Initial Development and Validation of the Bullyharm: The Bullying, Harassment, and Aggression Receipt Measure

Background and Purpose: Bullying is a significant social problem that can negatively affect the physical, social, emotional, and educational well-being of youth.  Research demonstrates that victims of bullying are more likely to report depression, anxiety, low self-esteem, psychosomatic problems, loneliness, suicidal ideation and behavior, tobacco and alcohol use, oppositional behavior, lower academic achievement, missing school, and dropping out of school. 

Instruments that accurately and reliably measure bullying are essential for the development and evaluation of interventions.  However, critical reviews of bullying instruments indicate problems with existing measures, including failure to include all types or dimensions of bullying behaviors, inadequate psychometric properties or a lack of information about psychometric properties, and item characteristics that threaten score validity, such as confusing wording or lack of appropriate recall time frames. 

This paper describes the development and preliminary validation of the BullyHARM (Bullying, Harassment, and Aggression Receipt Measure), which was designed to assess students’ experiences of being bullied by peers in school settings. 

 

Methods: The development of the BullyHARM involved a number of steps and methods designed to promote validity of scores obtained from the instrument.  Scale development steps included a review of the bullying literature and existing measures, development of an item pool, expert review of items, cognitive interviewing with members of the target population (N = 5), readability testing, and data collection from a large sample.  A sample of 275 middle school students was used to examine descriptive statistics, psychometric properties, and the factor structure of the BullyHARM.

Results: The BullyHARM consisted of 22 items and 6 subscales.  Internal consistency reliability for the overall scale is good (α = .93) and for each subscale: physical bullying (α = .85), verbal bullying (α = .85), relational or social bullying (α = .82), cyber-bullying (α = .91), property bullying (α = .83), and sexual bullying (α = .85).  Mplus was used to perform a confirmatory factor analysis (CFA) to evaluate first-order and second-order factor models of the BullyHARM.  Results showed that the first-order factor model met all of the pre-stated model fit criteria (CFI ≥ 0.97, TLI ≥ 0.97, RMSEA ≤ 0.05, and WRMR ≤ 0.90) whereas the second-order factor model only met two of the criteria.  In addition, results of the DIFFTEST procedure comparing the models showed that the second-order factor model significantly worsened model fit, χ² (9) = 32.66, p < 0.05.  The findings indicate that the first-order model should be retained.

Conclusions and Implications: The steps used to develop the BullyHARM were designed to ensure that the scale measured all relevant dimensions of bullying victimization, and that scores were valid for members of the target population.  Analyses of scores from 275 students indicate that the BullyHARM has good properties regarding content validation and responded-related validation, and is a promising instrument for measuring students’ experiences of being bullied at school.