A Systematic Review of Validated Measures of Bystander Related Constructs

Background and purpose: As bystander intervention has proliferated as a promising approach to prevent multiple forms of harm across various settings, so too have the techniques used to measure this complex phenomenon. Simultaneously, theoretical frameworks have evolved to shed innovative light on the intricate constructs influencing when, why, and how a bystander decides to intervene. As a result, there is limited uniformity within or across fields regarding bystander behavior measurement best practices, specifically those ensuring measures capture the intricate nuances of bystander behaviors and related constructs. This lack of consistency is of particular concern, given that these measures are used to design and evaluate the effectiveness of interventions to reduce harm. Thus, this study aimed to identify all validated US-based measures of bystander-related behaviors and constructs to identify gaps and strengths in measurement and provide the field of bystander intervention with recommendations regarding measurement advancements. Three primary research questions informed the analysis: 1) What are the general characteristics of the scales? 2) What are the item-specific characteristics of the scales? And 3) What are the characteristics of the samples used to validate the measures? Methods. Using PRISMA-P guidelines, electronic databases were searched, and emails were solicited to identify articles containing bystander scales. To meet inclusion criteria, articles had to (a) measure at least one bystander behavior or related construct in the context of imminent harm, b) be written in English, and c) be statistically validated on US-based samples. The initial search produced 8,559 articles. Following title, abstract, and full-text screening, 24 scales met all criteria for inclusion. Results. Regarding scale characteristics, most scales measured bystander behavior within the domain of interpersonal violence (67%), most scales assessed intent/willingness/likelihood to intervene (50%), and most scales captured behavior from the “taking action” step of Darley and Latane’s situational model. Regarding item-specific characteristics, on average, scales contained 14 items, and most scales (79%) provided Likert-style response options. Regarding sample characteristics, scales were validated using samples comprised of predominantly young (21.8 years old), White (75%), and heterosexual (89%) women (64%). Conclusions and implications. Overall, this systematic review demonstrates that, despite recent rapid growth in the field of bystander intervention, measurement of bystander behaviors and related constructs remains limited. Results indicate a need to develop validated measures that span the breadth of domains of harm and bystander-related constructs. Additionally, there is an acute need to increase the diversity of measurement validation samples, as the majority of available measures were validated on young, cisgender, heterosexual, White females. Given that individuals from marginalized communities have increased exposure to violence, different opportunities for intervention, distinct barriers to action, and unique perspectives on viable and safe intervention strategies, validated measures recognizing these variations must become available. This advancement will allow for a deeper understanding of the bystander behavior process and the creation of a multitude of tailored interventions to confidently prevent harm for all.