Modified Aggression Scale

Abstract

The Modified Aggression Scale (MAS) is a specialized psychometric assessment instrument designed to measure aggressive behaviors, anger reactivity, and prosocial compensatory behaviors among adolescents, particularly middle school students in grades 6 through 8 (ages 11–15). Originally derived from the Aggression Scale formulated by Pamela Orpinas (1993) at the University of Texas Health Science Center at Houston and further modified by Kris Bosworth and Dorothy L. Espelage (1995) as part of the Teen Conflict Survey at Indiana University, the instrument captures both violent behavioral manifestations and positive interpersonal alternatives within school environments. The assessment consists of 22 self-report items distributed across four distinct subscales: Fighting (5 items), Bullying (4 items), Anger (5 items), and Caring/Cooperative Behavior (8 items). Respondents quantify their behavior over a designated recall period of the past 30 days using a 5-point frequency scale ranging from “No opportunity” / “Never” (scored as 1) to “5 or more times” (scored as 4).

Extensive psychometric investigations published across developmental psychology and violence prevention literature—including documentation in the Centers for Disease Control and Prevention (CDC) Compendium of Assessment Tools—demonstrate that the MAS possesses robust internal consistency, with subscale Cronbach’s alpha coefficients typically ranging from .73 to .87, and an overall aggression construct alpha exceeding .85. Confirmatory factor analytic investigations support an oblique multidimensional structure that effectively differentiates overt physical conflict, relational and verbal bullying, internal affective arousal, and prosocial peer interactions. This article provides a comprehensive academic review of the Modified Aggression Scale, delineating its historical emergence, theoretical foundations in Social Cognitive Theory and Social Information Processing models, psychometric reliability and validity data, scoring mechanics, and practical applications in adolescent violence prevention research and school psychology.

Keywords

Modified Aggression Scale, youth violence, adolescent aggression, bullying measurement, peer conflict, anger reactivity, prosocial behavior, psychometrics, Middle School violence prevention, Teen Conflict Survey

Authors

The Modified Aggression Scale emerged from collaborative and sequential research programs focused on youth violence prevention in school settings:

• Pamela Orpinas, Ph.D., M.P.H. — Professor Emerita of Health Promotion and Behavior, College of Public Health, University of Georgia. Dr. Orpinas developed the foundational Aggression Scale in 1993 during her doctoral dissertation research at the University of Texas Health Science Center at Houston, School of Public Health, evaluating skills training and social influence curricula for violence prevention in urban middle schools.
• Kris Bosworth, Ph.D. — Professor and Smith Endowed Chair in Substance Abuse Education, College of Education, University of Arizona; formerly Director of the Center for Adolescent Studies at Indiana University Bloomington. Dr. Bosworth co-developed the Teen Conflict Survey, which integrated and expanded Orpinas’s aggression items with anger and prosocial behavioral indicators.
• Dorothy L. Espelage, Ph.D. — William C. Friday Distinguished Professor of Education at the University of North Carolina at Chapel Hill; previously at the University of Florida and the University of Illinois Urbana-Champaign. Dr. Espelage co-authored the 1995 Teen Conflict Survey modification and has published foundational longitudinal studies examining the psychometric structure of peer aggression, bullying perpetration, and school climate using this instrument.

Purpose

Adolescence represents a critical developmental period marked by rapid neurobiological maturation, shifting social affiliations, and heightened vulnerability to interpersonal conflict. During the middle school transition, the emergence of overt aggression, social exclusion, relational bullying, and physical confrontations poses substantial risks to academic attainment, socioemotional health, and physical safety. In response to the escalation of youth violence in the 1990s, educational researchers and public health entities required an empirically validated, brief, and developmentally tailored instrument capable of assessing both antisocial conduct and prosocial alternatives among early adolescents.

The primary purpose of the Modified Aggression Scale is to provide school psychologists, public health interventionists, and academic researchers with a standardized metric to:

• Quantify Specific Aggressive Behaviors: Accurately register the frequency of physical combat, verbal intimidation, and relational victimization behaviors occurring within the immediate peer ecosystem over a concrete 30-day temporal window.
• Disentangle Affective Arousal from Behavioral Enactment: Differentiate internal emotional states (trait-like or sustained emotional anger) from overt behavioral conduct (physical fighting, teasing, threatening peers).
• Measure Prosocial and De-escalation Competencies: Track positive social strategies (such as helping peers stay out of fights, offering compliments, cooperative problem solving, and walking away from volatile disputes) to evaluate whether violence prevention interventions increase protective behavioral repertoires alongside reducing hostile behaviors.
• Evaluate Universal and Targeted Interventions: Serve as an outcome measurement battery for middle school violence prevention curricula (e.g., universal social-emotional learning programs, conflict resolution training, and peer mediation initiatives), determining programmatic efficacy across longitudinal baseline and follow-up intervals.
• Identify At-Risk Cohorts for Secondary Prevention: Assist multi-tiered systems of support (MTSS) in identifying students who exhibit elevated levels of reactive physical combat or predatory bullying, thereby facilitating targeted socioemotional skills training before disciplinary exclusion becomes necessary.

Psychological Construct

The Modified Aggression Scale assesses an interrelated yet conceptually divergent matrix of interpersonal behaviors and emotional states. Rather than conceptualizing aggression as a unidimensional drive, the MAS operationalizes the construct as a multi-component behavioral complex encompassing physical violence, peer harassment, emotional dysregulation, and protective prosocial buffering.

1. Fighting (Overt Physical Aggression)

The Fighting subscale operationalizes overt, high-intensity physical conflicts between peers. Physical aggression involves behavioral acts directed toward another individual with the immediate intent to cause bodily harm, pain, or discomfort. The items capture reactive combat (e.g., hitting back when struck first), instigation (e.g., encouraging others to fight), overt violent acts (e.g., pushing, shoving, slapping, or kicking), and anger-induced physical confrontation. Uniquely, the subscale incorporates a behavioral inhibition item (“I walked away from a fight”), which is reverse-coded to capture the failure or success of conflict de-escalation in situ.

2. Bullying (Relational and Intimidating Peer Aggression)

The Bullying subscale captures repetitive, power-asymmetric social conduct characterized by humiliation, verbal degradation, and explicit coercion. Rooted in developmental peer relations research, this construct distinguishes between impulsive physical brawling and systematic peer victimization. The items assess teasing peers, weaponizing social humor to demean targets for peer group amusement, name-calling, and verbalized threats of physical harm. This captures both direct verbal aggression and indirect relational hostility aimed at degrading another student’s social status.

3. Anger (Affective Reactivity and Dysregulation)

The Anger subscale addresses the subjective affective and physiological state that frequently precipitates or accompanies hostile behavioral acts. Anger is defined within this operationalization as persistent negative emotional arousal, subjective feelings of hostility, and low threshold for frustration. The subscale monitors the chronic prevalence of anger (“frequently get angry”, “angry most of the day”), the direct behavioral translation of anger into physical brawling, and displaced aggression—specifically, being mean to secondary parties or taking emotional frustration out on innocent bystanders.

4. Caring and Cooperative Behavior (Prosocial Counter-Aggression)

The Caring/Cooperative Behavior subscale measures active prosocial competence, perspective-taking, positive communication, and bystander intervention. In contemporary developmental psychopathology, prosocial skills are recognized not merely as the inverse of aggression, but as an independent behavioral system capable of co-occurring with or buffering against hostile behavior. Items assess conflict de-escalation mediation (helping someone avoid a fight), affective assertion (communicating positive and negative emotional states constructively), active bystander defense (protecting a victim from a bully), interpersonal affirmation (giving compliments), collaborative problem solving, and institutional compliance (avoiding disciplinary infractions across home, school, and community environments).

Theoretical Framework

The architecture of the Modified Aggression Scale is anchored in multiple foundational paradigms within developmental psychology and behavioral science:

1. Social Cognitive Theory of Aggression

Grounded in the work of Albert Bandura (1986), Social Cognitive Theory asserts that human behavior is shaped by triadic reciprocal causation among cognitive/affective factors, behavioral patterns, and environmental influences. The MAS reflects Bandura’s modeling and vicarious reinforcement dynamics: students learn aggressive strategies—such as encouraging peers to brawl, verbally ridiculing classmates, or retaliating with physical force—through the observation of peer norms, media models, and domestic environments. Concurrently, the inclusion of the Caring/Cooperative subscale operationalizes self-regulatory efficacy, assessing an adolescent’s perceived behavioral agency to choose non-violent alternatives (e.g., walking away or defusing peer conflict) in an emotionally charged peer climate.

2. Social Information Processing (SIP) Model

The Social Information Processing model articulated by Nicki R. Crick and Kenneth A. Dodge (1994) provides a precise cognitive etiology for the behaviors measured by the MAS. According to Crick and Dodge, behavioral responses to peer provocation unfold through a six-step cognitive sequence:

1. Encoding of external and internal cues (e.g., noticing an accidental push in the school hallway);
2. Interpretation of cues (e.g., hostile attribution bias, inferring deliberate malicious intent);
3. Clarification of goals (e.g., status preservation, revenge, or conflict avoidance);
4. Response access or construction (retrieving behavioral scripts from long-term memory, such as striking back vs. verbal assertion);
5. Response decision (evaluating outcome expectancies and self-efficacy);
6. Behavioral enactment.

Within this framework, the MAS captures variations across these processing stages. Items assessing retaliatory violence (“I hit back when someone hit me first”) and displaced emotional hostility (“I took my anger out on an innocent person”) represent deficits in cue interpretation and response decision-making, while the prosocial items capture advanced cognitive problem-solving scripts.

3. The General Aggression Model (GAM)

Formulated by Craig A. Anderson and Brad J. Bushman (2002), the General Aggression Model integrates affective, cognitive, and arousal variables into a unified framework. The GAM posits that situational inputs (e.g., peer insults, physical proximity) interact with individual personological traits (e.g., chronic hostility, irritability) to influence an individual’s present internal state across three interconnected routes: cognition, affect, and arousal. The MAS explicitly isolates this internal state via the Anger subscale and measures its appraisal and decision outcomes through the Fighting and Bullying subscales.

4. Socio-Ecological Systems Theory

Pioneered by Urie Bronfenbrenner, Ecological Systems Theory emphasizes that youth behavior cannot be evaluated in isolation from its immediate microsystems (family, school classrooms, peer cliques). The MAS items operationalize behaviors embedded directly in the social fabric of the middle school microsystem (e.g., “I made fun of them to make other students laugh”, “I protected someone from a bully”), capturing how peer status dynamics and normative school climates govern behavioral frequency.

Validity

Empirical evaluations of the Modified Aggression Scale across diverse, large-scale adolescent samples demonstrate robust validity across multiple measurement domains:

Construct and Structural Validity

Construct validity has been established by confirming expected latent relationships between aggressive behavior, affective anger, and prosocial orientation. The Fighting and Bullying subscales exhibit strong, statistically significant positive intercorrelations (typically ranging between r = .55 and .68, p < .001), indicating that while they share common variance related to general antisocial conduct, they represent distinguishable behavioral forms. Conversely, the Caring/Cooperative Behavior subscale correlates negatively with both Fighting (r = -.24 to -.38) and Bullying (r = -.28 to -.42), confirming that prosocial competencies function as theoretically coherent opposing constructs without being a perfect inverse of aggression.

Convergent and Concurrent Validity

The convergent validity of the MAS has been documented against multiple validated behavioral, clinical, and disciplinary indices:

• School Disciplinary Records: In validation cohorts documented by Bosworth and Espelage (1995) as well as Orpinas (1993), composite scores on the Fighting and Bullying subscales correlated significantly with official school disciplinary referrals (r = .38 to .49, p < .001) and out-of-school suspensions.
• Substance Use and Delinquency: Aggression and Bullying scores correlate moderately to strongly with externalizing scales measuring adolescent alcohol consumption, marijuana experimentation, and petty delinquency (r = .40 to .55).
• Hostile Attribution and Conflict Beliefs: In studies evaluating the Teen Conflict Survey, high scores on the Anger and Fighting subscales were strongly associated with endorsement of pro-violence attitudes (e.g., “beliefs supporting aggression”, r = .52, p < .001).

Discriminant Validity

The scale effectively discriminates between adolescent cohorts characterized by non-aggressive sociometric status, reactive fighters, and persistent bullying perpetrators. Furthermore, the Anger subscale demonstrates adequate discriminant validity from generalized depressive symptom inventories (e.g., the Children’s Depression Inventory); although shared negative affectivity yields moderate baseline correlations (r ≈ .35), anger accounts for unique, incremental variance in predicting overt fighting behaviors after controlling for internalizing symptoms.

Predictive and Longitudinal Validity

Longitudinal investigations conducted by Espelage and colleagues demonstrated that middle school students exhibiting elevated scores on the Bullying and Anger subscales during sixth grade exhibited significantly elevated odds of self-reported physical fighting, carrying weapons to school, and experiencing peer rejection in eighth grade. Moreover, the Caring/Cooperative Behavior subscale significantly predicted lower odds of subsequent juvenile justice involvement and higher teacher-rated school bonding over a two-year tracking period.

Reliability

The Modified Aggression Scale has undergone extensive empirical reliability testing across heterogeneous geographic, socioeconomic, and racial/ethnic youth populations:

Internal Consistency

Across numerous published studies, Cronbach’s alpha (α) coefficients have demonstrated that all four subscales meet or exceed standard psychometric thresholds for behavioral research:

• Fighting Subscale (5 items): Cronbach’s α typically ranges between .73 and .83. The internal consistency remains stable across both male and female subsamples, although baseline mean scores are routinely higher among boys.
• Bullying Subscale (4 items): Cronbach’s α ranges from .81 to .87, reflecting high item homogeneity and coherent measurement of verbal, social, and coercive intimidation.
• Anger Subscale (5 items): Cronbach’s α ranges between .76 and .82, indicating reliable operationalization of affective irritability and displaced hostility.
• Caring/Cooperative Behavior Subscale (8 items): Cronbach’s α consistently ranges from .74 to .80, reflecting solid internal coherence across diverse prosocial actions.
• Aggression Composite: When the 14 items comprising the Fighting, Bullying, and Anger subscales are aggregated to form a generalized index of aggressive tendencies, total internal consistency reaches an alpha of .88 to .91.

Test-Retest Reliability

Stability assessments over brief intervals (2 to 4 weeks) in non-intervention control cohorts have yielded test-retest reliability correlation coefficients (r) ranging from .70 to .85 across subscales. Over longer multi-month intervals (e.g., 6 to 12 months), stability coefficients remain moderate (r = .45 to .62), reflecting expected developmental plasticity, environmental changes in peer groups, and the impact of school-wide disciplinary or preventive interventions.

Factor Analysis

The internal dimensionality of the Modified Aggression Scale has been repeatedly evaluated via both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Structure

Initial principal components and principal axis factoring using oblique rotations (e.g., Promax or Oblimin, accommodating the theoretical intercorrelation among aggressive traits) consistently reveal a four-factor solution accounting for approximately 52% to 61% of the total cumulative variance in middle school samples:

• Factor 1: Caring/Cooperative Behavior: Explains the largest singular proportion of variance (typically 22–26%), with primary item loadings ranging from .51 to .74. Prosocial items load cleanly with minimal cross-loadings onto the hostile behavior dimensions.
• Factor 2: Bullying: Accounts for approximately 12–15% of variance, with salient loadings (.62 to .81) across items representing verbal mocking, teasing, and threats.
• Factor 3: Fighting: Accounts for 8–11% of variance, characterized by high loadings (.58 to .79) on physical assault, brawling instigation, and physical retaliation. Item 5 (“I walked away from a fight”, reverse-coded) consistently demonstrates the lowest, though acceptable, loading (.42 to .50).
• Factor 4: Anger: Accounts for 6–9% of variance, capturing items reflecting emotional agitation, persistent irritability, and displaced wrath (loadings .54 to .77).

Confirmatory Factor Analysis (CFA) Fit Indices

Structural equation modeling and CFA evaluations confirm that a four-factor correlated model provides a superior fit to the data compared to unidimensional or two-factor (Aggression vs. Prosocial) models. Representative structural fit indices from adolescent samples include:

• Comparative Fit Index (CFI): .92 to .96, indicating strong comparative baseline fit.
• Tucker-Lewis Index (TLI): .91 to .95.
• Root Mean Square Error of Approximation (RMSEA): .042 to .058 (with 90% confidence intervals bounded below .065), demonstrating low residual error.
• Standardized Root Mean Square Residual (SRMR): .038 to .049.

Measurement invariance testing across biological sex and grade levels (grades 6, 7, and 8) has confirmed metric and scalar invariance, supporting the validity of cross-group comparisons and developmental tracking throughout early adolescence.

Instrument / Measurement Tool

• Instrument Name: Modified Aggression Scale (MAS); also referenced in literature as the Aggression and Prosocial Subscales of the Teen Conflict Survey.
• Developer / Originating Authors: Pamela Orpinas, Ph.D. (1993); modified and extended by Kris Bosworth, Ph.D., and Dorothy L. Espelage, Ph.D. (1995).
• Instrument Type: Standardized self-report behavioral questionnaire.
• Target Population: Early adolescents / middle school students (grades 6 through 8, approximately ages 11–15); adaptable to high school cohorts.
• Administration Format: Group-administered pencil-and-paper survey or secure digital assessment platform; typically integrated within broader school health or behavioral batteries.
• Estimated Administration Time: 8 to 12 minutes for complete administration.
• Item Count: 22 items total.
  • Fighting Subscale: 5 items
  • Bullying Subscale: 4 items
  • Anger Subscale: 5 items
  • Caring/Cooperative Behavior Subscale: 8 items
• Recall Period: Past 30 days immediately preceding administration.
• Response Scale and Numeric Coding: Items are presented with a 5-option frequency scale assigned the following point values:
  • No opportunity = 1
  • Never = 1
  • 1 or 2 times = 2
  • 3 or 4 times = 3
  • 5 or more times = 4

  (Note: “No opportunity” and “Never” share an equivalent assigned numeric value of 1 point.)

• Scoring Mechanics and Subscale Metrics:
  • Fighting Subscale (Range: 5 to 20 points): Item 5 (“I walked away from a fight”) is reverse-coded prior to computation (i.e., a raw score of 1 becomes 4, 2 becomes 3, 3 becomes 2, and 4 becomes 1). The 5 items are then summed. Higher scores indicate greater physical fighting, instigation, and failure to de-escalate.
  • Bullying Subscale (Range: 4 to 16 points): Calculated by directly summing responses across all 4 bullying items. Higher scores denote increased involvement in teasing, verbal harassment, and intimidation.
  • Anger Subscale (Range: 5 to 20 points): Calculated by directly summing responses across all 5 anger items. Higher scores reflect greater affective irritability, persistent emotional arousal, and displaced hostility.
  • Caring/Cooperative Behavior Subscale (Range: 8 to 32 points): Calculated by summing responses across all 8 prosocial items. Higher scores indicate more frequent enactment of cooperative problem solving, emotional sharing, and bystander defense.
  • Total Aggression Index (Optional): Researchers frequently calculate an overall Aggression index by summing the Fighting, Bullying, and Anger subscales (total range: 14 to 56 points).

Permissions & Fee and Test Year

The foundational Aggression Scale was established in 1993 by Dr. Pamela Orpinas, with the modified multi-scale version published in 1995 by Dr. Kris Bosworth and Dr. Dorothy L. Espelage. The Modified Aggression Scale was subsequently cataloged and published by the United States Federal Government in the Centers for Disease Control and Prevention’s compendium: Measuring Violence-Related Attitudes, Behaviors, and Influences Among Youths: A Compendium of Assessment Tools (Dahlberg et al., 2005, pp. 169–170).

Because the instrument was developed and disseminated under federal research and public health initiatives through the CDC and the Educational Resources Information Center (ERIC ED486261), it resides in the public domain. There are no licensing fees, purchase royalties, or formal administrative costs required for non-commercial educational, clinical, or academic research use. Researchers and institutional practitioners are encouraged to maintain standard academic attribution by citing the primary validation manuscripts and the CDC compendium.

References

• Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology, 53(1), 27–51. https://doi.org/10.1146/annurev.psych.53.100901.135231
• Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall.
• Bosworth, K., & Espelage, D. L. (1995). Teen Conflict Survey. Center for Adolescent Studies, Indiana University.
• Bosworth, K., Espelage, D. L., & Simon, T. R. (1999). Factors associated with bullying behavior in middle school students. The Journal of Early Adolescence, 19(3), 341–362. https://doi.org/10.1177/0272431699019003003
• Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social information-processing mechanisms in children’s social adjustment. Psychological Bulletin, 115(1), 74–101. https://doi.org/10.1037/0033-2909.115.1.74
• Dahlberg, L. L., Toal, S. B., Swahn, M., & Behrens, C. B. (2005). Measuring violence-related attitudes, behaviors, and influences among youths: A compendium of assessment tools (2nd ed.). Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. https://www.cdc.gov/violenceprevention/pdf/YV_Compendium.pdf
• Espelage, D. L., Bosworth, K., & Simon, T. R. (2000). Examining the social context of bullying behaviors in early adolescence. Journal of Counseling & Development, 78(3), 326–333. https://doi.org/10.1002/j.1556-6676.2000.tb01914.x
• Orpinas, P. (1993). Skills training and social influences for violence prevention in middle schools: A curriculum evaluation [Doctoral dissertation, University of Texas Health Science Center at Houston, School of Public Health]. ProQuest Dissertations Publishing.
• Orpinas, P., & Frankowski, R. (2001). The Aggression Scale: A self-report measure of aggressive behavior for young adolescents. The Journal of Early Adolescence, 21(1), 50–67. https://doi.org/10.1177/0272431601021001003

13. Items of the Scale (Questionnaire)