to 24 or more drinks per day, and which includes all alcoholic beverages (beer, wine and spirits) combined (Greenfield & Kerr, 2008; Wechsler et al., 1995 ). Frequency was captured on a 7-point scale ranging from “never” to “every day or nearly every day”. This approach is very effective for measuring consumption among individuals who occasionally drink heavily (Rehm et al., 1999). Alcohol problems in the family in early life and impulsivity have been found to be associated with alcohol-related problems (Salom et al., 2015; Hittner & Swickert, 2006). Therefore, we included the number of first degree relatives with a history of alcohol problems as a control variable. We also utilized a 7-item scale of sensation seeking/impulsivity (henceforth referred to as impulsivity) as a covariate (Cronbach’s alpha = .82, mean =5.57; SD=4.59).

Analysis

In order to assess how context-specific drinking was associated with AHFO, we used weighted binary logistic regressions. Data were weighted to the general population of the U.S. using the 2010 Census, taking account of age, sex, ethnic group and geographic area. All analyses were stratified by gender. All context-specific drinking variables, socio-demographic controls, respondents’ drinking, family history of alcohol problems and impulsivity were entered simultaneously in the multivariate models. Variance inflation factors (VIFs) indicated that multicollinearity was modest (Craney & Surles, 2002). The largest VIF was for women’s drinking when friends dropped over with a value of approximately 6. For men, the largest VIF was for drinking during a quiet evening at home with a value of approximately 5.

RESULTS

Descriptive Statistics

Weighted descriptive values and significance tests for differences between men and women are provided in Table 1. Women reported significantly more marital and financial problems than men. Significantly more men experienced assaults than did women. Men had a significantly higher number of drinks in the past year than did women and drank significantly more in all contexts than did women. Significantly more women than men reported low incomes.

Multivariate analysis

Context-specific drinking

As shown in Table 2, women’s bar drinking only was significantly associated with more marital problems due to others’ drinking. Conversely, only drinking at a party at someone else’s home was significantly associated with less marital problems from others’ drinking. Notably, drinking when friends dropped over was significantly associated with more assault and more financial problems from someone else’s drinking.

Table 2.

Adjusted Odds Ratios (aOR) for Alcohol’s Harm from Others among Women

	Assault	Marital Problems	Financial Problems
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Bar drinking	1.03 (0.92, 1.16)	1.16 (1.05, 1.28) **	1.06 (0.91, 1.22)
Party drinking	0.95 (0.80, 1.12)	0.77 (0.63, 0.95) *	0.79 (0.59, 1.05)
Home drinking	0.89 (0.75, 1.05)	1.04 (0.96, 1.13)	0.79 (0.62, 1.00)
Friends drop over	1.17 (1.01, 1.35) *	1.08 (0.98, 1.20)	1.41 (1.14, 1.73) **
Overall drinking volume	1.00 (1.00, 1.00)	1.00 (1.00, 1.00)	1.00 (1.00, 1.00)
# of first degree relative problem drinkers	1.46 (1.10, 1.92) **	1.59 (1.34, 1.88) ***	1.22 (0.95, 1.56)
Impulsivity	1.13 (1.07, 1.20) ***	1.07 (1.03, 1.12) **	1.06 (1, 1.12)
Age	0.96 (0.94, 0.98) ***	0.98 (0.96, 0.99) ***	0.97 (0.95, 0.98) ***
Minority race/ethnicity	1.09 (0.62, 1.92)	0.94 (0.62, 1.43)	0.52 (0.29, 0.93) *
Married/living with someone	1.03 (0.60, 1.75)	0.39 (0.25, 0.60) ***	0.56 (0.33, 0.94) *
LS than HS education	0.38 (0.13, 1.08)	1.16 (0.67, 1.81)	1.81 (0.96, 3.41)
Unemployed	0.75 (0.44, 1.30)	1.27 (0.88, 1.84)	1.48 (0.90, 2.44)
Income (referent >$60,000)	**	***	*
Income missing	1.7 (0.74, 3.92)	0.31 (0.14, 0.69) **	1.00 (0.45, 2.21)
<$20,000	2.57 (1.30, 5.09) **	1.04 (0.66, 1.66)	1.58 (0.85, 2.94)
$20,001-$60,000	0.79 (0.38, 1.63)	0.45 (0.29, 0.72) **	0.63 (0.33, 1.20)
Constant	0.04***	0.14***	0.09***

^*p < .05;

^**p < .01;

^***p < .001

Note. C.I. = Confidence interval.

As shown in Table 3 for men, similar to women, bar drinking was significantly associated with more marital/financial problems from other drinkers. Drinking at a bar, party, and during a quiet evening at home each were significantly associated with more assault by someone who had been drinking. However, drinking when friends dropped over was negatively associated with assault by other drinkers for men.

Table 3.

Adjusted Odds Ratios (aOR) for Alcohol’s Harm from Others among Men

	Assault	Marital/Financial Problems
	aOR (95% CI)	aOR (95% CI)
Bar drinking	1.06 (1.02, 1.11) **	1.08 (1.03, 1.12) **
Party drinking	1.07 (1.02, 1.12) **	0.95 (0.85, 1.07)
Home drinking	1.08 (1.04, 1.11) ***	1.02 (0.97, 1.08)
Friends drop over	0.96 (0.93, 0.99) **	0.99 (0.92, 1.06)
Overall drinking volume	1.00 (0.99, 1.00)	0.99 (1.00, 1.00) *
# of first degree relative problem drinkers	0.78 (0.55, 1.09)	2.01 (1.61, 2.51) ***
Impulsivity	1.17 (1.12, 1.23) ***	1.02 (0.97, 1.08)
Age	0.96 (0.94, 0.98) ***	0.98 (0.96, 1.00) *
Minority race/ethnicity	2.29 (1.41, 3.70) **	0.38 (0.19, 0.78) **
Married/living with someone	1.43 (0.83, 2.45)	2.64 (1.45, 4.81) **
LS than HS education	2.07 (1.16, 3.66) *	0.21 (0.05, 0.85) *
Unemployed	0.32 (0.18, 0.56) ***	0.98 (0.55, 1.76)
Income (referent >$60,000)	*	*
Income missing	2.16 (0.87, 5.37)	0.18 (0.04, 0.86) *
<$20,000	2.47 (1.31, 4.68) **	0.33 (0.13, 0.80) *
$20,001-$60,000	1.46 (0.79, 2.70)	0.86 (0.49, 1.52)
Constant	0.02***	0.06***

^*p < .05;

^**p < .01;

^***p < .001

Note. C.I. = Confidence interval.

We also ran a sensitivity analysis for the family/marital harms, restricting the sample to those who were married (results available upon request). Similar to the main analysis, for married women, bar drinking was significantly associated with more family/marital problems due to someone else’s drinking as was having a family history of alcohol problems and impulsivity. Drinking during a quiet evening at home was significantly associated with more marital problems among married women. Women’s overall drinking and context-specific drinking were not significantly associated with financial harm due to someone else’s drinking among married women. Results were consistent also for men, with family history of alcohol problems being positively and significantly associated with marital/financial harm in the subsample of married men. However, bar drinking was not significantly associated with marital/financial harm among married men.

Socio-demographics and harms

As shown in Table 2, for women, age was inversely associated with experiencing each of the harms from someone else’s drinking so that younger women reported more of these harms. Compared to White women, minority women were significantly less likely to report financial harm because of someone else’s drinking. Married/cohabitating women were significantly less likely to experience all types of harm from other drinkers, except for assault (where no significant difference was found), as compared to women who were not married or cohabitating. Compared to high income women, low-income women were significantly more likely to report assault by someone who had been drinking. Women’s overall drinking was not significantly associated with harms from others’ drinking. Women with more first degree relatives with alcohol problems were significantly more likely to experience both assault and marital problems (which included more general family problems) due to someone else’s drinking. Women’s impulsivity significantly increased the likelihood of experiencing assault and marital problems from other drinkers.

As shown in Table 3 for men, age was inversely associated with all types of harm from other drinkers. Compared to Whites, minority men experienced significantly more assault but less marital/financial problems. Married/cohabitating men reported more marital/financial problems from someone else’s drinking than men who were not married or cohabitating. Compared to men with more education, men with less than a high school diploma were significantly more likely to report assault but less likely to report marital/financial problems from other drinkers. Notably, unemployed men were significantly less likely than employed men to experience assault, whereas lower income men were significantly more likely to experience assault from other drinkers. Men’s overall drinking was associated with more alcohol-related assault. Similar to the findings for women, a family history of alcohol problems was associated with significantly more marital/financial problems from someone else’s drinking, and impulsivity was associated with more assault.

DISCUSSION

Our models, while not causal, included drinking contexts simultaneously to examine how the amount of drinking in each setting was associated with harm (thus assessing relative contributions of each context). While bar drinking increased men’s likelihood of assault, for women, drinking when friends dropped over increased the relative risk of experiencing assault from someone else’s drinking. For men, drinking when friends dropped over was negatively associated with assault, which indicates that assault by someone else who has been drinking may be distinct for men versus women. Men appear to be more likely to be assaulted in bar fights, whereas women may be more often assaulted by other drinkers in more private settings. It also could be that assault among women also involves sexual assault or other harms from partners or acquaintances rather than fights with strangers (more typical of men), but we do not have the data to more directly examine this possibility.

Additionally, findings indicate that bar drinking may contribute to conflict in the home and intimate relations in the family, as bar drinking increased the likelihood of marital problems due to other drinkers for both men and women. Family histories of alcohol problems increased family problems due to someone else’s drinking for both men and women, which suggests that early life experiences with problem drinkers could emerge as problems later in the life course in adult relationships. This is consistent with prior research linking family histories of alcohol problems with drinking and negative mental health outcomes, which could potentially increase both conflict within the family as well as the likelihood of continued exposure to other drinkers in interpersonal relationships (Salom et al., 2015).

For women, drinking at a party at someone else’s home was associated with less marital problems related to other drinkers, which suggests that this measure might be tapping into a more healthy lifestyle where women may drink to socialize with other women or drink with partners in a safer setting. Prior research using the 1984 National Alcohol Survey found that White women drank more often than Black women at parties at someone else’s homes whereas Black women drank more often when friends dropped over (Herd & Grube, 1993). In that study, race/ethnicity had indirect effects on drinking consequences through situational norms and drinking context, which suggests that drinking context might also help to explain racial/ethnic differences in alcohol’s harm to others (Herd & Grube, 1993). However, drinking when friends dropped over and during a quiet evening at home were combined in that analysis. Explanatory models which include both situational norms and drinking contexts to examine racial/ethnic and gender disparities in the consequences of alcohol use remain underdeveloped two decades after that study. A goal of our study was to contribute to this area of research so that when drinking context and norms are examined in the future, the social patterning of exposure to harm as it relates to context-specific drinking in both public and private settings across different domains of harms can be better integrated into analysis. We did find racial/ethnic differences in the relative risk of harms for women and men, which can be examined in future studies.

Further, the finding that men’s drinking at home was associated with more assault could be informed by future research integrating the source of harm with context. Our findings are consistent with prior research suggesting that those who drink heavily in one setting might drink more heavily across different settings (Nyaronga et al., 2009). Recent research from Australia indicate that a majority of drinking occurs in the home (Callinan et al., 2016). It also is possible that men who drink more at home are doing so with a heavy-drinking partner, which may increase their risk of assault.

Additional research is needed on drinking in private settings and risk of harm. In addition to helping to understand some of the results for men, this research could clarify why drinking when friends drop over might be associated with financial problems among women. Drinking when friends drop over also was associated with more assault among women. Perhaps experiencing financial problems overlaps with assault experienced by women — both harms may stem from women’s intimate partners.

Finally, more work is needed to develop interventions that can identify the source of harm in private settings and to provide rapid-response intervention. Promising areas for intervention might be trainings to help identify the signs of escalations in aggression and mobile technologies connecting those at risk of harm with an intermediary person to neutralize a potentially harmful situation. Interventions in bar settings have shown promise {Van Beurden et al., 2000)} – correspondingly, interventions in private settings could be helpful in reducing harm.

Overall, our results suggest that women and men’s drinking not only has distinct social and normative meanings, but also that harms stemming from someone else’s drinking are gendered and should be further examined in relation to the context within which drinking and harms occur. This is consistent with prior research finding that while men experience more assaults attributed to other drinkers than women, the patterning of harms is nuanced and can vary by drinking context and relationship to the harmer (Wells & Thompson, 2009;Greenfield et al., 2009;Karriker-Jaffe & Greenfield, 2014). The inclusion of marital and financial harms in this study extends prior research focusing primarily on assault, and it suggests that for both women and men, bar drinking can increase conflict involving other drinkers within the family. The examination of a broad range of alcohol-related harms provided a more holistic understanding of how drinking context can affect exposure to such harms. Multi-level models incorporating neighborhood factors with information on different drinking venues could further extend research on the social context of alcohol’s harm to others (Karriker-Jaffe & Greenfield, 2014).

Limitations

The data are cross-sectional, thus causal direction could not be determined. In the 2010 National Alcohol Survey, the relationship to the harmer and location of alcohol-related incidents was not assessed. Despite these limitations, this analysis of alcohol’s harm from others among a U.S. national sample indicates that context-specific drinking and alcohol’s harm from others involve both shared and distinct patterns for women and men.

CONCLUSIONS

Drinking in different social contexts may place women and men at an increased risk for a diverse range of alcohol-related harms from others’ drinking, even when taking into account their overall drinking. The characteristics of alcohol-related assault appear to differ for women and men, so that women may be more at risk of assault from intimate partners’ drinking and that of those in their immediate social networks, whereas men may be at a higher risk of assaults from strangers and newly encountered drinkers in alcohol-centered settings. Examinations of the overlap as well as the distinctions between different types of harms by gender could help to explain the gendered patterning of alcohol-related harm. Future research also is needed to identify the associations between drinking in a given context and specific instances and perpetrators of alcohol-related harm.

Prevention strategies to intervene on alcohol’s harm to others can be informed by taking drinking context, a range of alcohol-related harms, and gender into account. In addition, family history of alcohol problems and impulsivity of the victim appear to be important risk factors for exposure to alcohol’s harm from others. Identifying contextual factors that can moderate the relationships between these risk factors and harms are an important area for evidence-based prevention of alcohol’s harm to others.