Alcohol Misuse, Marital Functioning and Marital Instability: An Evidence-Based Review on Intimate Partner Violence, Marital Satisfaction and Divorce

Introduction

There have been calls for greater focus on relationships as a means of advancing public health and health outcomes¹ and marital relationships encompass one important aspect of social relationships. This is particularly impactful when considering alcohol use, as it is primarily a social behavior. Most adults drink with other people, and even those who frequently drink alone, also drink with other people.² As an intoxicant that disrupts attention, memory and other cognitive abilities, consuming alcohol, particularly at higher levels has the potential to impact social-interpersonal behaviors. There are also expectations about alcohol’s effect that can impact those behaviors. Moreover, the act of drinking to intoxication is itself a social act that has a may communicate something, such as the relative value of the relationship, the rejection of social norms about alcohol or some other specific communication. In the context of marriage, the act of drinking and the action of intoxication by both partners have been termed the “drinking partnership”,³ which impacts the relationship, sometimes in positive ways but often in deleterious ways.

The purpose of this article is to review the drinking partnership and its impact on different aspects of the marital relationship. We will begin with the extensive literature focusing on the impact of alcohol consumption on intimate partner violence. We will then consider the influence of drinking patterns on marital satisfaction and marital stability. The third section addresses the impact of marital functioning on the alcohol use of the partner. Finally, we will discuss the influence of divorce on alcohol use.

Alcohol Use and Intimate Partner Violence

Patterns of Alcohol Use and Intimate Partner Violence

Intimate partner violence (IPV) is defined as “physically, sexually, and psychologically harmful behaviors in the context of marriage, cohabitation, or any other form of union, as well as emotional and economic abuse and controlling behaviours.⁴ We confine our review to physical and psychological/verbal aggression because most of the research addressing the role of alcohol on IPV in marriage has focused on these two inter-related constructs. According to World Health Organization (WHO) estimates, worldwide, more than 1 in 4 women experience IPV in their lifetime, and 13% have experienced IPV in the past year. Estimates of the global prevalence of IPV against men are difficult to find. Archer⁵ presented data that suggests that in countries characterized by higher levels of women’s empowerment and higher levels of individuality (such as the United States and much of Western Europe), the overall rates of violence against men and women are roughly equal, although women are routinely subjected to more severe and injurious violence. In countries with lower empowerment and individuality, IPV by men exceeds IPV by women.

The Distal Relationship Between Patterns of Alcohol Use and IPV

Of all the studies addressing excessive alcohol use and marital problems, the vast majority focus on IPV. Indeed, there have been so many studies that there are now meta-analyses of meta-analyses.⁶ Studies of the alcohol–violence relationship can be divided between distal studies, those that examine patterns of alcohol consumption over some extended time period, and proximal studies, those that examine alcohol use in the time period shortly before the occurrence of IPV. Although explanations regarding alcohol and IPV have usually focused on proximal association, most of the research has examined the cross-sectional relationship between patterns of alcohol consumption and the occurrence of at least one episode of partner violence. As with the worldwide prevalence data, there is considerably less research across cultures regarding the impact of women’s drinking on IPV against men.

Research has been conducted throughout the world with extraordinarily consistent results showing male alcohol consumption is a moderately strong and pervasive risk factor for IPV against women. The WHO⁷ multi-country study (Brazil, Cambodia, Canada, Chile, Colombia, Costa Rica, El Salvador, India, Indonesia, South Africa, Spain, and Venezuela) found that men’s alcohol use was an independent predictor of women’s IPV experience. Johnson and Kishor⁸ found a significant relationship in all six countries that they surveyed in which partner drinking was assessed: Cambodia, Colombia, Dominican Republic, Haiti, Nicaragua, and Peru. A study of all 28-member states of the European Union assessed 42,000 women and found that men’s drunkenness was predictive of IPV against women in all but three of the countries.⁹ Finally, three studies of women across many African countries have found a significant relationship between partner alcohol use and IPV.^10–12 At this point, we may conclude that men’s excessive alcohol use is associated with IPV against women throughout the world with few exceptions.

The consistency observed in the worldwide studies is also observed in meta-analyses of survey studies, mostly conducted in the United States and Canada. Dating back to 1985,¹³ excessive alcohol use has been identified as a risk factor for aggressive behavior and IPV in numerous meta-analyses. In fact, in 2017, Duke et al,⁶ conducted a meta-analysis of 32 meta-analyses of alcohol, drugs and violence. The association of alcohol use and violence perpetration was significant and had a medium effect size. Importantly, although this study found a significant relationship for both men’s and women’s alcohol use, the association was stronger for male perpetration than for female perpetration, a finding replicated in a more recent meta-analysis.¹⁴

While studies have examined the impact of both men’s and women’s alcohol use, there have been a few that have examined the configuration of couple’s drinking patterns, or the drinking partnership, as it is related to IPV. In one of the earliest studies addressing drinking patterns and IPV, Leadley et al¹⁵ evaluated the impact of drinking partnerships on relationship satisfaction and IPV in a large national sample of married and cohabiting couples. They found that discrepant drinking patterns were correlated with relationship distress and IPV. While the finding regarding discrepant drinking and relationship distress has been replicated several times (a topic discussed below), the relationship between discrepant drinking and IPV is more complicated. For example, Cunradi et al¹⁶ assessed couples with respect to alcohol, cannabis, and tobacco and found that discrepant drinking couples in which the wife is the heavy drinker were the only couple type to have elevated risk for IPV controlling for other cannabis and tobacco use by the couples as well as other control variables. From a somewhat different perspective, Testa and colleagues¹⁷ examined husband and wife alcohol dependence (as scored by the Alcohol Dependence Scale, ADS).¹⁸ The likelihood of at least one episode of husband perpetrated IPV was increased by high scores on the ADS for either husband or wife or both. When the criterion variable was the frequency of IPV, high scores by either husband or wife were again predictive, but couples in which both had high ADS scores demonstrated a substantially higher frequency of husband-perpetrated IPV episodes.

There have also been several studies demonstrating that the pattern of drinking is predictive of subsequent IPV. In general, studies that have examined this association over a short period of time have supported an association, while studies with a longer period of time have not, perhaps because the drinking patterns changed over the course of the study. In one of the earliest studies, Leonard and Senchak¹⁹ found that husbands’ scores on a measure of alcohol dependence at the time of marriage predicted husband aggression to wife in the first year of marriage. There is also evidence for a prospective relationship over the transition to parenthood,²⁰ and throughout the early years of marriage.²¹ From a more clinical perspective, there has also been evidence of a longitudinal relationship among domestic violence offenders²² and among men in treatment for alcohol use disorders.²³

The Proximal Relationship Between Alcohol Use and IPV

Evidence for a proximal relationship between excessive alcohol use and IPV arises from two different research approaches, experimental studies of alcohol and aggressive behavior and studies of the presence/absence of drinking and the occurrence of violence.

Experiments have provided contexts for participants to behave in a physically or verbally aggressive manner after receiving alcohol, a placebo beverage, or no alcohol. Studies of physical aggression have typically involved one participant and one confederate who participate in a task that allows for one or both to administer different levels of noise or electric shock (eg aggression) to the other person. Early meta-analyses of these studies concluded that alcohol administration led to higher levels of aggression than no alcohol or placebo and placebo did not differ from no alcohol conditions.²⁴ A later meta-analysis replicated the impact of alcohol and found that higher doses of alcohol led to more aggression than lower doses.²⁵ The vast majority of these studies focused on male participants interacting with other males in the task. However, in recent years studies have utilized these paradigms with couples to demonstrate that alcohol increases aggression towards one’s intimate partner, although often in interaction with situational and personality variables. For example, Watkins et al²⁶ found that alcohol increased the level of noise that male participants administered to their intimate female partner under provocation. In a somewhat similar study, Subramani et al²⁷ allowed heavy drinking men and women with a recent history of marital aggression to compete and set shocks for their partner. They found that alcohol administration led to higher shock settings than the no alcohol control, but this effect was present primarily among the more problematic drinkers in this heavy drinking sample. More recently, surveying all the experimental approaches studying alcohol and male to female aggression, Crane et al²⁸ concluded that alcohol did indeed increase male to female aggression.

Among some of the experiments included by Crane et al²⁸ were several that involved a more ecologically appealing paradigm of aggression, conflictual discussions between married or cohabiting opposite sex couples. Although many of these have found that alcohol increases negative conflictual behaviors (interrupting, disagreeing, insulting), this has not uniformly been the case. For example, Jacob and Krahn²⁹ allowed “alcoholic”, “depressed”, and control men and their wives to discuss and work toward resolving conflictual issues, once when the couple was allowed access to their typical alcoholic beverage and once in the absence of alcohol. Alcoholic men and their wives exhibited more negative conflict behaviors than control couples. There was also a trend toward more negativity among the alcoholic men and their wives on a drink night than on a no drink night. Leonard and Roberts³⁰ allowed young married couples in which the husband had or had not engaged in marital aggression to participate in similar conflict discussions. All the couples engaged in one discussion in which no alcohol was provided. Then, the husbands were randomly assigned to receive alcohol, a placebo, or no alcohol. Aggressive couples engaged in more negative behaviors, and all couples increased their negative behaviors when the husband received alcohol. Couples in which the husband received a placebo or no alcohol for the second discussion, did not increase their negativity. Finally, Testa et al³¹ conducted a similar study, however after an initial no alcohol condition, either the husband, the wife, neither or both received an intoxicating dose of alcohol and then discussed the second conflict area. In contrast to the Leonard and Roberts³⁰ study, alcohol led to decreases in negativity. One possible reason for this is that the Testa et al³¹ study required both husbands and wives to acknowledge at least monthly consumption of 5 drinks (4 for women) on a single occasion. This was not a requirement for the Leonard and Roberts³⁰ study. As a result, the Testa et al³¹ participants may have been largely congruent, heavy drinkers, while the Leonard and Roberts³⁰ participants were likely incongruent drinkers. In later sections of this paper, we discuss the implications of this difference further, but for now we will note that congruent drinking is associated with better outcomes than incongruent drinking patterns.

One final study design involves event-based analyses comparing the risk for IPV following alcohol consumption with the risk for IPV at times during which alcohol was not consumed. These event-based studies include retrospective reports of conflict situations, as well as daily diary studies tracking alcohol consumption and conflict interactions. In the earliest of these, Leonard and Quigley³² found that husbands, but not wives, were more likely to be drinking during the most serious physical aggressive episode than during the most serious verbal episode. Murphy et al³³ studied alcoholic men and their partners and found that conflict situations involving physical aggression involved higher estimated blood alcohol levels for husbands than conflict situations that did not involve alcohol. In a daily diary study, Testa and Derrick³¹ reported that the likelihood of a physically aggressive episode was increased nearly fourfold when the perpetrator had been drinking within the previous four hours. Surprisingly, Testa et al,³⁴ using the same subjects, found that alcohol consumption was linked to increased emotionally intimate behaviors. To resolve this apparent discrepancy, Derrick et al³⁵ examined between and within subjects’ moderators of aggression and intimacy outcomes. The results suggested that alcohol increased aggression for couples with high conflict in the past 30 days and following times of relationship discord earlier in the day. Alcohol increased intimacy for couples with more intimate interaction and lighter drinking in the past 30 days, and lighter drinking prior to the events.

Moderators of the Alcohol-IPV Relationship

From the beginning of alcohol-violence research, it has been acknowledged that any effect of alcohol on violence would be stronger for some individuals, and perhaps null for other individuals. However, research into moderating factors in the distal relationship of alcohol and IPV was initially slow to develop. Within the experimental research, there was considerable interest in moderation. This research suggested that aggression facilitating personality constructs such as hostility and anger³⁶ and situational factors such as social pressure to aggress³⁷ interacted with alcohol to predict aggressive behavior in lab studies, although, as noted, these studies largely focused on male-to-male violence. Moderation has also been a major focus in recent studies of couples in lab studies. As noted above, Watkins et al²⁶ found that alcohol did not increase aggression in the absence of provocation but did increase aggression following provocation. Subramani et al²⁷ found that alcohol increased aggression for problematic heavy drinkers but not for heavy drinkers with minimal alcohol problems. Further analyses reported moderation by trait anger and psychological flexibility³⁸ and emotion differentiation.³⁹

In terms of moderation for the distal relationship, the earliest research supported moderation for relationship issues such as marital satisfaction⁴⁰ and several individual factors related to aggression more generally such as hostility⁴¹ and stress.⁴⁰ In the ensuing years, the number of constructs with evidence of moderation has expanded tremendously, and has included constructs such as jealousy,⁴² coping style⁴³ and psychopathology.⁴⁴ In general, much of the moderation literature suggests that alcohol and constructs that are associated with hostility or impulsivity have a synergistic relationship predicting violence. This has been interpreted within the dominant cue⁴⁵ and alcohol myopia model of alcohol and aggression.⁴⁶ These models suggest that by impairing cognition, alcohol leads individuals to be guided by the most salient factors in the situation, with salience being influenced by both situational and personality factors. However, some studies have found moderation effects suggesting that among individuals with very high hostile tendencies, the occurrence of IPV is not increased by alcohol.^47,48 This has been interpreted within inhibition conflict,¹³ multiple thresholds⁴⁹ and “I^3” or “I-cubed” models of aggression.⁵⁰ Briefly, these models suggest that aggression occurs when aggression-facilitating factors overcome the threshold imposed by aggression-inhibiting factors, and that alcohol has its strongest effect for individuals and situations where the facilitating and inhibiting factors are relatively equal. This suggests that IPV-moderating factors may differ by the nature of the sample of individuals being studied and by the other contextual aspects of the relationship within and around the couple.

Alcohol Use and Marital Satisfaction

Inasmuch as research has supported the role of excessive alcohol consumption as a contributing cause of IPV, one might assume that a similar relationship would hold for marital satisfaction and marital stability. To be sure, IPV does predict marital satisfaction. However, the effect is small to moderate.⁵¹ Moreover, as noted above, the acute effects of alcohol are temporally associated with both aggression and intimacy. In addition, many factors contribute to the association between marital satisfaction and health outcomes, and this is a relationship with bidirectional causality.⁵² That is, whereas marital satisfaction predicts substance use behaviors, substance use also predicts marital satisfaction.⁵³ Marital satisfaction is impacted both by positive features—such as responsivity or intimacy—and by negative features—such as conflict or aggression. However, research is unclear on whether the positive features, or negative features, are more crucial to driving this association.⁵³

It is clear that marital satisfaction is low among couples in which one member has an Alcohol Use Disorder (AUD).^54–56 However, determining the precise role of AUD has been more elusive for several reasons. First, there is substantial comorbidity between AUD and other disorders including mood disorders (anxiety, depression) and personality disorders (eg antisocial) which also impact marital satisfaction. For example, Whisman⁵⁷ (1999) found that both mood and anxiety disorders were independently associated with lower levels of marital satisfaction. However, he did not find a relationship between AUD and marital happiness after controlling for these disorders. In fact, after controlling for mood and anxiety disorders, even the broader category of substance use disorder was not significantly related to marital happiness. Second, it is common that spouses of those with an AUD also have psychiatric disorders, including AUD, a fact which much of the earlier research neglected.⁵⁸ Third, there is some research that suggests that the effect of AUD on marital satisfaction may differ between men and women. For example, Cranford and colleagues⁵⁹ (2011) found that wives’ lifetime AUD had a direct, negative association with their own, and their husbands’ longitudinal marital satisfaction. However, the same pattern of findings did not hold for husbands’ lifetime AUD. In sum, while it seems probable that AUDs negatively impact marital satisfaction, research has not been entirely consistent.

One of the more consistent findings with respect to excessive drinking and marital satisfaction has been that concordance of drinking within the couple is of critical importance. Early research by Roberts and Leonard³ found that couples who drank frequently and together were as satisfied as couples who were light social drinkers, and more satisfied that couples in which the husband was a heavy drinker and the wife was not. Examining concordance more directly, Mudar et al⁶⁰ found that couples in which both reported frequent intoxication and couples in which both reported rare intoxication had similarly high marital satisfaction scores, which were significantly higher than couples in which only one reported frequent intoxication. Interestingly, concordance with respect to frequency and quantity of alcohol consumption has been found related to higher marital satisfaction in a large nationally representative sample in New Zealand.⁶¹ This relationship has also been reported among female couples⁶² and young male couples.⁶³ To some extent, it may be that maritally satisfied couples are more likely to simply do things together. Indeed, Homish and Leonard⁶⁴ found that it was concordant drinkers who drank together experienced higher satisfaction, but not those concordant drinkers who drank apart. So, there is some validity to this notion. Nevertheless, there is longitudinal data that suggests that concordant drinking is more than simply a marker of marital satisfaction. In a longitudinal study, Homish and Leonard⁶⁵ found that discrepant levels of heavy drinking were predictive of lower marital satisfaction. Foulstone et al⁶⁶ also found that discrepant drinking was longitudinally predictive of lower levels of marital satisfaction, an effect that was stronger among couples with children. Although less commonly studied, discrepancies with respect to alcohol problems have also been linked to worse relationship functioning.⁶⁷

Although concordant drinking patterns are associated with higher marital satisfaction among non clinical samples, the role of concordant drinking patterns among those with AUD is less clear. One study has examined this specifically and has evaluated the relationship concurrently and longitudinally. Mattson, Lofgreen and O’Farrell⁶⁸ assessed a sample of couples in which the man was seeking treatment for an AUD. They found that higher conjoint frequencies of intoxication at baseline were associated with higher marital satisfaction. However, in contrast to non-clinical samples, higher conjoint frequencies were longitudinally predictive of declines in marital satisfaction. This suggests that the relationship between satisfaction and concordant-discordant intoxication may break down among couples with serious alcohol problems and/or over time. More research is necessary to examine this issue.

The mechanisms by which specific elements of marital satisfaction and related behaviors are shaped substance use have become a more predominant focal area in recent years. Not only is the concordance/discordance important but recent work has identified the perceptions of one’s partner’s drinking as influential on marital satisfaction. For example, perceiving a partner to be a heavy drinker, after controlling for actual alcohol use and alcohol-related consequences, is associated with lower levels of marital satisfaction. There was also a moderating role of gender, such that perceiving partner drinking as problematic resulted in poorer marital quality for men, but not for women.⁶⁹

In a review by Rodriguez et al,⁷⁰ moderators and mediators of the relationship between alcohol use and relationships (defined in their review as “marital distress”) are identified. Gender is a primary moderator, as there are differences in rates of diagnosis, reasons for use, treatment, and relapse between men and women. In particular, women are more likely to cite marital problems as a reason for drinking and are more likely to relapse from Alcohol Use Disorder (AUD) with a partner. Men, however, are more likely to relapse from AUD when alone.⁷⁰ Windle and Windle⁷¹ build on this by identifying partner conflict and support as moderators of alcohol use and marital satisfaction. In this work, alcohol use with higher perceived conflict was associated with more alcohol problems, and this pattern of alcohol use and perceived conflict was related to marital satisfaction for men. In contrast, alcohol use with higher perceived partner support was associated with fewer alcohol problems and higher marital satisfaction for both men and women. With respect to mediators, Rodriguez and colleagues⁷² identify three primary ones in the relationship between alcohol use and marital distress: negative partner interactions, drinking to cope, and negative perceived partner support. More recent research has identified the role of regulation strategies as a mediator; punishing behaviors, intended to decrease drinking or enhance the relationship dynamics, in fact have a damaging effect on the relationship.⁷¹ This is an important emerging area in which more research is clearly needed.

A second emerging trend has been expanding the nature of the populations of married couples. Whereas the research described above has been conducted among civilians, many of these findings have been replicated in military-affiliated samples, as well. Partners of military service members who show concern over their partner drinking were more likely to have poor relationship functioning.⁷³ In a sample of US Reserve and National Guard (USAR/NG) soldiers and their partners, increasing marital satisfaction was significantly associated with fewer alcohol problems.⁷⁴ Marital satisfaction also buffers the effects of combat exposure on alcohol problems among USAR/NG soldiers, wherein soldiers who perceived combat exposure to be moderate-to-highly traumatic were less likely to have alcohol problems in the presence of high marital satisfaction.⁷⁵ In contrast, there are mixed findings on the effects of marital satisfaction in samples of never-deployed USAR/NG soldiers. One study demonstrated that marital satisfaction was not significantly associated with alcohol problems, frequency of getting drunk, typical number of drinks, or percent of drinking days among never-deployed soldiers.⁷⁶ However, more recent research shows that marital satisfaction is a significant moderator in the relationship between non-deployment emotions (ie, a measure of guilt, perceived value, connectedness, and camaraderie) and alcohol problems.⁷⁷

Much of the prior research on marital satisfaction and alcohol use has focused on younger populations and/or newly married couples.^74,75,77 More recent research, however, has focused on this relationship among older adult couples. Birditt and colleagues⁷⁸ found that drinking concordance among older adults was associated with increased blood pressure over time, particularly in the presence of negative marital satisfaction. Building on this, a subsequent paper found that light drinking and concordant drinking were associated with longer lives among older adult couples.⁷⁹ Additional data from couples 50 years and older in the Health and Retirement study provides evidence that the relationship between alcohol use and marital satisfaction varies by gender, with older adult couples increasing their drinking in the presence of women’s reports of positive marital satisfaction.⁸⁰ Research by Stewart and colleagues⁸¹ also underscores the roles of age and gender as moderators in the relationship between marital satisfaction and alcohol use. Understanding the interplay in drinking concordance, alcohol use, marital satisfaction, and health outcomes is particularly salient given the increasing prevalence of binge drinking and AUD among older adults.⁷⁹ Further, the findings among older adults may indicate that the patterns observed among younger couples may be sustained into older age. Certainly, there is room for an improved understanding of this relationship.

A final direction that has emerged in recent years is the examination of the bidirectional nature of the relationship between alcohol use and marital satisfaction. Interdependence theory provides a theoretical foundation for this bidirectional relationship; it emphasizes the idea that, whereas individuals are important, the relationships between individuals are also just as important to examine. A small but growing literature applies interdependence theory to the relationship between alcohol use and marital satisfaction. Research using interdependence theory indicates that alcohol and marital problems exacerbate one another.⁶⁹ Patterns of substance use are important predictors of changes in marital functioning over time. Rodriguez and colleagues⁶⁹ point out that most research teams are more comfortable with modeling intrapersonal effects and less comfortable with interdependent effects. Focusing solely on intrapersonal effects, however, loses data from the dynamic interaction between partners. Thus, future research should collect data from both partners in a relationship and utilize techniques such as Actor-Partner Interdependence Model to elucidate the dynamic and complex relationship between alcohol use and marital satisfaction. Whereas dyadic perspectives are increasingly common in relationship research, it is less common in substance use research, even when the focus is on the effects of substance use in a relational context.

It should be noted that marital satisfaction is measured in a number of ways, and the primary articles that are discussed in this review use different measures. For example, some studies used the Marital Adjustment Test (MAT)⁸² including those focused on military samples.^74–77 This 15-item instrument measures overall marital satisfaction and adjustment of husbands and wives to each other and total scores range from 2 to 158. Other studies, however, used a variety of other measures, including: the 4-item Couples Satisfaction Index,⁸³ the 6-item Quality of Marriage Index (QMI),⁸⁴ and the 3-item Kansas Marital Satisfaction Scale (KMSS).⁸⁵ These brief scales are largely unidimensional and consequently appear to assess the very broad construct of satisfaction, without tapping more precise constructs involved in marital functioning. The Health and Retirement Study recommends the use of two separate variables; a 3-item scale for positive marital quality and a 4-item scale for negative marital quality.^80,86 As noted above, Bulanda, Curl and Roberts⁸⁰ found among older couples, wife’s positive marital quality was related to increased heavy drinking over time for both wives and husbands. Therefore, differences in the nuances of the operational definition and measurement of marital satisfaction may be underlying the results. Future research should utilize more detailed measures of marital satisfaction and marital functioning.

Finally, this current review focuses primarily on non-clinical samples; it is important to note that there is a literature base indicating that improved marital functioning leads to reduced drinking,⁸⁷ but this is beyond the scope of this review.

The Influence of Alcohol on Divorce

All Reasons for Divorce

Understanding risk factors that destabilize the marriage and contribute to marital dissolution may help clarify the role of alcohol in marital dynamics. Alcohol’s impact on marital functioning and subsequent divorce may be complicated by the complexity and nuance within the relationship literature. The risk for divorce may be moderated by religious differences, employment status, psychiatric status, age at marriage, or substance use other than alcohol.^88–91 Excessive drinking has been identified as a causal factor for divorce in studies over several decades. A study sample of divorces filed between 1937 and 1950 reported excessive alcohol consumption as the causal factor for marital dissolution in 21% of cases.⁹² Comparably, in a qualitative analysis of data collected between 1980 and 1997, Amato and Previti⁹³ found drinking and drug use to be the third most common reason for divorce, behind infidelity and incompatibility. Because these studies rely on open-ended questions about the reasons for divorce, it is difficult to determine whether excessive drinking, problems arising from drinking, or AUD all impact divorce.

Some studies assessing predictors of divorce have drawn conflicting conclusions regarding associations between alcohol use and divorce. For example, Fu and Goldmann’s⁹⁴ longitudinal analysis suggests no association between excessive drinking and divorce, while smoking emerged as a significant risk to marriages in the study sample. Similarly, Sanchez and Gager⁹⁵ identified no correlation between husbands’ alcohol use and marital dissolution. Qualitative analyses aiming to determine the causal factors of divorce may be distorted by the subjectivity of study participants, as partners tend to feel pressured to deflect blame onto the opposite partner.⁹⁶ Additionally, contradictions in study findings may be explained by the absence of an AUD diagnosis in one or both partners, discordant drinking patterns, and genetic risk factors for divorce.

Alcohol Use Disorder (AUD)

Individuals who meet the criteria for AUD may be at a higher risk of experiencing divorce compared to those without an AUD diagnosis. Several studies have associated excessive alcohol consumption resulting in intoxication with marital disruption and divorce,^84,88,91 and AUD has been identified as an independent risk factor for marital dissolution, regardless of other substance use or psychiatric comorbidities.^88,92 For example, Cranford⁸⁸ found that the rate of divorce for study participants with an AUD diagnosis meeting DSM-IV criteria was 18.2% higher compared to those with no lifetime AUD diagnosis. Demographic differences, such as ancestry and ethnicity, may moderate AUD as a predictor for divorce. The association between AUD and marital dissolution may be more significant for people of European ancestry compared to people of African ancestry.⁸⁹

Concordant Vs Discordant Drinking

Alcohol use in married couples may contribute to marital dissatisfaction and, ultimately, marital dissolution, especially within couples whose drinking habits contradict one another. As described earlier, research on excessive drinking and marital satisfaction consistently has found that the extent of excessive drinking is less relevant to marital satisfaction than the match or mismatch of the couple’s individual drinking patterns. Accordingly, discordant drinking patterns may be the critical predictive variable in marital dissolution.^97–99 More specifically, Ostermann et al⁹⁹ found that couples with one heavy drinker were more likely to divorce than spouses who both drank heavily or abstained from alcohol. In this middle-age sample, discordant drinking predicted divorce even after controlling for prior marriages, number of years married, and alcohol problems. Torvik et al’s⁹⁷ study was conducted with a sample of middle-aged Norwegians. They found a reduced risk of divorce among concordant abstainers and concordant heavy drinkers. In addition, they found that among discrepant drinkers, couples with a heavy drinking wife and light drinking husband were at the greatest risk. They also found that concordant abstainers had a lower risk than concordant heavy drinkers. The final study, Leonard et al⁹⁸ followed a younger, newlywed sample for nine years. They found that discordant heavy drinking was predictive of separation/divorce after controlling for antisocial behaviors, depression, employment and parental status. Notably discordant cannabis and discordant tobacco use were not predictive. This study did not find a particular risk for discordant couples in which the wife was a heavy drinker, nor a difference between concordant heavy drinkers and concordant non-heavy drinkers as observed by Torvik and colleagues. However, Leonard et al⁹⁸ findings were in the same direction as Torvik et al’s,⁹⁷ likely non-significant due to a smaller sample size.

In short, the deleterious effect of discordant drinking has been observed in newly-married couples as well as couples well into their marriage. It further appears that in opposite-sex couples, the risk of marital dissolution may be moderated by which partner, husband or wife, consumes more alcohol. Finally, these effects also hold when statistical controls are applied to a variety of other important predictors of divorce. Complementary to the findings in these studies, concordance within married couples in other domains, such as religion or shared phobias, also appears to have a protective effect on improving marital quality and reducing the risk of divorce.^89,100

Environmental/ Genetic Risk Factors

Genetic and environmental factors have been shown to influence the probability of individuals getting married and divorced.¹⁰¹ Some evidence suggests that heritable traits may indicate a higher risk of divorce linked to alcohol dependence. Waldron et al¹⁰² measured marital experiences in twin pairs to determine the covariation between alcohol dependence and marital dissolution. Results from this study indicate that genetics played an important role in this association and heritable traits were predictive of early marital separation. Similarly, Jerskey et al¹⁰¹ point out that environmental and genetic risk factors for early marital separation may be independent because the concordance of behaviors and beliefs may be protective, and the same factors in discordance increase the risk of divorce. They also suggest that particular personality traits, such as low harm avoidance, may be associated with an increased risk of AUD and increased marital dysfunction.

Salvatore et al¹⁰³ measured the correlation of genetic and environmental risk factors on AUD and divorce by comparing the experiences of twins, similar to the study design by Waldron and others.¹⁰² In a Swedish study sample born between 1940–1965, data analyses revealed statistically significant covariation between AUD and marital dissolution as a function of genetic and environmental influences. In a recent study, Thomas et al⁸⁹ tested the association between AUD and divorce in a high-risk population. Results from this study contradict the previous literature with no genetic associations identified as predictors for divorce; however, environmental background was identified as a risk factor for divorce. The majority of the literature regarding alcohol and divorce focuses on¹ marriages between a “husband” and a “wife”, and few studies have evaluated the influence of alcohol in same-sex marriages and LGBTQ+-identifying individuals. Moreover, serious alcohol problems or an alcohol use disorder may delay the age at first marriage and reduce the length of time the marriage survives.¹⁰² While out of the scope of this review, it should be noted that early age at marriage has been associated with a higher risk of divorce.⁹⁰ It is unclear if delays in the first marriage related to AUD are protective or predictive of future marital dissolution.

The Influence of Divorce on Alcohol Use Disorder (AUD)

Early survey research documented that divorced men and women were at greater risk for excessive drinking and alcohol problems than were married men and women.¹⁰⁴ As discussed in a previous section, this may reflect the deleterious effect of excessive drinking on marital stability. However, inasmuch as marriage and marital satisfaction are protective for excessive alcohol use and alcohol use disorder, the removal of these protective factors through divorce may lead to increased drinking and alcohol problems. It is also the case that alcohol is used to cope with stressful or difficult experiences, and divorce would certainly qualify as both for many. In addition, there are a number of confounding factors related to alcohol use as well as declining relationship satisfaction that are both independently related to these outcomes, such as family background, economic instability, as well as personal history of alcohol use/alcohol use disorder. Similar to findings discussed in previous sections, there is the clear likelihood that bidirectional and confounding processes may result in a relationship between divorce and excessive drinking, and parsing the different elements of this relationship is not straightforward. Consequently, cross-sectional studies of this relationship are not necessarily informative.

Fortunately, with strong epidemiological-style research studies, multiple waves of data address this challenge, particularly as studies have followed couples through divorce and after. For example, Kendler et al¹⁰⁵ found that divorce was strongly associated with the onset of an alcohol use disorder for both men (hazard ratio [HR] = 5.98) and women (HR = 7.29). Further, they modeled the temporal relationship between divorce and first-onset of AUD among both men and women. These figures demonstrate for both men and women that the prevalence of AUD among those who will become divorced increases in the five to seven years prior to divorce; they also experience a bump in the year following divorce and remain elevated for the fifteen years after divorce. Additional research supports this association, that is, divorce does increase alcohol consumption (whether heavy drinking or AUD). More specifically, divorce does increase heavy alcohol consumption among both men and women in the five years post-divorce.¹⁰⁶

There are some additional nuances to this relationship. For example, Agahi et al¹⁰⁷ used latent trajectory analysis to examine patterns of drinking prior to and after divorce. They found that only those who were heavy drinkers prior to divorce increased their drinking after divorce. Those who either were decreasing heavy drinking or did not engage in heavy drinking at all prior to divorce did not increase drinking in the post-divorce period. In a somewhat similar vein, Smith et al¹⁰⁸ argued that ending a marriage to a problem drinking spouse might have a very different effect that ending marriage to a spouse who was not a problem drinker. Using data from the first two waves of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), they found that women ending their marriage to a problem drinker had less heavy drinking and fewer alcohol problems than women remaining married to a heavy drinker. In contrast, women divorcing a man who was not a problem drinker had an higher risk of heavy drinking and AUD than those remaining married to a non-problem drinker. Because the prevalence of problem drinking was lower among women, they could not determine whether men divorcing women with alcohol problems would also experience a decline in their own alcohol use and problems.

Critically, it’s important to consider whether gender may have a more interactive effect. In heterosexual marriages, women have more rapid declines in marital satisfaction over the early years of marriage than do men.¹⁰⁹ Moreover, in studies of couples in marital therapy, women are significantly less satisfied than men.¹¹⁰ Perhaps as a consequence, women are more likely to initiate divorce than their male partners; some estimates indicate 70% of divorces are initiated by women.¹¹¹ When considering marriage and its dissolution, scholars posit that this operates within an “expected utility framework”,^112,113 or that there is some maximized utility to being married, rather than existing independently. Therefore, for marriage to end with divorce, there must be a greater advantage to the person who initiates the divorce, relative to remaining married. Indeed, research bears this out and demonstrates that for the divorce initiator, life satisfaction increases and depression decreases in the five years post-divorce.¹¹⁴ Non-initiators, as one may expect, do experience significant decreases in life satisfaction and increases in depression in the three years post-divorce.

When this is combined with gender, such that women are the more prevalent divorce initiators, and thus men would be more commonly the non-initiator and at risk for reduced life satisfaction and increased depression, it becomes even more important to consider patterns with alcohol use. This is particularly true for men who may engage in more heavy drinking to cope with stress.^115,116 In addition, women tend to have lower prevalence of AUD and hazardous drinking than men.^117,118 Although recent data suggests that men’s and women’s hazardous drinking is converging,¹¹⁹ at present, divorce appears to present a unique set of risks for men and subsequent alcohol use. This is consistent with Reczek et al¹²⁰ who used individual-level data, couple-level data, and marital histories from a longitudinal study to examine heavy alcohol use in older adults. They found that after divorce, men’s alcohol consumption increased while for women, alcohol consumption decreased.

Further, other studies have found associations between increased drinking among men. For example, more men increased heavy drinking than women, though the number of women who were drinking heavily did modestly increase,¹¹⁶ while another found those who continued heavy drinking after divorce were more likely to be men.¹¹⁷ Still others found that although divorced individuals were at risk for drinking fewer days per month, they had greater odds of at-risk drinking, even after controlling for exogenous factors like high school experiences, social participation, and living with an alcoholic (either as an adult or as a child).¹²¹ In this study, divorced individuals were also at higher odds of possible alcohol dependence, even when controlling for those same factors. When examining risks associated with divorce by gender, men were at greater odds of at-risk drinking, while women were at greater odds of having alcohol dependence in this study.

Critically for this study, they were explicit on including potential confounding factors, such as home environment. This is an important consideration as such influences can have major impact on alcohol consumption, hazardous drinking, and AUD and should be accounted for in the relationship between divorce and alcohol. Specifically, Kendler et al¹⁰⁵ did account for potential confounders and included low parental education, prior deviant behavior, and family history of AUD as covariates in their model, finding only a modest attenuation (for men, hazard ratio dropped from 5.98 to 5.09; for women, from 7.29 to 6.31).

In order to control for shared genetic liability as well as some exogenous variables, such as home environment, Dinescu et al¹²² used a same-sex twin pairs sample to examine relationships between drinking, marriage, and divorce. They found that married and cohabitating twins did not have different alcohol consumption patterns but divorced co-twins did consume more alcoholic beverages. Unfortunately, these data were cross-sectional and thus it’s unclear whether the relationship is causal (ie whether divorce caused the increased consumption). Other work has demonstrated the importance of family roles, expanding beyond marriage/divorce, to include parenting, as well as socioeconomic status, both of which may play a role in increased drinking.¹¹⁸ In addition, other environmental factors may also play a synergistic role, such as¹²³ Kendler and colleagues who found that being divorced (v. married) in a neighborhood with an alcohol outlet had significantly increased risk for AUD, compared to those without close proximity.

Another potential confounding factor is that alcohol is often used as a coping mechanism for stressful situations or difficult feelings. For example, even within marriages, marital conflict was associated with alcohol-related problems, and attachment anxiety was related to higher level of drinking to cope.¹²⁴ Being the non-initiator in a divorce may be associated with greater feelings of sadness, stress, and thus may be more likely to look for coping mechanisms, while the initiator may find divorce to be an immediate improvement in satisfaction or a reduction in stress.

Summary

For many years, the effects of drinking and alcohol problems on relationship functioning were studied from the perspective of men’s drinking. There were isolated attempts to study women’s drinking, however, the focus on drinking and drinking problems from the perspective of “drinking partnerships” evolved gradually beginning about 25 years ago. The literature reviewed in this paper demonstrates the necessity of addressing the behavior of both members of the couple. Indeed, it is difficult to imagine future research ignoring what seems like an obvious point, that dyadic behaviors and relationship trajectories may be influenced by both members of a couple. In particular, we have shown several instances in which gender has an interactive effect, and this merits much closer study.

In the context of IPV, the research continues to support the conclusion that acute intoxication is a contributing cause of IPV although it is neither a necessary nor sufficient cause. Alcohol problems may contribute to IPV because they are markers of frequent acute intoxication and because they can result in very significant family stresses. With respect to marital satisfaction and divorce, the key predictor seems to be whether the members of the couple can establish patterns of consumption that foster closer relationships. Alcohol consumption which is not excessive, and which occurs primarily when the couple is together, can serve a positive function. It is important to note that this does not necessarily mean that concordant patterns of heavier drinking are healthy or functional for the entire family.

As we have noted at various points in this review, there is a need for more focused attention on these issues among same sex couples, a need that is very slowly being addressed. Further, an emphasis on marital satisfaction may help clinicians identify the risk of IPV, marital dissatisfaction, or other deleterious outcomes, and further, may be a step towards increasing the timeliness of interventions and potentially prevent downstream health morbidities in couples. Finally, there is also a need to address this in socio-culturally diverse couples. Alcohol is used very differently and holds different cultural meanings across the world. The broad cultural understanding of alcohol must influence its impact on and within relationships.