Peer defending as a multidimensional behavior: Development and validation of the Defending Behaviors Scale☆
Introduction
It is estimated that youth witness peer victimization (as bystanders) once every 7 min (Craig & Pepler, 1998). Recent Canadian estimates indicate that over the past month, over 50% of Grade 7–8 students witnessed bullying at least once in their school hallway or on school grounds (Lambe, Hudson, Craig, & Pepler, 2017). Bystanders are part of the complex social dynamics of bullying and can alter the situation in both negative and positive ways. While some bystanders assist with the bullying or remain passive, others intervene to support their victimized peers. Bystanders who intervene in bullying are known as “defenders,” and defending includes any prosocial behavior intended to assist the individual being victimized (Salmivalli, 2010). Observational research indicates that peer defending is typically effective in stopping bullying (Hawkins, Pepler, & Craig, 2001). Consequently, many intervention programs aim to reduce peer victimization by increasing peer defending (Polanin, Espelage, & Pigott, 2012). While the benefits of peer defending have been established (e.g., better adjustment for victimized youth; Sainio, Veenstra, Huitsing, & Salmivalli, 2010), little is known about the types of behaviors youth enact when they defend their peers. Previous research has almost exclusively focused on examining defending as an overarching category, rather than a multidimensional behavior. Using prosocial theories as a guide, the current study aimed to fill this gap by developing a theoretically informed and psychometrically sound instrument to assess peer defending behaviors.
Defending was first assessed as part of the group dynamics of bullying (Salmivalli, Lagerspetz, Björkqvist, Österman, & Kaukiainen, 1996). Since then, defending has received a marked increase in research attention (for a review see Lambe, Della Cioppa, Hong, & Craig, 2019). The majority of the bullying literature considers defending a homogenous role – youth are categorized as defenders or are assigned to another participant role (e.g., bully, victim, outsider). For example, the Participant Roles Scale (Salmivalli et al., 1996) uses peer nominations to assign youth to a particular role based on their highest score. This approach has provided insight into the complex group dynamics involved in bullying, as well as an understanding of the personal characteristics that are common for youth cast in the defender role. For example, youth who are nominated as defenders tend to be girls, are popular and well-liked by their peers, are highly empathic both in terms of understanding people's emotions (cognitive empathy) and vicariously experiencing other people's emotions (affective empathy (Blair, 2005)), and have a sense of moral duty (Lambe et al., 2019). Peer nomination procedures, such as those typically used to categorize defenders, are subject to less bias (e.g., social desirability) than self-report measures. This is beneficial as youth may over-report desirable behaviors (e.g., defending) and under-report undesirable behaviors (e.g., bullying). Categorical role approaches (e.g., Participant Roles Scale) have provided a strong foundation for our understanding of this complex behavior; however, such approaches have measurement limitations.
The defender scale originally put forth by Salmivalli et al. (1996) consisted of 20 items. A large number of items is needed to adequately capture the diverse ways in which youth can defend, including telling an adult, comforting the person being victimized, attacking the aggressive person, and telling the others to stop bullying (Salmivalli et al., 1996). Yet, subsequent modifications of this scale (e.g., Salmivalli, Lappalainen, & Lagerspetz, 1998) have reduced this diversity in item content with little justification, with some defending scales consisting of only three of the original 20 items (e.g., Pozzoli & Gini, 2010). To our knowledge, factor analytic methods have not yet been used to determine what items best represent the underlying construct of defending, and whether or not these items represent a single underlying factor. While brief scales may be useful for differentiating between bullying roles (e.g., bully, victim, defender, reinforcer, outsider), they lack the specificity needed to fully capture defending behaviors. A comprehensive defending scale is needed to best assess the full underlying construct.
Most research that measures defending using the Participant Roles Scale (Salmivalli et al., 1996) assigns youth to a categorical role, which may further limit our understanding of defending. Approximately 13% of youth are not assigned to a bullying participant role because they have equivalent scores on multiple roles (Pouwels, Lansu, & Cillessen, 2016; Salmivalli et al., 1996). In other words, bullying participant roles are dynamic – youth who defend in one situation may be more likely to be victimized in another situation. Depending on the peer context, defenders may be more likely to be victimized, or to defend aggressors rather than victimized youth (Huitsing, Snijders, Van Duijn, & Veenstra, 2014). The categorical role approach can produce different proportions of defenders depending on the method used to assign participant roles (Goossens, Olthof, & Dekker, 2006; Sutton & Smith, 1999). Using participants' highest overall role score, 44% of children are assigned to the defender role; in contrast, using standardized scoring (i.e., highest scale and higher than the class mean), 27% of children are assigned to the defender role (Sutton & Smith, 1999). This discrepancy suggests that many youth who enact defending behaviors may be assigned to other roles using traditional measurement techniques. Thus, the traditional measurement of defending may not be accurately reflecting all youth who defend their peers. A continuous measure of defending that does not categorize youth may best capture all youth who engage in defending behaviors.
Current research that specifically focuses on defending is exploring the heterogeneity of this construct. In a qualitative study examining bystander responses to teen dating violence and bullying, youth reported 16 broad categories of possible responses; however, the authors organized these behaviors along a direct-indirect continuum of behavior (Casey, Storer, & Herrenkohl, 2018). Befriending or checking-in with the person being victimized were among the most indirect defending behaviors, whereas removing the victimized person or fighting the perpetrator were among the most direct defending behaviors (Casey et al., 2018). While there are not yet agreed upon terms to define defending behaviors, we use the term “direct defending” to refer to behaviors that involve the defender directly confronting the aggressive situation (e.g., asking the aggressor to stop, pushing the aggressor away), whereas “indirect defending” refers to defending that does not involve the defender confronting the aggressive situation themselves (e.g., comforting the person who was being victimized, seeking help from a parent). Similar definitions have been used in other qualitative research; direct defending behaviors are more likely when youth perceive themselves to be more socially powerful than the aggressor, whereas indirect defending behaviors are more likely when youth report low self-efficacy (Forsberg et al., 2018). In sum, the qualitative research indicates that youth use a variety of defending behaviors to intervene against bullying.
These distinctions in defending behaviors are further supported by quantitative work. Reijntjes et al. (2016) argue that there are both theoretical and empirical grounds to distinguish between types of defending behaviors – victim-oriented defending (e.g., consoling the person being victimized) and bully-oriented defending (e.g., actively intervening against the aggressor). Victim-oriented defending represents indirect defending, whereas bully-oriented defending represents direct defending. Although limited by single-item measurement, their research demonstrated significant differences between victim-oriented, bully-oriented, overall defenders (high on both behaviors), and non-defenders. For example, overall and bully-oriented defenders scored highest on popularity, whereas overall and victim-oriented defenders scored highest on peer acceptance (Reijntjes et al., 2016). Subsequent research using a similar, single-item scale found that indirect defending is associated with altruistic concern for victimized youth, whereas direct defending is associated strategic goals and personal gain (Pronk, Olthof, Goossens, & Krabbendam, 2019). Thus, direct and indirect defending behaviors appear to represent unique underlying constructs.
In addition to direct and indirect forms of defending, other researchers argue that aggressive defending represents another distinct form of defending (Frey, Pearson, & Cohen, 2014; Meter, Ma, & Ehrenreich, 2019). While Salmivalli et al.'s (1996) original defender scale contained both prosocial and aggressive forms of defending, subsequent modifications to this scale (e.g., Pozzoli & Gini, 2010; Salmivalli et al., 1998) have omitted these items with no justification. Observational research indicates that aggression is commonly employed by children as a means of defending their peers (Hawkins et al., 2001). In a study of defending among college students, Meter et al. (2019) demonstrated that relationally and verbally aggressive defending behaviors demonstrated unique associations with direct and indirect defending, as well as with moral disengagement. Together, these findings suggest that defending includes both prosocial and aggressive behaviors.
Examining defending as a heterogeneous behavior has provided a greater understanding of how youth defend their peers, however, previous research (Meter et al., 2019; Pronk et al., 2019; Reijntjes et al., 2016) is limited by its use of single-item measures (i.e., one item to measure each behavior). Using a single item to asses direct and indirect defending may be providing us with a biased conceptualization, as there are theoretically many different ways youth can defend using both direct and indirect behaviors. For example, boys are more likely to use physical aggression as their primary defending behavior, whereas girls are more likely to use verbal assertion (Hawkins et al., 2001). While these defending behaviors are arguably both direct in nature, they may represent different underlying constructs. It is unknown whether defending behaviors are best differentiated by direct and indirect factors, or whether more factors are needed to understand the multidimensional nature of defending. While single-item measures can successfully measure unambiguous constructs, they typically lack breadth to measure complex psychological constructs like peer defending. Scale development should be guided by theory and include enough breadth in content to successfully represent the construct (Carpenter, 2018). A theoretical approach to understanding the distinctions between direct and indirect forms of defending may help overcome these limitations to more fully understand the heterogeneity within defending behaviors.
Previous research indicates that direct and indirect defending behaviors are likely distinct (Meter et al., 2019; Pronk et al., 2019; Reijntjes et al., 2016). However, there is not yet a strong theoretical rationale behind this distinction. Defending is a form of prosocial behavior that occurs specifically in response to witnessing peer victimization; thus, we propose that prosocial theory can be applied to understand differences in defending behaviors. Contemporary theories of prosocial behavior argue that helping is a broad category that subsumes many related, yet unique, types of behavior (Eisenberg & Spinrad, 2014). In other words, prosocial behavior is multidimensional.
While types of prosocial behaviors are related, they can be distinguished through unique antecedents, correlates, and outcomes (Eisenberg & Spinrad, 2014). Prosocial theory posits that types of helping behaviors can be differentiated by many factors, including personal resources (Padilla-Walker & Fraser, 2014), physiological and affective arousal (Hoffman, 1989; Miller, 2018), and socialization experiences (Eisenberg & Spinrad, 2014). Furthermore, prosociality is defined by the intent to help another, allowing for the inclusion of aggressive defending. Indeed, punishing transgressors is a common way in which bystanders intend to help victimized individuals (O'Gorman, Wilson, & Miller, 2005; Vitaglione & Barnett, 2003). Altruistic/third-party punishment can be conceptualized as a moral, helping behavior that serves to maintain social norms (Dimitroff et al., 2019). We propose that defending subtypes can be distinguished in similar ways as other forms of prosocial behavior. In other words, individual differences in personal resources (e.g., social self-efficacy), affect (e.g., reactive aggression, empathy), and socialization experiences (e.g., social support) may distinguish between those who are more likely to use direct and indirect types of defending.
Taken together, defending may be best conceptualized as a multidimensional behavior, consisting of both direct and indirect types of defending. Conceptualizing and measuring defending as a multidimensional behavior has important implications for adults working with young people, such as school administrators. Bullying prevention and intervention programs often encourage peer defending without having a clear understanding of what defending behavior looks like or how different defending strategies may impact youth in different ways. We aimed to fill this gap by developing the Defending Behaviors Scale (DBS) and establishing psychometric evidence for the measure.
Our first research goal was to examine the underlying factor structure of the DBS. We hypothesized that a multifactor model with separate direct and indirect defending factors would best fit the data. Our second research goal was to establish the psychometric properties of each of these defending subtypes. Specifically, we aimed to describe the nature of each factor (e.g., prevalence and gender differences), examine reliability, and examine validity by testing relationships with existing constructs (i.e., bullying, victimization, empathy, aggression, social self-efficacy, social support, and prosocial behavior). We hypothesized that each factor would be both reliable and valid, as evidenced by high internal consistency reliability (>0.70), test-retest reliability (>0.70), and unique associations with established measures (convergent and discriminant validity; Carpenter, 2018; Morgado, Meireles, Neves, Amaral, & Ferreira, 2017). Specifically, we hypothesized that direct defending would be positively associated with social self-efficacy and reactive aggression (Pronk et al., 2019; Reijntjes et al., 2016), and would be more common among boys (Reijntjes et al., 2016). We also hypothesized that indirect defending would be positively associated with empathy and social support (Lambe et al., 2019; van Noorden, Haselager, Cillessen, & Bukowski, 2014) and would be more common among girls (Reijntjes et al., 2016). Links with victimization and bullying (offline and online), and prosocial behavior were examined as exploratory research questions. Lastly, given the established gender differences in defending, we aimed to test measurement invariance across gender.
Section snippets
Participants
A total of 572 participants in Grades 6–8 completed the DBS. Data collection occurred in two parts: a school sample (n = 334) and a community sample (n = 238). Using chi-square tests, these samples did not differ in terms of gender (χ2 = 2.79, p = .25) or grade (χ2 = 1.98, p = .37). The samples also did not differ in terms of age (t = 1.49, p = .14). Thus, sample characteristics are presented for the total sample. Overall, participants ranged in age from 11 to 14 years old, with an average age
Inter-item correlations and descriptives
Item correlations and descriptive statistics for the initial 20-items are shown in Table 1. Item correlations ranged from 0.13 to 0.77, with the majority being moderate in range. In general, items reflecting aggressive defending (2, 6, 10, 14, 18) demonstrated the weakest correlations with the other defending items; however, these items were all moderately correlated with each other. At the item level, item 4 (“I was friendly to the person being picked on”) was the most frequently endorsed
Discussion
Defending represents any prosocial action intended to help a victimized peer. While previous research supports a distinction between direct and indirect forms of defending, the current study was the first to rigorously examine the psychometric properties of a multi-item scale of defending – the Defending Behaviors Scale (DBS). While results supported a multidimensional model of defending behaviors, the factors were more distinct than direct and indirect behaviors. A four-factor model of
Acknowledgements
The authors thank Irene Hong, Thomas Vaughn-Johnston, Natalie Mangialardi, Theo Vissers, Veronica Johnson, Emily Yule, and Aliya Ali for their assistance with data collection.
Funding
Laura J. Lambe was supported by a Social Science and Humanities Research Council (SSHRC) Vanier Canada Graduate Scholarship. Wendy Craig was funded by the Canadian Foundation for Innovation and the Social Sciences and Humanities Council.
References (70)
- et al.
A two-method investigation of early adolescents’ responses upon witnessing peer victimization in school
Journal of Adolescence
(2012) Responding to the emotions of others: Dissociating forms of empathy through the study of typical and psychiatric populations
Consciousness and Cognition
(2005)- et al.
Standing up to bullying: A social ecological review of peer defending in offline and online contexts
Aggression and Violent Behavior
(2019) - et al.
Does defending come with a cost? Examining the psychosocial correlates of defending behaviour among bystanders of bullying in a Canadian sample
Child Abuse & Neglect
(2017) - et al.
Defenders of victims of peer aggression: Interdependence theory and an exploration of individual, interpersonal, and contextual effects on the defender participant role
Developmental Review
(2015) Physiological mechanisms of prosociality
Current Opinion in Psychology
(2018)- et al.
How do men and women help? Validation of a multidimensional measure of prosocial behavior
Journal of Adolescence
(2017) - et al.
Altruistic punishing and helping differ in sensitivity to relatedness, friendship, and future interactions
Evolution and Human Behavior
(2005) - et al.
How much is it going to cost me? Bidirectional relations between adolescents’ moral personality and prosocial behavior
Journal of Adolescence
(2014) - et al.
Measurement invariance conventions and reporting: The state of the art and future directions for psychological research
Developmental Review
(2016)
Bullying and the peer group: A review
Aggression and Violent Behavior
Examining the validity and reliability of the cyber-aggression and cyber-victimization scale
Computers in Human Behavior
Bystander behavior in bullying situations: Basic moral sensitivity, moral disengagement and defender self-efficacy
Journal of Adolescence
Emotion regulation flexibility
Cognitive Therapy and Research
Exploratory structural equation modeling
Structural Equation Modeling
Predictors of student defenders of peer aggression victims: Empathy and social cognitive factors
International Journal of Behavioral Development
Differences in proactive and reactive aggression in traditional bullies and cyberbullies
Journal of Aggression, Maltreatment & Trauma
Sociocognitive and behavioral correlates of a measure of prosocial tendencies for adolescents
Journal of Early Adolescence
The multidimensionality of prosocial behaviors and evidence of measurement equivalence in Mexican American and European American early adolescents
Journal of Research on Adolescence
Ten steps in scale development and reporting: A guide for researchers
Communication Methods and Measures
Mapping a continuum of adolescent helping and bystander behavior within the context of dating violence and bullying
American Journal of Orthopsychiatry
A first course in factor analysis
Social self-efficacy in adolescence: Relations with self-concept, social adjustment, and mental health
Canadian Journal of Behavioural Science
Pow! Boom! Kablam! Effects of viewing superhero programs on aggressive, prosocial, and defending behaviors in preschool children
Journal of Abnormal Child Psychology
A cross-national profile of bullying and victimization among adolescents in 40 countries
International Journal of Public Health
Observations of bullying and victimization in the school yard
Canadian Journal of School Psychology
Measuring individual differences in empathy: Evidence for a multidimensional approach
Journal of Personality and Social Psychology
Perceptions of the frequency and importance of social support by students classified as victims, bullies, and bully/victims in an urban middle school
School Psychology Review
Bullying Participant Behaviors Questionnaire (BPBQ): Establishing a reliable and valid measure
Journal of School Violence
Third-party punishment following observed social rejection
Emotion
Gender and helping behavior: A meta-analytic review of the social psychological literature
Psychological Bulletin
Multidimensionality of prosocial behavior
Willingness to intervene in bullying episodes among middle school students
The Journal of Early Adolescence
Exploratory factor analysis
Students’ views of factors affecting their bystander behaviors in response to school bullying: a cross-collaborative conceptual qualitative analysis
Research Papers in Education
Cited by (39)
Trajectories of defending behaviors: Longitudinal association with normative and social adjustment and self-perceived popularity
2023, Journal of School PsychologyEffects of the KiVa anti-bullying program on defending behavior: Investigating individual-level mechanisms of change
2023, Journal of School PsychologyLongitudinal associations of social-cognitive and moral correlates with defending in bullying
2022, Journal of School PsychologyCitation Excerpt :Furthermore, following many studies in this field, we measured defending in general, without distinguishing particular types of defending behavior or different targets of this behavior. It has been recently claimed the importance of differentiating, for example, between bully-oriented defending and victim-orienting defending (e.g., Lambe & Craig, 2020). Moreover, bystanders' responses to bullying might vary according to who is engaging in bullying or whom is being bullied (e.g., a same vs. opposite sex peer, a friend vs. an acquaintance; Oldenburg et al., 2018).
- ☆
This article is part of the special issue Bystanders in Bullying and Peer Victimization; Edited by Dr. Lyndsay Jenkins and Dr. Dr. Wendy Troop-Gordon.