Research Review: Omega-3 supplementation to reduce antisocial behavior - a systematic review and meta-analysis of randomized controlled trials.

Bystander intervention (BI) is a common approach to promoting prosocial behavior and represents one of the most widely-examined social-psychological phenomena in the past century. Rigorous evaluation efforts and meta-analytic findings provide compelling evidence that BI is an efficacious and effective strategy for preventing sexual violence by increasing awareness of contextual risk cues, enhancing motivation for proactive intervention, and providing opportunity for skill acquisition and practice (e.g., Coker et al., ; Jouriles et al., ; Kettrey & Marx, ; Park & Kim, ); however, the scope of this evidence remains limited, as most evaluations focus on creating shifts in bystander attitudes, intentions, confidence, and self-efficacy following program participation (Bush et al., ). Similarly, existing BI prevention programs have been criticized as primarily impacting to intervene, as opposed to influencing rates of BI action or reducing violence within communities (Porat et al., ). Given the well-documented intention–behavior gap in behavior-change research (Sheeran & Webb, ), a focus on bystander attitudes and intentions provides limited insight into whether—and under what conditions—intervention actually occurs. Seminal findings from Darley & Latané’s () research emphasized numerous individual-difference and situational barriers to intervening, including diffusion of responsibility, evaluation apprehension, and pluralistic ignorance. More recent research has identified a plethora of other barriers and facilitators to BI (Bennett & Banyard, ; Oesterle et al., ; Orchowski et al., ); however these findings are limited in their real-world applicability. Given the complex, dynamic, and multifaceted nature of sexual violence, effective prevention efforts require empirical research that elucidates the real-world, proximal mechanisms and contextual factors shaping actual BI behavior; ecological momentary assessment (EMA) is uniquely positioned to address this need by capturing real-time data within the high-risk contexts in which BI decisions occur. Research has begun to move beyond descriptive cataloguing of barriers and facilitators to BI (e.g., Park et al., ) to more advanced experimental approaches that account for time-varying effects (e.g., Clay-Warner et al., ; Huang et al., ). These studies demonstrate that changing contextual and situational factors (e.g., gender, severity, etc.) influence likelihood intentions of using BI. Yet even these experimental approaches are unable to successfully replicate the complexities of real-life situations, making it difficult to determine the proximal determinants of BI. Although experimental analogues can assess the temporal dynamics of situational processes and BI in a controlled setting, they fundamentally fail to replicate the acute stress, anxiety, and fear inherent in real-world emergencies, often rendering them ineffective at eliciting authentic levels of affective and emotional intensity (Gray et al., ). Even when examining real-time substance use, existing studies have operationalized BI using outcomes that remain psychologically proximal but behaviorally distal, rather than capturing intervention as it occurs in-vivo (Leone et al., ; Wiseblatt et al., ). The Situational Model (Burn, )—which is one of the most widely used explanatory frameworks for BI—outlines a sequence of decisions that a bystander must make before acting to prevent or stop sexual violence: (a) noticing the event, (b) interpreting it as problematic, (c) assuming responsibility, (d) knowing how to intervene, and (e) taking action. While existing prevention programs often draw on this model to address barriers to intervention, critics have noted its limited applicability to more complex scenarios involving friends or acquaintances, wherein social dynamics and relational factors may inhibit intervention (Banyard, ). In response, alternative theoretical models—such as the Theory of Action Coils (Banyard, ) and the Bystander Intervention for Problematic Alcohol Use Model (Mennicke et al., , )—have emphasized the convergence of decision-making processes, situational complexity, and event-level characteristics in promoting bystander behavior. Although these innovative frameworks offer meaningful theoretical and conceptual contributions to our understanding of BI, theoretical models that explicitly incorporate acute alcohol intoxication as a momentary event-level barrier—despite its frequent co-occurrence within high-risk contexts for sexual violence (Abbey et al., ; Cleveland et al., )—remain underdeveloped. Metatheoretical frameworks such as the I³ Model (Finkel & Hall, ), however, may provide a useful organizational template for researchers to examine the interactive effects of affective, behavioral, and cognitive processes on BI under varying levels of intoxication (Finkel & Eckhardt, ; Parrott & Eckhardt, ). Therefore, advancing BI theory and prevention requires identifying for , and , proximal co