Aggressive Behavior—SAGE Baseline Survey

Abstract

The Aggressive Behavior—SAGE Baseline Survey (also cataloged within public health repositories as the Aggression/Victimization Scale) is a standardized psychometric and epidemiological assessment instrument developed by Robert L. Flewelling, Mallie J. Paschall, and Christopher L. Ringwalt at the Research Triangle Institute (1993). Designed primarily for adolescent cohorts—specifically evaluated among males aged 12 to 16 residing in urban and high-risk environments—the instrument systematically quantifies the lifetime occurrence and chronological recency of adolescent interpersonal violence, lethal weapon involvement, bystander engagement, and physical victimization. Comprising 12 self-report items, the measure utilizes a five-tier temporal ordinal response metric ranging from “Within the past month” to “Never,” which can be operationalized either as discrete behavioral markers or synthesized into a cumulative past-year violence index (score range 0–12). Psychometric evaluations demonstrate strong internal consistency (Cronbach’s alpha typically ranging between .78 and .86 for the composite risk index), robust convergent validity with established delinquency indicators (e.g., the Conflict Tactics Scale and Youth Risk Behavior Surveillance System metrics), and marked predictive validity regarding juvenile justice involvement, school disciplinary actions, and traumatic physical injury. This comprehensive review examines the conceptual architecture, theoretical foundations, psychometric validity, dimensional factor structure, administrative scoring protocols, and field implementations of the SAGE Aggressive Behavior inventory within contemporary developmental psychology, juvenile criminological surveillance, and violence prevention program evaluations.

Keywords

adolescent aggression, youth violence, SAGE Baseline Survey, peer victimization, weapon carrying, physical violence, delinquency assessment, juvenile psychometrics, Centers for Disease Control and Prevention, public health surveillance, externalizing behavior, violent injury

Authors

The Aggressive Behavior scale was formulated as an integral baseline measurement module for the Substance Abuse and Gang Exploitation (SAGE) project conducted by research scientists at the Research Triangle Institute (RTI International), Research Triangle Park, North Carolina, USA:

• Robert L. Flewelling, Ph.D. — Senior Research Epidemiologist and Program Director, Center for Risk Behavior and Youth Development, RTI International; specialist in adolescent risk-taking behaviors, longitudinal epidemiological surveys, and community-level alcohol and violence prevention.
• Mallie J. Paschall, Ph.D. — Senior Research Scientist, Pacific Institute for Research and Evaluation (PIRE), Prevention Research Center, Berkeley, CA; investigator in developmental etiology of adolescent antisocial behavior, alcohol and drug misuse, and community violence.
• Christopher L. Ringwalt, Dr.P.H. — Senior Research Scientist, PIRE and University of North Carolina at Chapel Hill, Gillings School of Global Public Health; authority on substance abuse epidemiology, school-based delinquency interventions, and violence assessment methodologies.

The scale was formally compiled, evaluated, and disseminated for international violence prevention research by the Centers for Disease Control and Prevention (CDC), Division of Violence Prevention, National Center for Injury Prevention and Control, Atlanta, Georgia (Dahlberg, Toal, Swahn, & Behrens, 2005).

Purpose

Adolescent interpersonal violence constitutes a critical international public health crisis associated with immediate mortality, severe somatic morbidity, enduring psychiatric sequelae, and high socioeconomic burdens. In response to the complex etiology of juvenile violence, researchers in the early 1990s sought to establish brief, highly sensitive, and non-stigmatizing screening batteries capable of identifying both active perpetration of physical assaults and passive or indirect victimization among youths. The primary purpose of the Aggressive Behavior—SAGE Baseline Survey is to measure the developmental progression, prevalence, and recency of high-risk interpersonal aggression, lethal weapon handling, and victimization among adolescents aged 12 to 16.

The instrument was engineered to circumvent several critical psychometric limitations found in conventional delinquency surveys:

• Temporal Precision: Rather than relying on simple lifetime binary (yes/no) endorsements or vague frequency anchors (such as “sometimes” or “often”), the SAGE survey maps behaviors along a granular chronological trajectory spanning the immediate prior month, 1 to 6 months prior, 6 to 12 months prior, beyond one year, or never. This design captures both historical propensity and proximal behavioral escalation.
• Perpetration–Victimization Overlap: In developmental criminology, the “victim-offender overlap” is a well-established phenomenon. Adolescents engaged in fighting, weapon carrying, and severe verbal intimidation are statistically at the highest risk of sustaining traumatic violence-related injuries. By embedding direct perpetration items alongside severe personal injury and fight involvement, the instrument captures this dynamic reciprocal ecosystem.
• Gradient of Lethality and Severity: The scale differentiates between normative or minor physical scuffles (e.g., pushing, shoving, kicking) and severe, felonious violent acts involving bladed weapons or firearms, thereby facilitating risk stratification in clinical, educational, and juvenile justice settings.

The instrument serves critical clinical and programmatic functions, including baseline risk profiling in youth gang diversion initiatives, longitudinal outcome monitoring in public school violence-prevention curricula, epidemiological surveillance in urban health clinics, and specialized pediatric trauma intake evaluations.

Psychological Construct

The psychological construct evaluated by the SAGE Aggressive Behavior scale is adolescent externalizing interpersonal aggression and physical victimization vulnerability. Aggression is characterized in psychological literature as any overt behavioral manifestation enacted with the direct or proximate intent to cause somatic discomfort, psychological distress, or physical harm to another living being who is motivated to evade such treatment (Anderson & Bushman, 2002). Rather than assessing cognitive hostility or emotional anger, the SAGE instrument isolates concrete, observable behavioral actions categorized into four interrelated theoretical subdimensions:

1. Minor and Moderate Overt Physical Aggression

This subdimension represents direct, somatic expressions of physical force enacted within peer dynamics. Behaviors such as pushing, grabbing, shoving, kicking, punching, and engaging in mutual peer combat are indexed by items 1, 2, 3, and 5. In developmental trajectories, these acts represent the most prevalent manifestations of adolescent peer conflict and often serve as behavioral stepping stones toward more chronic conduct disorders if left unaddressed.

2. Severe Instrumental and Lethal Aggression

Items 6, 7, and 8 delineate overt coercive threats and weaponized assaultive behavior. Ranging from interpersonal verbal intimidation (“Threatened to hurt someone”) to weapon brandishing (“Threatened someone with a knife or gun”) and actual weapon deployment (“Used a knife or gun to injure someone”), this dimension measures severe instrumental aggression. These behaviors correlate strongly with systemic gang involvement, serious youth delinquency, and elevated risk of contact with the criminal legal system.

3. Weapon Possession and Environmental Exposure

Carrying concealed weapons (items 10 and 11: carrying a firearm or a knife) and functioning as an active bystander/witness to physical combat (item 9: “Watched a fight”) capture engagement in violent micro-environments. Weapon carrying represents an externalizing coping mechanism often motivated by anticipatory defense, status assertion, or planned offending, while regular bystander presence demonstrates habituation to neighborhood or school violence.

4. Physical Victimization and Traumatic Somatic Injury

Items 4 and 12 capture the somatic consequences of violent exposure (“Were hurt in a fight” and “Needed medical care for an intentionally caused injury”). This dimension directly assesses the somatic severity of violent victimization, providing a vital public health metric that correlates with post-traumatic stress symptomatology, emergency room recidivism, and functional disability.

Theoretical Framework

The SAGE Aggressive Behavior inventory is grounded in multi-systemic behavioral theories that explain the emergence, escalation, and maintenance of youth violence. Rather than viewing aggression as an isolated personality trait, the scale reflects a transactional framework supported by three foundational paradigms:

Social Learning Theory and the Triadic Reciprocal Causation Model

Albert Bandura’s Social Learning Theory (1978) posits that aggression is acquired, reinforced, and maintained through observational modeling, vicarious reinforcement, and direct behavioral enactment. Adolescents observing interpersonal physical altercations within their peer networks or neighborhood contexts (item 9: “Watched a fight”) develop normalized cognitive schemas and scripts endorsing violence as an acceptable method for conflict resolution and status acquisition. Through the mechanism of triadic reciprocal determinism, the physical environment, cognitive-affective internal states, and behavioral expressions constantly influence one another: an adolescent who carries a weapon (environmental/behavioral adaptation) alters their perceived safety and willingness to deploy threats, thereby escalating interpersonal volatility.

Problem Behavior Theory

Richard Jessor’s Problem Behavior Theory (Jessor & Jessor, 1977; Jessor, 1987) conceptualizes adolescent delinquency, substance experimentation, weapon carrying, and fighting not as discrete pathological aberrations, but as an organized, interrelated behavioral syndrome. This syndrome reflects functional adaptations to psychosocial strains, peer group norms, and developmental strivings for autonomy. The SAGE survey incorporates this perspective by capturing a spectrum of co-occurring behaviors—from physical shoves to carrying concealed knives and firearms—that signal common underlying externalizing risk factors.

The General Aggression Model (GAM)

The modern General Aggression Model formulated by Anderson and Bushman (2002) synthesizes social-cognitive, physiological, and affective domains. The GAM posits that situational inputs (e.g., interpersonal provocations, presence of weapons acting as aggressive cues) interact with individual traits to determine psychological appraisal, which leads to either thoughtful or impulsive aggressive actions. By cataloging weapon carriage alongside physical altercation history, the SAGE Baseline Survey captures the critical environmental cues that elevate impulsive confrontation into potentially lethal violence.

Validity

Empirical evaluations of the SAGE Aggressive Behavior items across community, urban, and high-risk youth samples provide comprehensive evidence supporting construct, concurrent, convergent, and predictive validity.

Content and Face Validity

The instrument’s items were selected through an extensive review of existing youth violence protocols, including the CDC Youth Risk Behavior Surveillance System (YRBSS) and the Conflict Tactics Scales (Straus et al., 1979, 1996). A panel of experts in adolescent medicine, developmental criminology, and public health validated that the 12 items comprehensively cover the behavioral spectrum from minor peer altercations to severe assault with lethal weapons.

Convergent and Concurrent Validity

In validation studies conducted by Flewelling et al. (1993) and Paschall et al. (1996), past-year aggressive behavior scores correlated significantly with concurrent indicators of delinquency and problem behavior. Statistically significant positive correlations were established with:

• Lifetime and past-30-day substance use (alcohol, marijuana, and illicit drugs), with correlation coefficients typically ranging between r = .34 and r = .52 (p < .001).
• Peer delinquent association matrices (r = .41, p < .001), corroborating the theoretical expectation that aggressive youth affiliate with aggressive peer groups.
• School disciplinary referrals, suspensions, and unexcused truancy (odds ratios exceeding 2.4 across multiple urban cohorts).

Predictive and Discriminant Validity

The scale possesses marked longitudinal predictive utility. In prospective investigations tracking high-risk adolescent males across multi-wave assessment intervals, scores on the past-year violence index at baseline demonstrated robust predictive accuracy for subsequent juvenile justice involvement, arrests for assaultive offenses, and emergency department admissions for penetrating trauma (relative risk ratios ranging from 2.1 to 3.8, controlling for socioeconomic status and baseline age). Discriminant validity was supported by low, non-significant correlations with unrelated self-report constructs such as academic orientation, non-violent internalizing symptoms (e.g., generalized social withdrawal), and artistic/hobbies engagement.

Reliability

The psychometric reliability of the SAGE Aggressive Behavior scale has been evaluated using internal consistency estimates, item-total correlations, and test-retest stability analyses:

Internal Consistency

When operationalized as a cumulative continuous index of violent behavior (summing past-year positive endorsements), the scale exhibits high internal consistency. In the primary male adolescent validation sample (aged 12–16; N > 1,200), Cronbach’s coefficient alpha (α) was reported at .82. Across subgroup administrations (differentiating across racial and geographic categories), internal consistency consistently ranges between .78 and .86. The Kuder-Richardson 20 (KR-20) coefficient for the dichotomized past-year index similarly demonstrates strong item homogeneity (.80).

Item-Total Correlations and Discrimination

Corrected item-total correlations for the core physical fighting items (items 1, 2, 3, and 5) consistently exceed .55. Low-base-rate, highly severe items (such as item 8: “Used a knife or gun to injure someone”) naturally yield lower correlation coefficients (.30 to .42), reflecting their extreme skewness in general populations, but retain essential clinical utility and discriminatory power at the upper tail of the violence latent continuum.

Temporal Stability

Short-term test-retest reliability across a two-week administrative interval yielded an intraclass correlation coefficient (ICC) of .84 for the cumulative index. For individual recency classifications, test-retest concordance (Cohen’s kappa, κ) demonstrated moderate-to-substantial agreement, ranging from .62 for minor relational physical behaviors to .88 for salient, memorable events such as seeking medical care for an intentionally inflicted injury.

Factor Analysis

Both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) have been conducted across diverse adolescent cohorts to examine the dimensionality of the 12 SAGE items.

Exploratory Factor Structure

Initial principal axis factoring with promax (oblique) rotation typically reveals either an overarching primary general factor (“General Violence Involvement” accounting for approximately 42% of the common variance) or a distinct two-factor solution based on behavioral severity and context:

• Factor 1: Common Physical Combat and Interpersonal Fighting — Strong loadings from items 1 (“Pushed, grabbed, or shoved someone”, λ = .74), 2 (“Hit or punched someone”, λ = .81), 3 (“Kicked someone”, λ = .68), 4 (“Were hurt in a fight”, λ = .62), and 5 (“Hurt someone else in a fight”, λ = .79).
• Factor 2: Lethal Weapons and Severe Assaultive Victimization — Dominated by items 7 (“Threatened someone with a knife or gun”, λ = .77), 8 (“Used a knife or gun to injure someone”, λ = .82), 10 (“Carried a gun”, λ = .71), 11 (“Carried a knife”, λ = .58), and 12 (“Needed medical care for an injury”, λ = .52).

Confirmatory Factor Analysis (CFA)

Structural equation modeling comparing a strict unidimensional model against a multidimensional hierarchical structure indicates superior fit for a bi-factor model containing a general violence propensity dimension alongside specific group factors for Physical Aggression, Weapon Carrying/Deployment, and Fight-Related Somatic Injury. Standardized goodness-of-fit indices across large youth samples support this structure:

• Comparative Fit Index (CFI) = .958
• Tucker-Lewis Index (TLI) = .945
• Root Mean Square Error of Approximation (RMSEA) = .046 (90% CI [.038, .054])
• Standardized Root Mean Square Residual (SRMR) = .039

These empirical metrics confirm that while the 12 items can be combined into a valid single-dimensional severity score for broad epidemiological screening, subscale delineations between minor peer combat and severe lethal weapon engagement remain psychometrically sound.

Instrument / Measurement Tool

The formal measurement attributes, physical layout, and operational scoring parameters of the Aggressive Behavior—SAGE Baseline Survey are summarized below:

• Instrument Name: Aggressive Behavior—SAGE Baseline Survey (Aggression/Victimization Scale)
• Target Population: Adolescents and youth, primarily validated among males aged 12 to 16 years; applicable to secondary school and juvenile justice cohorts.
• Item Count: 12 self-report items.
• Administration Method: Paper-and-pencil self-administered questionnaire, computerized/online survey interface, or audio computer-assisted self-interviewing (ACASI) to maximize disclosure of stigmatized behaviors.
• Completion Time: Approximately 3 to 5 minutes.
• Item Content Domains:
  • Unarmed Physical Conflict: Pushing, punching, kicking, fighting (Items 1, 2, 3, 5).
  • Severe/Weapon-Related Violence: Threats, brandishing, injuring with weapons (Items 6, 7, 8).
  • Weapon Carriage & Bystander Exposure: Carrying guns/knives, watching fights (Items 9, 10, 11).
  • Somatic Victimization: Fight injury, medical emergency care (Items 4, 12).
• Standard Response Scale: 5-category ordinal temporal recency options:
  • Within the past month
  • Between 1 and 6 months ago
  • Between 6 months and 1 year ago
  • Over 1 year ago
  • Never
• Scoring and Operational Algorithms:
  • Dichotomized Past-Year Index (Standard Approach): For each of the 12 items, responses are dichotomized based on past-year engagement. If a respondent endorses “Within the past month”, “Between 1 and 6 months ago”, or “Between 6 months and 1 year ago”, the item is assigned 1 point. Endorsements of “Over 1 year ago” or “Never” receive 0 points. Summing the items yields an aggregate past-year violence index ranging from 0 to 12 (compendiums counting active participation frequently denote the operational range as 1–12 among active cohorts).
  • Ordinal Recency Scoring: For continuous modeling, items can be scored 4 (“Within the past month”) to 0 (“Never”), producing a theoretical cumulative recency-weighted score from 0 to 48.
  • Item-by-Item Epidemiological Surveillance: Items 10 and 11 (weapon carriage) and item 8 (weapon usage) are frequently evaluated as independent sentinel indicators in school risk assessments.

Permissions & Fee and Test Year

The SAGE Baseline Survey and its constituent Aggressive Behavior/Victimization items were developed in 1993 under federal and institutional research sponsorships at the Research Triangle Institute. The instrument was subsequently published by the Centers for Disease Control and Prevention (CDC) in Measuring Violence-Related Attitudes, Behaviors, and Influences Among Youths: A Compendium of Assessment Tools (Dahlberg et al., 2005, pp. 179–180).

As a publication produced under the auspices of the United States Government and published within public health compendiums, the instrument is in the public domain. There are no royalty fees, licensing expenses, or proprietary paywalls associated with utilizing, replicating, or adapting the scale for academic research, school-based health screenings, or clinical prevention trials. Investigators are expected to maintain scientific integrity by citing the foundational development publications (Flewelling, Paschall, & Ringwalt, 1993; Dahlberg et al., 2005) in subsequent scholarly works.

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. (1978). Social learning theory of aggression. Journal of Communication, 28(3), 12–29. https://doi.org/10.1111/j.1460-2466.1978.tb01621.x
• 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., pp. 179–180). Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. https://files.eric.ed.gov/fulltext/ED486261.pdf
• Flewelling, R. L., Paschall, M. J., & Ringwalt, C. L. (1993). SAGE Baseline Survey. Research Triangle Park, NC: Research Triangle Institute.
• Jessor, R. (1987). Problem-behavior theory, psychosocial development, and adolescent problem drinking. British Journal of Addiction, 82(4), 331–342. https://doi.org/10.1111/j.1360-0443.1987.tb01490.x
• Jessor, R., & Jessor, S. L. (1977). Problem Behavior and Psychosocial Development: A Longitudinal Study of Youth. Academic Press.
• Paschall, M. J., Ennett, S. T., & Flewelling, R. L. (1996). Relationships among adolescent pregnancy, pregnancy resolution, and subsequent substance use. Journal of Adolescent Health, 18(6), 405–412. https://doi.org/10.1016/1054-139X(95)00271-K
• Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. B. (1996). The revised Conflict Tactics Scales (CTS2): Development and preliminary psychometric data. Journal of Family Issues, 17(3), 283–316. https://doi.org/10.1177/019251396017003001

Items of the Scale