Proximal Effects of Alcohol on Intimate Partner Violence Perpetration: Recommendations for Investigating “The How Question”

Abstract

Objective:

In this Perspectives article, we outline a collaborative research program aimed at a more refined understanding of the proximal factors involved in alcohol-facilitated intimate partner violence (IPV) perpetration.

Method:

We provide a summary of our research experiences and offer recommendations regarding studies aiming to understand what we term “the how question,” which focuses on applying research methods that allow for the investigation of the potential causal mechanisms that underlie the well-established link between heavy alcohol use and IPV perpetration.

Results:

We suggest several “gold standard” research methods that allow researchers to yield useful information about how alcohol may cause IPV perpetration, but that are infrequently applied in this research area.

Conclusions:

We conclude by reviewing several factors limiting our understanding of the proximal determinants of alcohol-facilitated IPV, including issues relating to sampling, definition, inclusiveness, and the dyadic nature of IPV.

Flashback to the year 2010: Both authors are conducting independent, yet complementary, research that generally focused on risk factors for alcohol-related interpersonal violence (e.g., Eckhardt, 2007; Eckhardt & Crane, 2008; Parrott & Giancola, 2004; 2006). This work involved a range of methodological approaches, including laboratory-based methods of alcohol administration and aggression assessment. Through some unknown motivational force, contact was made and a phone call (truly a flashback to life in 2010!) was arranged to test the waters about a more formal collaboration to examine the factors that might account for alcohol- facilitated intimate partner violence (IPV). This communication revealed a shared interest in integrating our methods to better understand what we have come to call “the how question”; that is, how does acute alcohol intoxication actually cause IPV perpetration? Our discussion of “the how question” quickly expanded to include other critical questions: What theoretical models are available to allow for a better prediction of the mechanisms that mediate this association? What are the most rigorous methods for examining those mechanisms? Can we test these research questions among individuals at high risk for IPV? Although the IPV field had made substantial progress establishing the basics of the alcohol-IPV perpetration association and identifying alcohol as a contributing cause of IPV perpetration (Leonard, 2005), we discovered that there were few studies that actually addressed “the how question.”

Since that first phone call in 2010, our research programs have employed a variety of research designs and assessment methodologies to uncover theoretically-based mediating mechanisms that might provide important causal clarifications about how alcohol actually causes IPV. In this Perspectives article, we aim to articulate some of what we have learned in hopes that it is helpful to researchers with similar interests. We first provide an overview of our current understanding of the alcohol-IPV association and review the models that aid in our understanding of these findings. Next, we describe in more detail the research methods that allow for a more mechanistic determination of alcohol’s effects on IPV (i.e., “the how question”). Finally, based on insights from our work in this area, we offer suggestions for future research.

Alcohol-Facilitated IPV Perpetration: Empirical and Theoretical Overview

A Brief Review of the Alcohol-IPV Perpetration Association

IPV is a global public health crisis that affects the lives of millions of people each year and is associated with numerous negative consequences, including mental and physical health problems, increased use of legal and housing services, and financial burden of up to US$5.8 billion annually (Black et al., 2011; Leemis et al., 2022). Acute alcohol intoxication and a pattern of heavy alcohol use are among the most robust correlates of IPV perpetration (Leonard, 2008) that remain significant even after controlling for a range of perpetrator and dyadic contextual factors (Leonard & Senchak, 1993; Pan et al., 1994). This association is consistent across research methodologies and measurement approaches (Parrott & Eckhardt, 2018), which led researchers to conclude that alcohol use is a clear contributing cause of IPV perpetration (Leonard, 2005; Leonard & Quigley, 2017).

Theoretical Models of Alcohol-Facilitated IPV Perpetration

While the field has made progress in demonstrating that alcohol is a contributing cause of IPV, we have a relatively poor understanding of how alcohol use actually functions to increase risk for IPV (Eckhardt et al., 2015). There is much to learn by shifting away from the broader question of whether heavy alcohol use relates to IPV perpetration (it does, full-stop) towards a closer look at the models and methods that can inform our understanding of the causal mechanisms that account for the alcohol-IPV association. Historically, these models fall into one of three categories: (1) the spurious model (i.e., the alcohol-IPV association is a 3^rd-variable problem; Zubretsky & Digirolamo, 1996), (2) the indirect effects model (the alcohol-IPV association is explained by downstream negative consequences of the perpetrator’s alcohol use, such as relationship conflict; Leonard & Jacob, 1988), and (3) the proximal effects model, which focuses on how alcohol pharmacodynamics impact cognition, emotion, and behavior (Bowen, 2011; Foran & O’Leary, 2008; Leonard & Quigley, 1999). The spurious and indirect effects models are not well-supported, as the alcohol-IPV association remains significant even after controlling for a long list of moderating factors (for a review, see Eckhardt et al., 2022). In contrast, ample evidence supports the proximal effects model for many forms of aggressive behavior (Abbey, 2023; Giancola et al., 2010; Shorey et al., 2014). This approach aligns well with research questions focused on the causal mechanisms of alcohol-facilitated IPV perpetration (i.e., “the how question”): How does acute alcohol use affect cognitive (e.g., attention, attitudes), affective (e.g., negative affect, anger), and physiological processes (e.g., arousal) in ways that increase the proclivity to act aggressively towards an intimate partner? Such research questions are at the forefront of our collaborative work (e.g., Bresin et al., 2020; Eckhardt et al., 2021) and other scholars in this field (Abbey et al., 2000; Shorey et al., 2017; Watkins et al., 2015).

The proximal effects approach is typically informed from the etiologic standpoint of Alcohol Myopia Theory (AMT; Steele & Josephs, 1990), which suggests that alcohol’s pharmacological properties specifically disrupt attentional processes. The consequences of alcohol consumption, particularly at higher levels, involve narrowed attention, restriction in the cues the drinker can perceive, and a reduced ability to meaningfully interpret the information the drinker does perceive. During couple conflict, provocative cues (e.g., an insulting comment from one’s partner) are typically more salient and temporally recent compared to inhibitory cues (e.g., legal or relationship consequences of conflict escalation). As a result, proximal alcohol use that closely precedes conflict situations increases the likelihood of IPV by narrowing the drinker’s attention onto the most recent and salient provocative cues to a greater extent than distal, inhibitory cues (Fleming et al., 2013; Gable et al., 2016). All told, AMT has garnered extensive empirical support as an explanation for a range of alcohol-related behaviors, including IPV (Giancola et al., 2010; Massa et al., 2019).

Thus, the alcohol-IPV association has been confirmed across multiple methodologies, and is best understood through a proximal effects, AMT-informed, framework (Eckhardt et al., 2015). The AMT framework facilitates researchers’ examination of the immediate processes that may account for how acute alcohol intoxication facilitates specific acts of IPV in specific contexts. Identification of these in-the-moment processes has enormous significance for (1) understanding the practical, proximal dynamics of abusive interactions among relationship partners, and (2) developing etiological models of partner abuse that can inform efforts to design intervention programs that effectively reduce the occurrence of IPV perpetration.

Alcohol-Facilitated IPV Perpetration: A Review of Methodologies

How, then, should researchers endeavor to uncover these in-the-moment mechanisms? What are the specific research methods available to researchers who seek to examine the putative mechanism(s) that underlie alcohol’s immediate effects on IPV? Below, we review the methodologies that, in our opinion, represent the most rigorous options for testing the proximal effects of alcohol on IPV perpetration.

Alcohol Administration Methods

Our laboratory-based work is built upon the innovative and foundational research of many noteworthy scholars in this area who pioneered the use of standardized alcohol administration methods to advance addiction research generally (George et al., 2012; Martin & Sayette, 1993) and the alcohol-aggression literature specifically (Duke et al., 2011; Giancola & Zeichner, 1997; Leonard, 1984). Importantly, these methods are grounded in a multitude of ethical and regulatory protections to maximize participant safety and health (see recent guidelines from the National Institute of Alcohol Abuse and Alcoholism [NIAAA, 2023]).

As noted in an excellent overview of laboratory-based alcohol administration methods (Abbey, 2023), if the goal is to maximize the broader internal validity of research that investigates whether acute alcohol use impacts a specific behavioral outcome, then a standardized alcohol administration method is without comparison. This approach involves random assignment of participants to beverage conditions, application of standardized algorithms to determine equivalent alcohol dosing amounts across participants, and eligibility criteria to effectively isolate alcohol consumption as the proximal cause of changes in an outcome measure. Thus, the use of laboratory-based alcohol administration methods allows for (1) a closer estimation of the causal effect of a specific alcohol dose on aggression-related outcomes, and (2) the assessment of concomitant biological and psychological processes hypothesized to comprise the mechanistic causal link(s) between alcohol and aggression (for a quantitative review of lab studies on alcohol and violence, see Crane et al., 2016). However, despite the successful integration of laboratory-based alcohol administration and aggression paradigms to study different forms of aggression perpetration (e.g., “general” physical aggression: Giancola, 2004; physical IPV: Eckhardt et al., 2021; sexual violence: Abbey, 2023) and sexual victimization (e.g., Parks et al., 2016), these methods are sometimes criticized for being an artificial context for alcohol consumption, as they often lack the key situational cues present in other consumption venues (e.g., bar settings; parties). We fundamentally disagree with these criticisms and point the reader to what we view as timeless rebuttals to such criticisms of “artificiality” in laboratory- based research (Berkowitz & Donnerstein, 1982; Mook, 1983). Setting that criticism aside, we offer several considerations for incorporating alcohol administration methods into IPV research.

In any alcohol administration study, decisions regarding the inclusion of a control group(s) are critical. For studies on interpersonal aggression, we recommend the use of a no- alcohol control group (told no alcohol, receive no alcohol) over a placebo control group (told alcohol, receive no alcohol) for several reasons (e.g., Eckhardt et al., 2021; Parrott & Zeichner, 2002; Testa et al., 2014). First, a meta-analysis (Bushman & Cooper, 1990) concluded that the ideal manipulation to estimate alcohol’s effect on laboratory-based aggression involves the alcohol versus no-alcohol control comparison, because (1) placebo and no-alcohol control groups do not significantly differ in aggression, and (2) placebo, relative to no-alcohol control, groups are more likely to produce compensatory responses that reduce aggression (thereby artificially increasing the observed effect of alcohol on aggression). Second, placebo participants tend to behave in a manner more consistent with participants in a no-alcohol control condition than an alcohol condition (Testa et al., 2006). Third, the purpose of the placebo-control group, which is to control for the effects of participants’ expectations about how they should behave when intoxicated, is not well-supported (e.g., Watkins et al., 2023). While we are not dismissing the utility of placebo control groups altogether, we suggest that the placebo control group is simply not the most rigorous option if one is studying aggression as the outcome.

In our experience, there are many key procedures to a well-run alcohol administration study that oftentimes don’t appear in journal articles. Here, we highlight three¹. First, in our research with couples, it is standard practice to take and store participants’ phones during the experiment to prevent distraction and partner communication during the experimental protocol. Many participants struggle with not being able to check or interact with their phone. Thus, providing research staff with scripted discussion points and options for ensuring that phones can be used briefly for critical reasons related to family or work is beneficial. Second, in just about every alcohol administration study we have conducted, a small number of participants experience adverse, but mild, physical reactions to consuming alcohol (e.g., feeling nauseous or actually vomiting). Just as ethical guidelines require us to inform participants of these potential risks (which we do in the written consent form and verbal description of the study early in the consent process), we also ensure that research staff expect and are prepared for these risks. Such preparations must include written procedures that specify exactly what staff should do in these situations (e.g., having gender-aligned staff that can accompany the participant into a restroom; a clear line of communication procedures). Preparations also include having the resources to make the participant comfortable (e.g., non-alcoholic beverages, mouthwash, cleaning materials).

Third, there is…The Waiting. Participants who consume alcohol must typically wait 3–5 hours before they can depart the lab due to the immutability of alcohol metabolism processes. During this time, participants’ breath alcohol concentration (BrAC) is regularly assessed to ensure that they are discharged at an appropriate BrAC, which can vary depending on a variety of considerations (NIAAA, 2023). Thus, participants must be informed during the consent process of the potential for long wait times (we also have participants sign a separate statement that they will not leave prior to predetermined BrAC); however, it is of equal importance for research staff to be trained on protocols that (1) promote participant comfort and safety during the wait period, and (2) detail procedures for participants who want to leave before they reach a predetermined BrAC. Regarding the former, the overarching consideration is to ensure that participants do not incur any risks that directly result from their intoxication level and/or detoxification period. For example, we prefer to have partners detoxify in separate rooms (e.g., to reduce the likelihood of couple conflict in the lab); however, for some couples, detoxing together is a positive experience and thus appropriate. The waiting space should be comfortable and quiet in case participants nap, and there should be sources of entertainment to help pass the time (e.g., a television that provides streaming content). Food should be provided at no cost to the participant, especially since they likely were asked to fast prior to the study, and research personnel should have a readily available list of restaurants that can deliver food quickly.

Safety considerations necessitate that at least two research personnel be present for any session and that they can see the participant at all times (e.g., via closed circuit video camera) in case there is an adverse reaction or situationally inappropriate behavior. In our studies, while we have experienced very few problems with participants, a small number of individuals have engaged in inappropriate behaviors (e.g., masturbation; property destruction) or politely (or angrily!) demanded to leave the laboratory prior to the discharge BrAC. Research personnel must therefore be fully trained on IRB-approved protocols for executing a discharge prior to a predetermined BrAC. While situations such as these are uncommon, it is nevertheless important for researchers to have thought through these and other scenarios that might be expected among individuals with a history of heavy drinking and aggression who are in unstable and conflictual relationships. The myriad approaches and considerations cannot all be detailed here; however, interested researchers can determine best practices for their research and setting via direct conversations with colleagues who have alcohol administration experience.²

So, given all these methodological challenges and ethical considerations, why do we do alcohol administration research? And why in the world would we choose to integrate such methods with comparably structured and predictable laboratory-based aggression assessment paradigms? For us, the answer is straightforward: These methods are simply the best way to model the causal mechanisms that underlie how acute alcohol use affects aggression, i.e., the how question. Using structured alcohol administration methods is, in our view, the only approach that exactly replicates the pharmacologic processes that operate when humans consume alcohol in naturalistic settings; every alternative method is one step removed from being able to speak directly to alcohol’s acute and temporal effects. “Gold Standard” Methodologies to Assess the Proximal Effect of Alcohol on IPV Perpetration

In March 2023, the National Institutes of Health issued a Notice of Special Interest (NOT-AA-23–003) titled “Advancing mHealth Interventions for Understanding and Preventing Alcohol-Related Domestic Violence.” A core purpose of this call is to stimulate the use of research designs that best allow for a modeling of the proximal and temporal associations between alcohol and IPV. In the sections below, we provide examples of these “gold standard” designs in the study of alcohol-facilitated IPV, including three experimental methods (i.e., couple observation methods, vignette/scenario-based methods, and laboratory-based aggression assessment methods) and intensive longitudinal designs (e.g., daily diary, ecological momentary assessment). Rather than providing a comprehensive review of all studies in this area, our aim is to instead provide a brief review of each design and to identify the core elements of each approach that allow for the proximal assessment of the mechanisms that may account for the alcohol-IPV association.

Couple observation methods.

Building on early laboratory work (Gottman, 1979; Margolin et al., 1988), the couple observation approach involves having couples discuss common topics of disagreement in a laboratory setting. In the context of understanding how acute alcohol intoxication may play a role in these conflicts (e.g., Leonard & Roberts, 1998), researchers typically have couples with varying IPV histories engage in two 10–15 minute recorded discussions (one at baseline; another after one or both partners are randomly assigned to a beverage condition) of the most major unresolved issue for each couple. Couples are instructed to discuss how each person is responsible for the issue, the challenges that the issue presents for the relationship, and how it might be resolved. The resulting video interactions are coded, using various structured measurement systems, for actor-partner interactions. These interactions might include actor responses to partner provocation (composites of hostility, criticism, contempt, and ridicule), actor responses to partner rejection, and behavioral indicators of actor and partner physical and psychological IPV (e.g., verbally aggressive or belligerent statements).

A relatively small number of studies have integrated alcohol administration methods with couple-observation approaches to assess alcohol-facilitated IPV (Haber & Jacob, 1997; Jacob & Leonard, 1988; Leonard & Roberts, 1998). Given the complexities involved in each individual method, this may be unsurprising, especially given the dynamics at work when the two methods are combined with intoxicated individuals (or couples: see Testa et al., 2014) engaged in difficult and often angry discussions with an intimate partner. In addition, it is often unclear whether this laboratory-based and time-limited approach accurately samples the types of partner-directed maltreatment that are present in the day-to-day lives of people in distressed relationships, even when alcohol intoxication is included. Despite these challenges, the couple observation approach allows researchers to extend their understanding of the alcohol-IPV association far beyond what would be provided by retrospective self-report methods, especially regarding the dyadic and interactive mechanisms that characterize alcohol-involved couple conflict.

Vignette/Scenario designs.

In this approach, after consuming an alcohol or control beverage, participants imagine a vignette-based provocative situation that is presented as a written narrative or as a video- or audio-recording; there are variations in whether the participant, or other fictional individuals, are the main character in the scenario. The participant provides a response to the vignette, which is recorded as a proxy for IPV perpetration; the response can be a spontaneous verbalization articulated while imagining the scenario, or it may involve the participant indicating their response to a questionnaire item. Given the focus on an individual participant, this approach offers a high degree of flexibility while also allowing for a more structured presentation of the provocation manipulation. The vignette/scenario approach is commonly used in research examining the role of alcohol in sexual violence perpetration; see Abbey’s (2023) comprehensive review of this body of work.

In the context of assessing the role of alcohol on non-sexual forms of IPV, the vignette/scenario approach has many benefits for addressing “the how question.” Specifically, researchers can carefully specify the instigating scenario that may be related to alcohol-related IPV perpetration and essentially guarantee that every participant is exposed to the same situation. Approaches such as the Articulated Thoughts in Simulated Situations method (ATSS; Davison et al., 1983) involve presenting structured provocative scenarios to individual participants and recording their verbalizations while imagining these scenarios (Eckhardt, 2007). Such a high level of experimental control is not possible in couple observation methods given the variability across couples regarding the content, nature, and dynamics of the provoking context. Researchers can also vary the thematic content of the vignettes/scenarios presented to participants to specify further the conditions that may yield more aggressive responding. The responses measured during scenario exposure can provide proxies for IPV-related behaviors, such as indicating how likely they would be to engage in different types of IPV if they were in the situation, as well as the ongoing, in-the-moment assessment of cognitive and affective processes that accompany IPV-related responding during varying conditions of alcohol consumption. Several studies have applied the vignette/scenario approach, particularly the ATSS method, in the context of alcohol administration to assess alcohol-related aggressive verbalizations that occur during in-the- moment anger arousal (Eckhardt, 2007; Stappenbeck & Fromme, 2014).

Laboratory-based aggression assessment methods.

Within the general interpersonal aggression literature, the most widely used laboratory-based measure of aggression is the Taylor Aggression Paradigm (TAP; Taylor, 1967), in which participants are told that they will compete against an opponent (seated in an adjacent room) on a reaction-time task. Depending on whether participants win or lose a reaction-time trial, they will either administer or receive an electric shock to/from their opponent. Of course, no opponent actually exists and the sequence of wins and losses as well as the intensity of the shocks set by the “opponent” is predetermined by the experimenter. The TAP is built upon on an over five decades-long foundation of research findings and all manner of methodological considerations, data analytic suggestions, and recommendations for interpretation (Giancola & Parrott, 2008; Hyatt et al., 2019).

Over the past several years, researchers have extended the TAP framework to examine aggression-related outcomes in the context of alcohol-involved conflicts among relationship partners that allows participants to act aggressively toward their (apparent) romantic partner in real time (Eckhardt et al., 2021; Watkins et al., 2015). The increased methodological complexity created by the presence of two partners in the same research laboratory creates new ethical considerations, particularly around safety and privacy. Regarding safety, we exclude couples who pose the greatest risk of violence in the lab or as a result of the lab experience (e.g., those who report fearing for their life or seeking shelter/services for domestic violence). Regarding privacy, at every point in our research protocol (i.e., from first point of contact through discharge), we ensure that each partner can communicate with the research team without their partner present. Partners receive individual communications (i.e., without cc-ing or leaving messages with their partner), we meet with each partner individually in the laboratory (particularly during consent procedures), and we protect a partner’s privacy when they do not want to continue the study. In conclusion, laboratory-based experimental studies are methodologically and ethically complex; however, we consider it a “gold standard” approach because it provides a standardized instigatory context for IPV-related behaviors to be assessed (i.e., delivering painful electric shocks to an intimate partner) and a clear answer to “the how question.”

Intensive Longitudinal Designs.

Intensive longitudinal designs are characterized by obtaining many repeated assessments of a behavior over some time period (Bolger & Laurenceau, 2013). The most common application of this design in the IPV literature (Shorey et al., 2017; Testa & Derrick, 2014) is the use of daily diary or ecological momentary assessments (EMA). In a daily diary study, participants complete a single, brief survey assessment (usually less than 5 minutes) each day, typically in the morning, which captures behaviors that occurred the prior day. In an EMA study, participants complete multiple survey assessments each day (usually 1–2 minutes per assessment), which capture behaviors or experiences that are occurring in the present moment (i.e., at the time they are completing the survey). With both methods, the time of day in which a given behavior was enacted is recorded, which allows for subsequent determinations of temporal sequencing. Thus, when applied to the alcohol-IPV association, intensive longitudinal methods allow researchers to identify the specific time of day in which participants consumed alcohol (and how much they drank) as well as whether IPV was perpetrated after the initiation of drinking. In addition, these methods also allow for the assessment of putative mediators (e.g., anger, hostile cognitions) during periods of conflict. As such, daily diary and EMA methods assess alcohol use and IPV perpetration in ecologically valid contexts while affording near real-time assessment of behavior, which is a strong method for determining temporal relations.

Key Barriers to Answering “The How Question”

We have outlined what we view as a key inflection point in the alcohol-IPV field, in which scholarly focus must shift from a question of “Does alcohol cause IPV perpetration?” to “How does alcohol cause IPV perpetration?” In addition, we have highlighted the most rigorous methods to be used in future research investigating “the how question.” However, key barriers remain that, if left unaddressed, will greatly limit our ability to understand fully the putative mechanisms of alcohol-facilitated IPV perpetration. In the sections that follow, we conclude our Perspectives paper by detailing some of these key barriers.

Methods to Examine Putative Mechanisms of Action Are Complex and Costly

We recognize that we are advocating for complex designs to assess the mechanisms of alcohol-facilitated IPV with the benefit of hindsight -- our research teams were able to use these approaches successfully, and the resulting findings have been informative. But we should acknowledge the potential barriers to more widespread use of these designs. First, researchers may be concerned about legal/ethical issues related to participants’ involvement in research on alcohol and violence, such as inappropriate behavior (see our earlier discussion on ‘the waiting’), partner safety (see our previous section on the TAP for how we address this), or other issues often raised by institutional review boards (Bender, 2017). For example, on occasion our IRBs will inquire as to whether researchers must report participants’ admissions of prior partner- directed aggression to the authorities. Consistent with the practice of most studies on alcohol- facilitated IPV of which we are aware (as well as Certificates of Confidentiality and our ethical responsibilities as researchers), we explicitly state in IRB protocols that we will not reveal participants’ self-reports of past IPV to anyone outside of the research project. In situations where participants directly threaten violence towards a specific person, of course we must follow all legal guidelines about mandated reporting. But our perspective is that more effective intervention and prevention methods will emerge if we more fully understand risk factors for IPV perpetration, which rely on creating experimental contexts wherein individuals with IPV histories can “reveal themselves” with all expected confidentiality protections in place.

Second, gold standard-type research designs are often quite expensive to conduct, require sizeable grants to support them, and can place severe demands on researchers’ time. In our experience, it may take 1–2 years before a well-conceptualized gold-standard research plan is funded, followed by however many months or years it takes to complete such complex studies. These time demands may disproportionately impact early career and/or pre-tenure researchers, who simply may not have years to spend on any given research project, no matter how gold standard it may be! That said, many scholars have navigated these tricky waters, usually by complementing a time-intensive, years-long alcohol-aggression project with an approach that prioritizes productive collaborations (i.e., regularly generates strong publications). For such an approach to be successful, scholars must (1) recognize the professional activities which are most valued by their institution, and (2) strike the right balance of being a ‘fox’ (i.e., focusing on many topics) and a ‘hedgehog’ (i.e., focusing on a single topic) (Taylor et al., 2006).

Sampling Methods

Our respective research programs have long-recognized the importance of who we include in studies of alcohol-facilitated IPV. In our past work (e.g., see Eckhardt et al., 2015), we argue that research on alcohol-facilitated IPV must use sampling methods that ensure representation of high-risk populations (e.g., defined by common measures of problem drinking, presence of an alcohol-use disorder, recent history of IPV perpetration). For instance, there is substantial literature regarding the cross-sectional and distal-longitudinal associations between alcohol and IPV perpetration in heavy episodic drinking samples (Smith et al., 2012). However, we know relatively little about the proximal and temporal nature of this association among this population, in part because the initial versions of federal guidelines regarding the use of alcohol administration methods with human participants strongly discouraged the use of such methods with heavy drinking individuals. While more recent guidelines include some provisions that would allow heavy drinking individuals to participate in an alcohol administration study (NIAAA, 2023), most studies in this area still recruit “social” or non- problematic drinking samples or do not oversample heavy episodic drinkers to permit meaningful analysis. Research on etiological mechanisms will have the greatest impact on intervention development to the extent that at-risk drinkers are oversampled, as they are the individuals for whom intervention development will typically have the greatest impact. Thus, sufficient numbers of at-risk drinkers must be purposefully sampled to address this problem.

As another major example of the importance of who we include in our studies, research on alcohol-facilitated IPV exhibits a relative lack of diversity in study samples (e.g., racial, ethnic, cultural) and rarely addresses research questions that are directed to the unique concerns of stigmatized populations. To illustrate, consider the significant cisheteronormative sampling bias in IPV research. Historically, sampling methods were designed with the assumption that participants identify as cisgender and heterosexual or used methods that necessarily exclude sexual and gender diverse people. These sampling biases are directly responsible for the marginalization of sexual and gender diverse people in this research literature. For example, while multiple reviews of research from the past 25 years document the prevalence and etiology of IPV in sexual and gender minoritized populations (e.g., Edwards et al., 2015: n = 96 studies; Trombetta & Rolle, 2023: n = 78 studies), only eight studies were identified by these reviews which examined the distal and/or proximal effects of alcohol use.

Related to this problem, analysis of research included in narrative and meta-analytic reviews of the alcohol-IPV link (Foran & O’Leary, 2008; Leonard, 2005) reveals that studies rarely assess perpetrators’ sex assigned at birth, gender identity, gender expression, and sexual orientation. This weakness, which is also robust across studies of sexual IPV perpetration toward sexual and gender minoritized people (Rothman et al., 2011), prevents researchers from ascertaining the diversity of identities of IPV perpetrators and is particularly problematic for understanding the effects of alcohol on IPV in sexual and gender diverse people. At the core of this problem is that, relative to non-stigmatized heterosexual and cisgender identities, stigmatized sexual and/or gender identities are associated with minority stressors that negatively impact health (Brooks, 1981; Meyer, 2003). When these identities are not assessed or considered, the effect of minority stressors on the alcohol-IPV relation are also not considered.

Collectively, these weaknesses in sampling methods weaken the diversity of study samples and make it difficult, if not impossible, for studies to focus on the unique concerns of stigmatized populations which inform etiological mechanisms that underlie alcohol-facilitated IPV across the continuum of individual risk and diversity of intimate relationships. Until this problem is addressed, efforts to develop effective, culturally-informed IPV treatment and prevention efforts will remain ill-informed and potentially misdirected.

Understudied Forms of IPV Perpetration

IPV is a multifaceted construct that can be expressed behaviorally in myriad ways, including physical, psychological, sexual, and stalking behaviors committed by an intimate partner with an intent to harm. However, the IPV literature has been criticized by scholars for overemphasizing physical assault (Hamby, 2014; Hamby & Turner, 2013) and often failing to consider other forms of IPV. In our view, this criticism also applies to the study of the alcohol- IPV link, which is largely “siloed” in terms of the IPV outcomes assessed. Most notably, studies that examine the effect of alcohol on physical or psychological IPV rarely include measures of sexual IPV or stalking. Likewise, studies that examine the effect of alcohol on sexual violence perpetration rarely consider the relationship context between the perpetrator and victim. Overall, these weaknesses limit the ability to consider simultaneously how alcohol use is proximally and temporally related to multiple forms of violence that likely co-occur.

To address these weaknesses, intensive longitudinal studies should strive to include assessments of multiple forms of IPV perpetration. However, assessing more forms of IPV can be easier said than done, as the additional items may conflict with the need for daily surveys to be brief and of minimal burden to participants. The solution is less straightforward for laboratory- based studies involving physical IPV assessment (e.g., the TAP). That stated, validated laboratory-based measures of sexual violence exist (Abbey, 2023; Abbey & Helmers, 2020; Franz et al., 2018) that can be adapted for the couples-context to assess alcohol-facilitated sexual IPV perpetration.

Examining Individuals in Isolation: A Call for Dyadic Approaches

IPV is a dyadic behavior that requires both partners. However, most research on alcohol- facilitated IPV perpetration – and most IPV research overall -- focuses on actor effects (i.e., the effect of Partner A’s alcohol use on Partner A’s behavior) without consideration of their partner’s self-reported characteristics. The reliance on a single respondent to report their own and/or their partner’s perpetration limits the validity of measurement and, consequently, generalizability. While a single-partner measurement approach has historically been a key weakness in the rigor of IPV research, we view the publication of Dyadic Data Analysis (Kenny et al., 2006) as a key inflection point in IPV research. In what we view as a “modern classic” text that is required reading for IPV scholars, Kenny and colleagues’ (2006) analysis of the scientific study of interpersonal relationships presents a compelling argument:“…most interpersonal concepts have been studied by examining individuals in isolation. Before we can have a genuinely interpersonal social science, our theories, research methods, and data analyses must take into account the truly interpersonal nature of the phenomena under study” (p. 3). In essence, they highlight the extreme disconnect between what researchers are (or had been) doing relative to what they should be doing.

The past 15 years has witnessed a substantial increase in the number of published studies in the alcohol-IPV field that address IPV via a dyadic approach involving analyses of both partners’ risk factors for IPV (Eckhardt et al., 2019). Relative to an analysis of only one partner’s characteristics, this approach more accurately models IPV perpetration risk by providing the unique ability to examine the effects of both partners’ characteristics on IPV perpetration while accounting for the other partner’s characteristics and perpetration. For example, studies indicate that Partner B’s problematic drinking is positively associated with Partner A’s IPV perpetration in both cisgender, heterosexual couples (Leone et al., 2016; Testa et al., 2012) and SGM couples (Do et al., 2021; Leone et al., 2022; Parrott et al., 2023). In addition, a dyadic approach affords the opportunity to examine the interaction between Actor and Partner risk factors on Actor IPV perpetration. For instance, studies indicate that Actor impulsivity (Leone et al., 2016) and Actor emotion dysregulation (Parrott et al., 2017) are positively associated with Actor IPV perpetration toward partners who were problematic, relative to nonproblematic, drinkers.

In our view, the most methodologically rigorous studies within IPV research more broadly (and alcohol-specific research specifically) adopt a dyadic approach afforded by the recruitment and assessment of both partners (Eckhardt et al., 2019). Yet, we recognize the challenges inherent to meeting this standard. In our collaborative work, we have first-hand experiences of the logistical difficulties in recruiting, screening, and enrolling both members of couple, particularly when in-person and/or time-intensive laboratory visits are required. Moreover, as noted previously, recruitment of both partners heightens risks of participant confidentiality, privacy, and safety that must be carefully and thoughtfully addressed in consultation with local IRBs. These potential risks may be heightened further when there is a history of IPV within the couple and/or one or both partners reports significant risk factors (e.g., heavy alcohol use).

Limited Focus on Neural Factors

Extant literature on alcohol-facilitated IPV focuses heavily on psychosocial etiological mechanisms (e.g., attention allocation, emotional processing). As such, there is a relative paucity of research that integrates a neuroscientific focus into this etiological work, despite evidence that the neural substrates of these mechanisms are overlapping (Denson et al., 2009, 2018; Gilam et al., 2018; Nikolaou et al., 2013). We see this as a critical weakness in the current rigor of this research. Research on the alcohol-IPV association which integrates neurocognitive theory and measurement tools will afford a better understanding of putative mechanisms and, in turn, better inform development of interventions that more directly target psychosocial and neurocognitive mechanisms. Recommendations such as these are consistent with past calls for research to incorporate neuroclinical measures into the addictions field (Kwako et al., 2016). Thus, to advance the field in this area, research is needed which integrates well-validated neurocognitive measures (e.g., neuroimaging, electroencephalography [EEG]) with the aforementioned recommendations for advancing our understanding of the proximal effects of alcohol on IPV perpetration. Recent examples that begin to integrate these ideas exist, including the use of fMRI to assess non-alcohol related IPV perpetration dyadically (Chester et al., 2021) and to assess alcohol-facilitated aggression, albeit not within romantic couples (Denson et al., 2018).