Abstract
Despite the fact that intimate partner aggression (IPA) is a widespread public health problem, empirically supported interventions for IPA are limited. Furthermore, existing interventions tend to be resource-intensive and may not adequately address the risk factors that serve to maintain IPA, resulting in challenges to intervention dissemination and implementation. Based on theoretical and empirical findings linking emotional clarity and IPA, this study represents a secondary data analysis to evaluate the preliminary efficacy of a brief web-based cognitive restructuring (CR) intervention to reduce psychological and physical IPA perpetration intentions by increasing emotional clarity skills. In all, 137 men were randomized into a CR intervention versus control condition and subsequently completed an aggression analog scenario. Results of structural equation modeling analyses showed a significant indirect effect of the intervention on psychological IPA intentions via increased emotional clarity. Specifically, men in the CR intervention condition evidenced greater emotional clarity as compared to men in the control condition, which, in turn, was related to lower intentions to perpetrate psychological IPA. The indirect effect from the CR intervention to physical IPA intentions via emotional clarity was not statistically significant. These findings provide promising initial support for the usefulness of the current CR intervention in reducing psychological IPA. Next steps in this line of research include expansion to a randomized controlled trial that tests intervention effects on real-world IPA perpetration across more diverse samples. Importantly, our findings highlight that CR skills can be delivered via a brief web-based intervention, which decreases potential barriers to dissemination and implementation.
Intimate partner aggression (IPA) is a prevalent public health concern. Young adults experience IPA at especially high rates, with approximately 23–38% of dating relationships involving physical IPA and 81–86% involving psychological IPA (Bell & Naugle, 2007). Despite the fact that IPA is a widespread problem with significant health impacts (Centers for Disease Control and Prevention, 2021), empirically supported interventions for IPA are limited (Stover et al., 2009). Furthermore, those interventions that do exist tend to be resource-intensive and may not adequately address the risk factors that serve to maintain IPA (Stover et al., 2009), resulting in challenges to the dissemination and implementation of IPA interventions. Because the consequences of male-perpetrated IPA tend to be particularly detrimental (Archer, 2000), interventions to reduce IPA perpetration among young men may be especially useful.
The I3 Model (Finkel, 2011, 2014; Finkel & Eckhardt, 2013) provides a theoretical framework to help explain risk factors of IPA. According to this model, risk for IPA is determined by three factors: instigation (i.e., experiences that increase one’s urges to aggress), impellance (i.e., dispositional factors that strengthen urges to aggress), and inhibition (i.e., one’s ability to resist or override urges to aggress). Emotion regulation is an empirically supported inhibitory factor that has been consistently and robustly associated with IPA (Neilson et al., 2023). Among men, emotional clarity, which reflects the extent to which individuals know the emotions they are experiencing (Gratz & Roemer, 2004), appears to be particularly predictive of IPA perpetration, possibly due to masculine socialization processes (Berke et al., 2019).
Although some individuals may inherently have stronger emotional clarity than others, research suggests that certain skills can be taught. Cognitive restructuring (CR) is a technique included in many cognitive-behavioral therapies that involves actively challenging and modifying one’s thoughts as a way to change how one feels about a given situation that influences subsequent behavior (Berking et al., 2008). In line with the I3 Model, CR can bolster inhibition through increasing emotional clarity, and may thus be a useful tool for reducing IPA.
The Present Study
This study represents a secondary data analysis to evaluate the preliminary efficacy of a brief CR intervention in reducing psychological and physical IPA perpetration. This intervention was developed and evaluated in a sample of young men with sexual assault perpetration histories. Because approximately half of the sexually aggressive incidents tend to involve alcohol, by the perpetrator, the victim, or both (Abbey, 2002), half of the participants in the original data collection underwent an alcohol administration procedure. Although participants were not selected based on a history of IPA, there are findings showing that individuals who perpetrate one form of aggression are likely to perpetrate multiple forms (Wilkins et al., 2014). Given known associations between CR, emotional clarity, and IPA (Neilson et al., 2023), the goal of this study was to preliminarily examine whether the intervention would also be helpful in reducing the likelihood of IPA through bolstered inhibition. Specifically, as suggested by the I3 Model, we hypothesized that improved CR abilities would increase men’s emotional clarity, thereby decreasing their likelihood of perpetrating IPA due to better control of aggressive urges. Importantly, the CR intervention under investigation is brief and web-based, and thus requires minimal resources. If proven successful, it has the potential to be widely disseminated and help overcome several of the barriers of existing interventions.
Method
Participants and Procedure
Participants were recruited through online and print advertisements, as well as the university registrar’s student list. The advertisements solicited single, male drinkers who were interested in dating women. Interested men completed an online screening survey to determine eligibility, then were called to confirm eligibility and schedule an appointment. Based on the goals of the larger data collection, which included evaluating mechanisms of change for CR interventions on sexual aggression likelihood, men (N = 209; 69.9% full-time or part-time students) were eligible if they were (1) between 21 and 30 years old, (2) had sexual intercourse with a woman at least once in the past year, (3) reported at least one incidence of sexual aggression perpetration in their lifetime, (4) had no history of alcohol problems, no medical conditions or medications that contraindicated alcohol consumption, (5) consumed at least three alcoholic drinks per week, and (6) had at least one instance of heavy episodic drinking in the past year. All recruitment and study procedures were approved by the University’s Institutional Review Board.
The larger study from which the current data were drawn involved a 2 (beverage administration) × 3 (intervention condition) design. Participants came to the laboratory and, after providing informed consent, completed a battery of background measures on a computer. Participants were then block randomized to receive one of two interventions (CR or mindfulness) or a control condition. Half of the participants in each condition completed a standard beverage administration procedure with a target breath alcohol concentration of 0.08%. Finally, all participants read a sexual aggression scenario and completed dependent measures, then were detoxed, paid, and released. Participants were compensated $15 per hour they spent in the laboratory, and the participants who consumed alcohol stayed in the laboratory until their Breath Alcohol Concentration (BrAC) was below 0.03%. Total session length ranged from 2 to 8 hours depending on whether or not detoxification was needed. For more detailed information on these study procedures, please see Davis et al. (2021).
Given our previous research (Davis et al., 2021) showing that mindfulness did not influence emotional clarity, this focused study included the 137 men who participated in either the CR or control condition and had complete data on the outcome variables. All participants were 21–30 years old (M = 24.96, standard deviation [SD] = 2.76). Participants were White 59.9% (n = 82), Asian 18.2% (n = 25), Black 5.8% (n = 8), and Multi-racial 11.0% (n = 15); 1.5% (n = 2) identified as some other race, and 3.6% (n = 5) declined to identify their race. In addition, 8.0% (n = 11) identified as Hispanic or Latino.
Materials and Measures
Intervention
The intervention consisted of a brief (5–10 minutes), web-based skills training completed before the beverage administration. The training for individuals in the CR intervention was based on the Antecedent, Belief, Consequences Model of Cognitive Behavioral Therapy (Ellis, 2008), in which participants were taught to identify their initial thoughts about an interpersonal interaction, labeled their subsequent emotions, developed more balanced alternative thoughts, and identified any new emotional responses. Participants in the control condition received information regarding basic nutrition that was similar in length and engagement as the CR intervention. Next, all participants were presented with a practice situation to practice applying the information they had just learned. Finally, all participants were told they would be reading a story in the next phase of the study. Those in the CR intervention condition were provided further instructions to practice applying the skills they had just learned while reading the scenario.
Aggression analog scenario
After completing background measures, the intervention or control condition, and beverage administration, participants were instructed to project themselves into a sexually explicit scenario. This scenario was written in the second person, with the participant as the protagonist, and described the protagonist’s interaction with a hypothetical woman (“Kim”) with whom he had previously had sexual intercourse. In the story, the protagonist has a conversation with Kim at a party, after which they go to Kim’s apartment. They engage in consensual kissing, during which Kim states that she does not want to have sex that night. Kissing and sexual activity continue, with Kim again clearly stating that she does not want to have sexual intercourse while pushing the protagonist away, after which the story ends. Dependent measures were assessed after conclusion of the story.
Emotional clarity
Emotional clarity during the scenario was assessed via two items (“I am confused about how I feel,” “I have no idea how I feel,” reverse-scored; α = .67) from a modified version of the State-Difficulties in Emotion Regulation Scale (S-DERS; Lavender et al., 2017). Participants rated the extent to which the statements on the S-DERS applied to their emotions in the present moment (1 = Not at all to 5 = Completely). Mean scores were created, with higher scores indicating greater emotional clarity.
IPA likelihood
IPA likelihood was assessed at the end of the story via three items tapping into psychological aggression (e.g., “How likely are you to insult or swear at Kim?”; α = .83) and four items tapping into physical aggression (e.g., “How likely are you to hit or slap Kim?”; α = .84) from a modified version of the revised Conflict Tactics Scales (Straus et al., 1996). Participants rated the likelihood of engaging in each aggressive act at that point in the story on 7-point scales (1 = Very Unlikely to 7 = Very Likely). Mean scores were created, with higher scores indicating greater psychological and physical IPA.
Data Analytic Approach
Structural Equation Modeling in Mplus Version 8 with Maximum Likelihood Robust estimation (to account for non-normality) was used. Our model included a dichotomous (0 = control vs. 1 = CR) intervention variable as the predictor, emotional clarity as the mediator, and psychological and physical IPA as outcomes (see Figure 1). Because there were no main or interactive effects of alcohol use, beverage administration (0 = no alcohol vs. 1 = alcohol) was included as a covariate in all analyses. Research materials related to this study are available from the corresponding author on reasonable request.
Mediation effects from intervention (CR vs, control) to psychological and physical IPA perpetration via emotional clarity.
Results
Descriptive Statistics
Men in the CR condition (n = 69) and men in the control condition (n = 68) did not differ significantly on psychological IPA perpetration likelihood (M = 1.18, SD = 0.61, range = 1–4.33 vs. M = 1.18, SD = 0.40, range = 1–2.67; t[135] = 0.25, p = .98) or physical IPA perpetration likelihood (M = 1.07, SD = 0.36, range = 1–3.50 vs. M = 1.12, SD = 0.44, range = 1–3.50; t[135] = −0.66, p = .51). However, men in the CR condition had significantly higher levels of emotional clarity as compared to men in the control condition (M = 4.62, SD = 0.58, range = 3.00–5.00 vs. M = 4.25, SD = 0.85, range = 2.00–5.00; t[117.60] = 2.99, p = .003). Correlations between psychological IPA likelihood and physical IPA likelihood (r = .71, p < .001) and between psychological IPA likelihood and emotional clarity (r = −.15, p = .04) were statistically significant, whereas there was no significant correlation between physical IPA likelihood and emotional clarity (r = −.07, p = .35).
Mediation Model
As shown in Figure 1, after controlling for beverage condition, there was a significant indirect effect from the intervention to psychological IPA likelihood via emotional clarity (β = −.06, p = .04). Specifically, men in the CR intervention condition evidenced better emotion regulation abilities than men in the control condition (β = .25, p < .001), which, in turn, was related to lower psychological IPA likelihood (β = −.24, p < .001). The indirect effect from the intervention to physical IPA likelihood via emotional clarity was not statistically significant (β < −.02, p = .23). Direct effects from the intervention to psychological IPA likelihood (β = .06, p = .46) and physical IPA likelihood (β = −.03, p = .71) also were not statistically significant.
Discussion
This study aimed to evaluate the preliminary efficacy of a brief web-based CR intervention to reduce psychological and physical IPA likelihood compared to a control condition among high-risk young adult men. Although the intervention was designed to reduce sexual assault perpetration likelihood, rather than psychological and physical aggression, the specific skills targeted in this intervention were found to be helpful for reducing psychological IPA likelihood via increased ability to clarify one’s emotions. This effect emerged independent of whether men were intoxicated or not (i.e., beverage administration was added as a covariate to the analysis), thereby increasing external validity given that many men may be intoxicated in sexual situations in the real world (Abbey, 2002). There were no effects of the intervention on physical IPA perpetration.
The findings regarding psychological IPA likelihood are in line with theoretical approaches (e.g., Finkel, 2011, 2014; Finkel & Eckhardt, 2013) and prior research (e.g., Neilson et al., 2023) linking emotional clarity and IPA. Using the CR skills taught in the intervention may have helped men in effectively overcoming aggressive urges due to improved ability to clarify what they were feeling, thereby resulting in decreased psychological IPA likelihood following the intervention. The absence of an effect of the intervention on physical IPA intentions may be related to the overall low base rate of physical IPA intentions in the sample. Future research with samples specifically selected on the basis of physical IPA risk is needed to further examine the potential usefulness of the current CR intervention for this outcome.
The current results should be interpreted in light of some limitations. First, our inclusion criteria limit generalizability of the present findings to single, male, sexually active drinkers aged 21–30 who are interested in dating women and have a history of sexual aggression perpetration. Thus, it is unknown whether results would generalize to other groups of individuals not considered here. Similarly, the majority of our sample identified as White; thus, generalizability to groups of more diverse backgrounds remains unknown. Second, although the laboratory-based procedures enabled rigorous testing of ER mechanisms and intervention effects, they preclude our ability to evaluate hypotheses in real-world conditions on actual IPA behavior. Next steps in this line of research include expansion to a randomized controlled trial that tests intervention effects on real-world IPA perpetration in a more diverse sample.
Despite these limitations, we find the current preliminary findings in support of this CR intervention for reducing IPA promising. If replicated, these results have important clinical implications. Specifically, CR skills may help individuals clarify their emotions, which, in turn, may reduce the risk that they engage in psychological IPA. The current CR intervention may provide multiple benefits over existing batterers interventions based on Cognitive Behavioral Therapy, which sometimes involve CR components, yet often evidence high recidivism rates (Stover et al., 2009). Specifically, the brief and web-based format of the current intervention decreases potential barriers to dissemination and implementation. Future research could examine intervention effects when administered at repeated intervals (or as “booster” sessions), thereby leading to potential decreases in recidivism.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interests with respect to the authorship and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research and/or authorship of this article: Funding was provided by the National Institute on Alcohol Abuse and Alcoholism to Cynthia A. Stappenbeck (R01AA027994) and Kelly Cue Davis (R21AA023811; 2R37AA025212).
