Buss-Perry Aggression Questionnaire (AGQ)

Abstract

The Buss-Perry Aggression Questionnaire (AGQ), also widely referenced in psychometric literature as the Aggression Questionnaire (AQ), is an internationally recognized self-report measurement instrument engineered to evaluate quantitative dimensions of human aggressive behavior. Originally constructed by Arnold H. Buss and Mark P. Perry in 1992 as a psychometric refinement of the legacy Buss-Durkee Hostility Inventory (BDHI) developed in 1957, the AGQ addresses major psychometric limitations of earlier dichotomous true-false instruments. The instrument consists of 29 self-report items distributed across a robust four-factor operational architecture: Physical Aggression (9 items), Verbal Aggression (5 items), Anger (7 items), and Hostility (8 items). Respondents rate each statement on a 5-point Likert scale ranging from 1 (“Extremely uncharacteristic”) to 5 (“Extremely characteristic”). Psychometric evaluations consistently confirm satisfactory internal consistency, with Cronbach’s alpha coefficients typically ranging from 0.72 to 0.85 across subscales and reaching approximately 0.89 for the overall composite score, accompanied by substantial nine-week test-retest stability coefficients (0.72 to 0.80). Confirmatory factor analytic investigations support its hierarchical structure, affirming that aggression operates as a multidimensional construct synthesizing instrumental/motor actions, affective/physiological activation, and cognitive/attitudinal biases. Validated across diverse psychiatric, forensic, cross-cultural, and non-clinical samples, the AGQ remains an indispensable gold standard in clinical psychopathology, behavioral genetics, social psychology, and risk assessment research.

Keywords

Buss-Perry Aggression Questionnaire, AGQ, Aggression Questionnaire, Physical Aggression, Verbal Aggression, Anger, Hostility, Psychometrics, Factor Structure, Behavioral Assessment, Hostile Attribution Bias, Self-Report Scales

Authors

The Buss-Perry Aggression Questionnaire was developed by:

• Arnold H. Buss, Ph.D. — Late Professor Emeritus of Psychology at the University of Texas at Austin. Renowned pioneer in behavioral genetics, temperament research, and the operational assessment of personality traits and aggressive dispositions.
• Mark P. Perry, Ph.D. — Department of Psychology, University of Texas at Austin. Co-developer and quantitative psychometrician on the 1992 questionnaire reconstruction.

Purpose

The overarching purpose of the Buss-Perry Aggression Questionnaire (AGQ) is to provide clinicians, behavioral scientists, and psychometricians with an empirically validated, fine-grained diagnostic and investigative scale that dissociates overt behavioral manifestations of violence from internal emotional reactivity and cognitive hostility. Prior to the development of the AGQ, clinical and social psychological assessments were predominantly reliant on the Buss-Durkee Hostility Inventory (BDHI; Buss & Durkee, 1957). Although the BDHI achieved widespread clinical adoption, subsequent psychometric evaluations demonstrated substantial shortcomings, including a binary (True/False) response format that suppressed trait variance, confounded factor structures, and mixed affective, attitudinal, and physical manifestations within unstandardized indices.

The AGQ rectifies these structural deficiencies by establishing four distinct yet correlated dimensions of aggression, enabling precise clinical profiling and differential research paradigms:

• Clinical and Psychiatric Assessment: In clinical settings, the AGQ facilitates the evaluation of impulse control deficits, intermittent explosive disorder, borderline personality pathology, antisocial conduct, and emotion dysregulation. Because anger and hostility can emerge independently of motoric violence, clinicians utilize the AGQ to isolate emotional triggers (Anger) and paranoid or cynical assumptions (Hostility) from motoric outbursts (Physical and Verbal Aggression).
• Forensic and Correctional Diagnostics: The AGQ serves as an evidence-based component of comprehensive violence risk appraisals in forensic institutions, correctional facilities, and probationary monitoring. Quantifying overt violent tendencies versus verbal provocation helps legal and correctional authorities design individualized rehabilitative programs.
• Experimental and Social Psychology: The instrument allows researchers to dissect behavioral aggression in response to laboratory stress inductions, frustration protocols, and cognitive bias interventions, ensuring that experimental effects are examined at the specific level of motor action versus latent cognitive appraisals.
• Treatment Outcome Monitoring: Owing to its continuous Likert scale, the AGQ is sensitive to subtle behavioral and emotional shifts occurring across psychotherapy (such as cognitive-behavioral anger management) and psychopharmacological interventions (including mood stabilizers, atypical antipsychotics, or selective serotonin reuptake inhibitors).

Psychological Construct

Human aggression is conceptualized within modern psychometrics not as a monolithic drive, but as a complex multidimensional construct spanning expressive motor behaviors, physiological affective arousal, and cognitive-attitudinal belief systems. Buss and Perry (1992) formalized this tripartite division into four specific operational subscales:

1. Physical Aggression (Motor/Instrumental Dimension)

Physical Aggression constitutes the overt, motoric execution of destructive or violent behavior toward other individuals or objects. It represents the behavioral endpoint of aggressive drive wherein an individual inflicts, attempts to inflict, or threatens physical harm. Items measuring this construct focus on the proclivity to engage in physical fights, use weapons or force to protect personal interests, and break objects under extreme emotional distress. Prototypical expressions include item 2 (“If I have to resort to violence to protect my rights, I will”) and item 8 (“Once in a while, I can’t control the urge to strike another person”). This dimension directly reflects behavioral disinhibition and externalizing behavioral pathology.

2. Verbal Aggression (Motor/Instrumental Dimension)

Verbal Aggression encompasses the communicative, vocalized expression of antipathy, confrontation, and hostility. Like physical aggression, it is motoric, yet its execution is symbolic and linguistic rather than physically destructive. This dimension assesses an individual’s tendency to engage in heated arguments, deliver insulting or threatening verbalizations, and openly proclaim hostile dissent. Expressive examples include item 4 (“I tell my friends openly when I disagree with them”) and item 11 (“I have threatened people I know”). Verbal aggression often functions as an escalatory precursor to physical altercations or as a non-contact manifestation of dominance and intimidation.

3. Anger (Affective Dimension)

Anger assesses the physiological, autonomic, and subjective emotional state that bridges cognitive perceptions of injustice with overt behavioral aggression. This subscale reflects autonomic nervous system arousal, emotional lability, frustration intolerance, and difficulties regarding temper regulation. Unlike physical and verbal actions, anger is an internal affective response characterized by feelings of fury, hotheadedness, and explosive urgency. Illustrative items include item 1 (“Some of my friends think I am a hothead”) and item 23 (“I sometimes feel like a powder keg ready to explode”). Anger provides the motivational and physiological fuel that can catalyze both verbal and physical acts.

4. Hostility (Cognitive Dimension)

Hostility captures the cognitive, attitudinal, and evaluative orientation of the individual toward their social environment. It is defined by cynicism, generalized suspicion, ill-will, social paranoia, and perceived injustice. Unlike immediate affective anger or acute behavioral violence, hostility represents a chronic, schematic worldview wherein the individual believes that others are fundamentally adversarial, exploitative, or untrustworthy. Prototypical items include item 3 (“When people are especially nice to me, I wonder what they want”) and item 17 (“At times I feel I have gotten a raw deal out of life”). Hostility constitutes a durable cognitive schema that biases social information processing toward malevolent attributions.

Theoretical Framework

The construction and validation of the Buss-Perry Aggression Questionnaire are theoretically anchored in modern paradigms of social cognition, behavioral genetics, and affective neuroscience. Specifically, the AGQ is structured upon three central conceptual pillars:

The Tripartite Model of Emotion and Behavior

Buss and Perry grounded their framework in the classical tripartite paradigm of psychological response systems: Cognitive, Affective, and Motoric. Motor components are partitioned into Physical Aggression and Verbal Aggression, both representing action-oriented instrumental responses. The affective component is encapsulated within the Anger subscale, accounting for autonomic reactivity and subjective emotional experience. The cognitive component is operationalized as Hostility, embodying perceptual biases and inferential schemas. By decomposing aggression into these interconnected domains, the theoretical model elucidates how an individual may harbor intense cognitive hostility without ever crossing the threshold into physical violence, or conversely, act violently out of uninhibited impulsivity in the absence of sustained hostile attitudes.

Berkowitz’s Cognitive Neoassociationistic Model

The AGQ strongly reflects Leonard Berkowitz’s cognitive neoassociationistic perspective on aggressive behavior. Berkowitz posited that aversive events (e.g., frustration, physical pain, social rejection) generate unconditioned negative affect. This negative affect simultaneously activates an associative network of primitive physiological reactions, motor impulses, and cognitive memories linked to fight-or-flight tendencies. In the context of the AGQ, chronic hostility primes cognitive associative nodes, rendering an individual hyper-responsive to benign environmental cues, which in turn sparks acute anger and subsequently unleashes physical or verbal violence.

The General Aggression Model (GAM)

Subsequent psychometric and experimental research integrates the AGQ within the General Aggression Model (GAM) developed by Craig A. Anderson and Brad J. Bushman. The GAM incorporates biological, personality, and environmental inputs to explain the internal states (affect, cognition, arousal) that guide appraisal processes and behavioral decision outcomes. Within this paradigm, the AGQ functions as a stable measurement of individual personality inputs: baseline cognitive hostility dictates the initial primary appraisal of intent, trait anger modulates physiological arousal and affective intensity, and physical/verbal aggressive habits govern the eventual impulsive or thoughtful behavioral action choice.

Validity

The psychometric validity of the Buss-Perry Aggression Questionnaire has been extensively scrutinized across clinical, student, correctional, and community cohorts internationally.

Construct and Structural Validity

Construct validity was initially established by Buss and Perry (1992) through rigorous exploratory factor analyses (EFA) on a large normative cohort of university undergraduates (N = 1,253). The four-factor solution demonstrated exceptional construct distinctiveness while confirming inter-factor relationships: Anger was found to be strongly correlated with both Physical Aggression (r = 0.43) and Hostility (r = 0.53), illustrating its role as the affective bridge connecting cognitive schemas to physical actions.

Convergent and Discriminant Validity

Convergent validity is documented via strong, predictable associations with theoretically related personality constructs and external behavioral inventories:

• Personality Dimensions: Within the Five-Factor Model (FFM) of personality, the AGQ subscales correlate negatively with Agreeableness (ranging from −0.45 to −0.60, with the strongest negative relationships observed for Anger and Hostility) and positively with Neuroticism (primarily Hostility and Anger, exhibiting correlations between 0.35 and 0.52).
• Impulsivity and Sensation Seeking: Physical Aggression and Anger demonstrate robust positive correlations (r = 0.38 to 0.48) with Barratt Impulsiveness Scale (BIS-11) scores, whereas Hostility aligns primarily with cognitive reflection deficits and trait anxiety inventories.
• Peer and Collateral Nominations: Buss and Perry (1992) paired self-reports with peer ratings, demonstrating that peer-observed aggressive behavior correlated significantly with self-reported Physical Aggression (r = 0.45) and Verbal Aggression (r = 0.36), confirming that the instrument minimizes severe social desirability bias.

Criterion and Predictive Validity

The scale effectively discriminates between violent and non-violent populations. Forensic evaluations demonstrate that incarcerated violent offenders score significantly higher on the Physical Aggression subscale compared to non-violent offenders and non-offender controls. Laboratory paradigms testing competitive provocation (such as the Taylor Aggression Paradigm) confirm that individuals with elevated baseline AGQ scores deliver significantly higher electro-tactile or acoustic punishment to ostensible opponents.

Reliability

The AGQ displays robust internal consistency and temporal stability across diverse cultural, language, and clinical contexts:

Internal Consistency

In the foundational validation study by Buss and Perry (1992), internal consistency was estimated via Cronbach’s alpha coefficients across an initial sample of 1,253 undergraduate participants:

• Physical Aggression: α = 0.85 (9 items)
• Verbal Aggression: α = 0.72 (5 items)
• Anger: α = 0.83 (7 items)
• Hostility: α = 0.77 (8 items)
• Total Aggression Scale: α = 0.89 (29 items)

Subsequent psychometric cross-validations in European, Asian, and Latin American translations have yielded comparable coefficients, consistently placing subscale alpha values between 0.70 and 0.86, and total score reliability between 0.88 and 0.93.

Test-Retest Temporal Stability

Buss and Perry evaluated the longitudinal temporal stability of the scale across a 9-week interval in a sub-sample of 372 participants. The resulting Pearson correlation coefficients confirmed durable trait stability:

• Physical Aggression: r = 0.80
• Verbal Aggression: r = 0.76
• Anger: r = 0.72
• Hostility: r = 0.72
• Total Score: r = 0.80

These findings substantiate that the AGQ measures stable personality traits rather than temporary, fluctuating situational states.

Factor Analysis

The internal structural validity of the AGQ has been an area of active psychometric inquiry, spanning Exploratory Factor Analysis (EFA), Confirmatory Factor Analysis (CFA), and Item Response Theory (IRT).

Original Exploratory Factor Analysis

Buss and Perry (1992) conducted a principal components analysis followed by oblique rotation (promax) on an initial item pool of 52 questions derived from the BDHI. Items with low primary loadings (< 0.35) or substantial cross-loadings were iteratively eliminated, yielding a parsimonious 29-item structure. The four retained factors accounted for the underlying common variance and exhibited modest-to-moderate inter-correlations, demonstrating distinct yet conceptually linked facets.

Confirmatory Factor Analytic Studies

Subsequent structural evaluations using Confirmatory Factor Analysis (e.g., Harris, 1997; Archer et al., 1995; Bryant & Smith, 2001) scrutinized the four-factor oblique model against alternative structural configurations, including a single-factor unidimensional model and a two-factor model (internal/emotional vs. external/motoric):

• Unidimensional Model: Consistently rejected across studies, producing inadequate goodness-of-fit indices (e.g., Comparative Fit Index [CFI] < 0.75; Root Mean Square Error of Approximation [RMSEA] > 0.10).
• Four-Factor Oblique Model: Demonstrates superior fit across independent samples (CFI typically 0.90 to 0.93; RMSEA ~ 0.05 to 0.06), verifying that human aggression is best modeled as a correlated four-dimensional construct.
• Higher-Order (Hierarchical) Model: CFA investigations frequently affirm a second-order model in which a single overarching latent trait (“General Aggression”) drives the four first-order latent factors. This empirical finding legitimizes the reporting of both specific subscale scores and a cumulative Total Aggression index.

Short-Form Derivations and Model Modifications

Methodological researchers, most notably Bryant and Smith (2001), observed that several original items displayed minor structural noise or modest cross-loadings. This led to the validation of a 12-item abbreviated version (AQ-12) containing three items per subscale, which exhibits excellent CFA model fit parameters across multiple cultural environments. Nonetheless, the full 29-item AGQ remains the standard diagnostic and research reference instrument.

Instrument / Measurement Tool

The operational administration, item composition, and scoring metrics of the Buss-Perry Aggression Questionnaire are structured as follows:

• Instrument Type: Standardized self-report psychometric rating questionnaire.
• Administration Format: Individual or group administration via paper-and-pencil or computerized/digital platforms.
• Target Population: Adolescents (aged 12 and above) and adults across clinical, forensic, and community contexts.
• Estimated Completion Time: Approximately 7 to 12 minutes.
• Item Count: 29 declarative statements.
• Response Scale: 5-point Likert scale formatted with the following response designations:
  • 1 = Extremely uncharacteristic
  • 2 = Somewhat uncharacteristic
  • 3 = Neither uncharacteristic nor characteristic
  • 4 = Somewhat characteristic
  • 5 = Extremely characteristic
• Reverse-Scored Items: Two items are framed in the anti-aggressive direction and must be reverse-coded (1 → 5, 2 → 4, 3 → 3, 4 → 2, 5 → 1) prior to calculating scale scores:
  • Item 9: “I am an even-tempered person.” (Reverse-scored; belongs to Anger subscale)
  • Item 16: “I can think of no good reason for ever hitting a person.” (Reverse-scored; belongs to Physical Aggression subscale)
• Subscale Item Distributions:
  • Physical Aggression (9 items): Items 2, 5, 8, 11, 13, 16 (reversed), 22, 25, 29. Possible score range: 9 to 45.
  • Verbal Aggression (5 items): Items 4, 6, 14, 21, 27. Possible score range: 5 to 25.
  • Anger (7 items): Items 1, 9 (reversed), 12, 18, 19, 23, 28. Possible score range: 7 to 35.
  • Hostility (8 items): Items 3, 7, 10, 15, 17, 20, 24, 26. Possible score range: 8 to 40.
• Total Score Calculation: Sum of all 29 items (after reverse-scoring items 9 and 16). Theoretical range: 29 to 145, with higher cumulative scores indicating greater general trait aggression.

Permissions & Fee and Test Year

• Year of Initial Publication: 1992 (following its precursor, the Buss-Durkee Hostility Inventory, published in 1957).
• Original Publication Venue: Journal of Personality and Social Psychology, published by the American Psychological Association (APA).
• Permissions and Licensing: The 29-item questionnaire was published in full within the original 1992 peer-reviewed academic journal article by Buss and Perry. Under standard academic fair use guidelines, the scale items may be utilized without fee for non-commercial academic research, student theses, and institutional scientific inquiry, provided appropriate academic attribution and citation are rendered. For commercial applications, diagnostic software integration, or proprietary clinical test distribution, permissions must be coordinated through the American Psychological Association (APA) rights and permissions department.

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
• Archer, J., Kilpatrick, G., & Bramwell, R. (1995). Comparison of two aggression inventories. Aggressive Behavior, 21(5), 371-380. https://doi.org/10.1037/0003-066X.45.4.494
• Bryant, F. B., & Smith, B. D. (2001). Refining the architecture of aggression: A measurement model for the Buss–Perry Aggression Questionnaire. Journal of Research in Personality, 35(2), 138-167. https://doi.org/10.1006/jrpe.2000.2302
• Buss, A. H., & Durkee, A. (1957). An inventory for assessing different kinds of hostility. Journal of Consulting Psychology, 21(4), 343-349. https://doi.org/10.1037/h0046900
• Buss, A. H., & Perry, M. (1992). The Aggression Questionnaire. Journal of Personality and Social Psychology, 63(3), 452-459. https://doi.org/10.1037/0022-3514.63.3.452
• Felsten, G., & Hill, V. (1999). Aggression Questionnaire hostility scale predicts anger in response to mistreatment. Behaviour Research and Therapy, 37(1), 87-97. https://doi.org/10.1016/S0005-7967(98)00115-4
• Harris, J. A. (1997). A further evaluation of the Aggression Questionnaire: Issues of validity and construct co-occurrence. Behaviour Research and Therapy, 35(9), 877-883. https://doi.org/10.1016/S0005-7967(97)00041-5

13. Items of the Scale (Questionnaire)