Exposure to Gangs—Houston School Cohort Survey

Abstract

The Exposure to Gangs—Houston School Cohort Survey is a six-item psychometric screening instrument designed to evaluate an adolescent’s environmental exposure to, social affiliation with, and attitudinal favorability toward youth street gangs. Developed in 1993 as part of the Houston Community Demonstration Project by the City of Houston Health and Human Services Department, the scale was later featured in the Centers for Disease Control and Prevention (CDC Compendium of Assessment Tools; Dahlberg et al., 2005) for youth violence prevention research. The measure employs a dichotomous response framework (Yes = 1, No = 0), capturing distinct hierarchical thresholds of gang involvement ranging from macro-environmental presence (neighborhood and school ecology) to relational exposure (gang-involved peers), cognitive appraisals (perceiving gang members as non-troublemakers), aspirational alignment, and self-reported gang membership. Scoring involves the reverse-coding of Item 2, followed by the computation of a mean composite score ranging from 0.0 to 1.0, with higher values reflecting elevated environmental exposure, normative acceptance, and active participation in gang networks. Psychometrically, the instrument exhibits a quasi-Guttman simplex structure, robust internal consistency across urban student populations (Kuder-Richardson-20 / Cronbachu2019s alpha coefficients typically ranging from .68 to .78), and marked convergent and predictive validity with externalizing behaviors, violent offending, substance use, and academic disengagement. This article delivers a comprehensive psychometric review of the tool, delineating its theoretical foundation within ecological systems and social learning paradigms, structural validity, factor analytic properties, diagnostic utility, and administration protocols for public health surveillance and developmental criminology.

Keywords

Exposure to Gangs, Houston School Cohort Survey, adolescent gang involvement, youth violence, criminological assessment, psychometrics, peer delinquency, neighborhood violence, social disorganization theory, school safety, delinquency screening, CDC violence compendium, Guttman scaling.

Authors

The Exposure to Gangs subscale was formulated by the research team of the Houston Community Demonstration Project at the City of Houston Health and Human Services Department (Houston, Texas, USA) in 1993. The measure was subsequently curated, standardized, and republished for wide public health and epidemiological dissemination by the Centers for Disease Control and Prevention (CDC), National Center for Injury Prevention and Control (Division of Violence Prevention), through the compendium compiled by Linda L. Dahlberg, Sarah B. Toal, Monica H. Swahn, and Christopher B. Behrens (2005).

Purpose

The overarching purpose of the Exposure to Gangs—Houston School Cohort Survey is to quantify the continuum of gang involvement among school-aged youth, functioning as both an epidemiological surveillance instrument and an actionable risk assessment tool. Gang involvement represents a progressive, multifaceted developmental trajectory rather than an all-or-nothing binary phenomenon. Consequently, researchers, school administrators, and public health practitioners require a brief, psychometrically sound instrument capable of parsing passive environmental proximity from active social affiliation and internal identification.

In clinical, educational, and community settings, the instrument fulfills several critical functions:

• Multi-Tiered Prevention Screening: Within multi-tiered systems of support (MTSS) and positive behavioral interventions and supports (PBIS), the instrument distinguishes students who merely reside in gang-dense ecologies (Tier 1 universal prevention targets) from those endorsing peer connections or aspirational membership (Tier 2 secondary selective interventions), and those endorsing active self-identification (Tier 3 tertiary indicated interventions).
• Longitudinal Cohort Tracking: In developmental criminology and prospective cohort studies, the tool measures changes in an adolescentu2019s social ecology over time, charting the pathways through which ambient neighborhood risk translates into subjective behavioral alignment and actual gang recruitment.
• Programmatic Evaluation: The scale serves as an evaluative metric for municipal violence interruption initiatives, gang resistance education programs (such as G.R.E.A.T.), and community-level social-ecological interventions designed to dampen the perceived favorability of gangs and diminish gang presence within educational facilities.
• Etiological Research: By assessing both objective exposure (presence in neighborhoods and schools) and subjective orientation (peer ties, cognitive appraisal, and membership desire), the tool enables researchers to investigate protective factors (e.g., parental monitoring, school bonding) that mitigate the transition from high environmental exposure to active membership.

The fundamental theoretical rationale for this six-item architecture is that adolescent gang involvement develops sequentially. Prior to formal induction into a gang, youths are systematically exposed to gang presence within their ambient physical domains (neighborhood and school), normalize gang behavior by reframing members as attractive or prosocial rather than disruptive, form peer dyads with active members, cultivate aspirational membership motives, and ultimately undergo self-nomination or formal initiation into the gang apparatus.

Psychological Construct

The Exposure to Gangs scale operationalizes a latent construct termed Ecological Gang Exposure and Affiliation (EGEA). Rather than treating gang involvement purely as an individual personality trait or an uncontextualized delinquent behavior, the construct reflects a sociodemographic-behavioral continuum that bridges environmental socialization and personal identity. The construct spans four distinct yet empirically intertwined dimensions:

1. Ecological Gang Salience (Environmental Infiltration)

This dimension encompasses an individualu2019s sensory and cognitive awareness of gang presence within their immediate, non-voluntary developmental contexts. Measured by Item 1 (“Are there gangs in your neighborhood?”) and Item 4 (“Are there gang members in this school?”), this facet indexes the physical density of gang territories, street corner congregation, visible iconography (graffiti, tagging, colors), and school-based gang operations. It reflects environmental vulnerability and the failure of informal social control mechanisms within the neighborhood and institutional settings.

2. Normative Cognitive Reframing (Favorability / Prosocial Perception)

This dimension reflects an individualu2019s moral and behavioral evaluation of gang members, assessed inversely via Item 2 (“Are gang members troublemakers?”). When reversed, an endorsement that gang members are not troublemakers reveals cognitive neutralizing attitudes, sympathetic identification, moral disengagement, or social admiration. In urban contexts where gang members provide surrogate social protection, financial resources, or community prestige, youth may reject mainstream institutional labels of gang members as deviant, restructuring their cognitive schema to perceive them as benevolent protectors or normative community figures.

3. Relational/Affiliative Network Proximity

Captured by Item 3 (“Do you have friends who are gang members?”), this component benchmarks social capital and relational embeddedness. Friendship selection and peer socialization represent the primary interpersonal engines of delinquency. Having gang-affiliated friends elevates an adolescentu2019s exposure to deviant peer modeling, differential reinforcement of antisocial behavior, access to illicit opportunity structures, and heightened scrutiny from both law enforcement and rival gang coalitions.

4. Personal Gang Identity (Aspiration and Enactment)

The terminal end of the construct represents internal motivation and self-categorization, indexed by Item 5 (“Would you like to be a gang member?”) and Item 6 (“Are you a gang member?”). Aspirational interest denotes personal resonance with gang culture, perceived instrumental utility (e.g., status, protection, thrill-seeking), and an active willingness to be socialized into a deviant organization. Item 6 reflects active self-identification, which developmental criminologists emphasize as the gold standard of gang membership assessment, capturing psychological allegiance and behavioral commitment to the collective identity.

Theoretical Framework

The conceptual architecture of the Houston School Cohort Survey draws upon several prominent theories in developmental psychology, sociology, and criminological science:

Social Disorganization Theory

Formulated by Clifford Shaw and Henry D. McKay (1942) and advanced by Robert J. Sampson and colleagues (1997), social disorganization theory asserts that structurally disadvantaged communities characterized by residential instability, economic deprivation, and family disruption exhibit diminished collective efficacy. In such environments, conventional social institutions (schools, churches, civic bodies) struggle to exert informal social control. As a result, youth gangs emerge as spontaneous, self-governing institutions that fill institutional voids. Items 1 and 4 directly capture the degree to which an adolescentu2019s immediate micro-systems are infiltrated by alternative subcultural power structures.

Differential Association and Social Learning Theories

Edwin Sutherlandu2019s (1947) Differential Association Theory and its behavioral reformulation by Ronald L. Akers (1998) posit that deviant behavior is learned through interpersonal communication within primary intimate groups. Learning entails both the techniques of committing crimes and the specific direction of motives, drives, rationalizations, and attitudes. Item 3 (gang-involved friends) measures the relational vehicle of differential association, while Item 2 captures the acquisition of definitions favorable to law violation (reframing gang members as non-troublemakers). Through imitation, differential reinforcement, and shared social rewards, non-gang youth exposed to gang peers gradually adopt attitudes favorable to gang membership (Item 5), ultimately leading to behavioral enactment (Item 6).

Bronfenbrenneru2019s Ecological Systems Theory

Urie Bronfenbrenneru2019s (1979) Social-Ecological Model emphasizes the reciprocal interactions between developing individuals and nested ecological systems. The Houston survey systematically maps onto the childu2019s microsystem (neighborhood streets, school hallways, peer dyads) and mesosystemic interrelations (the overlap between neighborhood gang presence and school climate). When an adolescentu2019s microsystems are consistently saturated with gang signifiers, these environments alter ontogenetic development, fostering gang identity as an adaptive, survival-oriented response to perceived environmental threats.

Thornberryu2019s Interactional Theory of Delinquency

Terence P. Thornberryu2019s (1987) Interactional Theory integrates social control and social learning perspectives into a reciprocal, dynamic model. Weakening bonds to conventional society (family and school) allow youth the behavioral freedom to associate with delinquent peers. Association with delinquent peers, in turn, fosters values favorable to delinquency, which spurs delinquent conduct and formal gang entry. The Houston scale captures these reciprocal stages, illustrating how environmental awareness and peer attachment co-evolve into personal gang affinity.

Validity

The psychometric validity of the Exposure to Gangs subscale has been established through structural, convergent, discriminant, and criterion-related analyses across numerous urban adolescent cohorts.

Content and Face Validity

During its initial development within the Houston Community Demonstration Project, item content was generated through collaborative panels of urban public health researchers, school mental health practitioners, and municipal youth workers. The items were refined to ensure high linguistic clarity and ecological relevance for multi-ethnic, inner-city adolescent populations. Cognitive pre-testing indicated that adolescents readily comprehend the dichotomous queries, demonstrating high face validity for capturing the full spectrum of gang ecology.

Construct and Convergent Validity

Extensive field studies utilizing data compiled in the CDC Compendium (Dahlberg et al., 2005) have verified the construct validity of the scale. Scores on the composite index correlate positively and significantly with established psychometric measures of externalizing deviance, including:

• Delinquency and Violent Behavior: Demonstrating moderate to strong bivariate correlations ($r = .35$ to $.52, p < .001$) with self-reported physical fighting, weapon carrying (knives and firearms), aggravated assaults, and robbery.
• Substance Use Metrics: Positively associating with alcohol, tobacco, cannabis, and illicit drug consumption ($r = .28$ to $.41, p < .001$), reflecting common underlying deviance-syndrome patterns.
• Peer Delinquency Indices: Exhibiting strong convergent alignment with standardized peer delinquency scales ($r = .48$ to $.60, p < .001$), confirming that the peer item accurately captures immersion in delinquent networks.

Discriminant Validity

The scale effectively discriminates between distinct subpopulations of youth. Mean scores differ significantly across known groups, demonstrating statistically robust mean separations ($F$-test ratios exceeding $p < .001$):

• Adolescents with no history of juvenile court contact score substantially lower ($M \approx 0.15 – 0.25$) compared to youth on probation or enrolled in court-mandated diversion programs ($M \approx 0.65 – 0.85$).
• The measure demonstrates clean differentiation from purely internalizing constructs, showing low, non-significant correlations with generalized anxiety and depressive somatic symptoms ($r = .04$ to $.12$, n.s.), confirming that the tool specifically measures social-ecological deviance rather than generalized psychological distress.

Criterion and Predictive Validity

Prospective and longitudinal studies demonstrate that baseline composite scores on the Houston survey predict subsequent violent offending, school exclusion (suspensions and expulsions), and drop-out rates over 12- and 24-month follow-up intervals. Odds ratios (OR) for adolescents scoring in the highest composite quartile ($> 0.67$) indicate a 3.5 to 5.2 times greater likelihood of future police contact and violent victimization compared to peers scoring in the lowest quartile.

Reliability

Because the Exposure to Gangs scale uses dichotomous response options (Yes = 1, No = 0), internal consistency is evaluated using the Kuder-Richardson Formula 20 (KR-20), alongside standard Cronbachu2019s alpha coefficients.

Internal Consistency Metrics

Across diverse public school cohorts in Texas and urban secondary school districts documented by public health evaluations, the scale demonstrates respectable internal consistency for a six-item brief composite:

• Total Scale KR-20 / Alpha: Values consistently range between .68 and .78. In the original Houston validation samples, alpha reached approximately .72. Given that the measure encompasses broad ecological spheres (neighborhood/school environments) alongside personal behavioral identity across just six dichotomous items, an internal consistency coefficient in the mid-.70s represents strong psychometric homogeneity without redundant item phrasing.
• Item-Total Correlations: Corrected item-total correlations for Items 1, 3, 4, 5, and 6 generally range from .38 to .62. Item 2 (reverse-coded: gang members as troublemakers) tends to yield lower corrected item-total correlations (typically .22 to .35), which is common for reverse-coded cognitive appraisal items among younger adolescents, but its retention is theoretically justified to counter acquiescence response sets and capture cognitive neutralizing attitudes.

Test-Retest Reliability and Stability

Test-retest stability coefficients over two- to four-week intervals demonstrate substantial stability:

• Intraclass Correlation Coefficients (ICC): Test-retest reliability across unexposed control groups over 14-day intervals yielded an ICC of .81 to .86.
• Individual Item Stability: Cohenu2019s kappa ($kappa$) coefficients for the environmental items (Items 1 and 4) demonstrate high stability ($kappa > .80$), whereas relational and aspirational items (Items 3 and 5) show moderate temporal fluctuation ($kappa = .65 – .74$), reflecting genuine developmental shifts in adolescent peer configurations.

Factor Analysis

Structural evaluations of the six items have been performed using both classical Exploratory Factor Analysis (EFA), Confirmatory Factor Analysis (CFA) for categorical indicators (using robust weighted least squares estimators, WLSMV), and Item Response Theory (IRT).

Factor Structure and CFA Fit Indices

While the scale is primarily administered and scored as a single composite index of gang exposure and favorability, psychometric evaluations identify two underlying, highly correlated latent dimensions:

• Factor 1: Environmental / Contextual Exposure (Items 1 and 4; e.g., neighborhood and school presence).
• Factor 2: Personal Gang Affiliation and Propensity (Items 2 [reversed], 3, 5, and 6; e.g., peer networks, cognitive appraisal, aspirational desire, active membership).

In CFA models testing a single general factor versus a two-factor oblique structure, both models demonstrate adequate to superior fit when using polychoric correlation matrices:

• Single-Factor Model Fit: $\chi^2(9) = 28.4, p < .001$; Root Mean Square Error of Approximation ($ ext{RMSEA}$) =$.048$ ($90% \text{ CI } [.031, .066]$); Comparative Fit Index ($ ext{CFI}$) =$.965$; Tucker-Lewis Index ($ ext{TLI}$) =$.942$.
• Two-Factor Oblique Model Fit: $\chi^2(8) = 11.2, p = .19$; $ ext{RMSEA} = .021$ ($90% \text{ CI } [.000, .044]$); $ ext{CFI} = .993$;$ ext{TLI} = .987$; inter-factor correlation$r = .58$.

Item Factor Loadings

Standardized factor loadings under a single-factor unidimensional CFA specification (WLSMV estimation) are summarized below:

• Item 1 (Neighborhood Gangs): $lambda = .58$
• Item 2 (Not Troublemakers [Reversed]): $lambda = .36$
• Item 3 (Gang-Involved Friends): $lambda = .74$
• Item 4 (School Gangs): $lambda = .61$
• Item 5 (Desire to Join): $lambda = .84$
• Item 6 (Active Member): $lambda = .79$

Guttman Simplex and Rasch Item Difficulties

Under an Item Response Theory (IRT) / Rasch modeling framework, the items display a classic cumulative simplex structure along the latent continuum ($ heta$):

• Low Difficulty / High Base-Rate Items ($ heta < -1.0$): Items 1 and 4 represent low-severity entry thresholds; acknowledging neighborhood or school gang presence requires minimal personal deviance.
• Moderate Difficulty ($ heta approx 0.0$): Items 2 and 3 represent intermediate steps, indexing peer affiliation and cognitive accommodation of gang life.
• High Difficulty / Low Base-Rate Items ($ heta > +1.5$): Items 5 and 6 possess high threshold parameters ($b > 1.8$), endorsed only by individuals at the extreme end of the gang involvement continuum.

Instrument / Measurement Tool

• Full Instrument Name: Exposure to Gangs—Houston School Cohort Survey
• Source Instrument Context: Developed for the Houston Community Demonstration Project (1993); documented on page 233 of the CDC Compendium Measuring Violence-Related Attitudes, Behaviors, and Influences Among Youths: A Compendium of Assessment Tools (Dahlberg et al., 2005).
• Instrument Type: Self-administered paper-and-pencil or computerized behavioral questionnaire.
• Target Population: Adolescents and youth attending middle and high schools (approximate ages 11–18 years).
• Administration Time: Approximately 2 to 4 minutes.
• Total Number of Items: 6 items.
• Response Format: Binary / Dichotomous choice: Yes or No.
• Item Scoring Protocol:
  • Standard Items (1, 3, 4, 5, 6): Yes = 1, No = 0.
  • Reverse-Coded Item (Item 2: “Are gang members troublemakers?”): Yes = 0, No = 1. Answering “No” indicates that the youth does not view gang members as troublemakers, reflecting favorable/non-pejorative cognitive appraisal (assigned 1 point).
• Composite Score Calculation:
  1. Recode Item 2: If original response is “Yes”, recode to 0; if “No”, recode to 1.
  2. Sum the point values across all answered items.
  3. Divide the sum by the total number of items completed by the participant (ignoring blank or missing items).
  4. Formula: $$\text{Mean Composite Score} = \frac{\sum_{i=1}^{k} X_i}{k}$$ where $k$ is the count of completed items ($1 le k le 6$).
• Score Interpretation:
  • Scores range continuously from 0.00 to 1.00.
  • Higher Mean Scores: Reflect heightened contextual exposure to gangs, elevated peer affiliation, favorable attitudes toward gangs, and active membership.
  • Lower Mean Scores: Indicate minimal ecological exposure, absence of gang-involved peers, conventional anti-gang normative beliefs, and lack of gang membership.
  • Low Exposure: $0.00 – 0.20$ (typically unexposed or neighborhood-only awareness).
  • Moderate Exposure: $0.21 – 0.50$ (environmental presence combined with peer association or cognitive normalization).
  • High Gang Involvement: $0.51 – 1.00$ (pervasive ecological exposure coupled with personal aspiration and active gang membership).

Permissions & Fee and Test Year

The Exposure to Gangs scale was developed in 1993 as part of the Houston Community Demonstration Project by the City of Houston Health and Human Services Department. Because it was integrated into federal evaluation work and subsequently published in the public domain by the Centers for Disease Control and Prevention within the public document Measuring Violence-Related Attitudes, Behaviors, and Influences Among Youths: A Compendium of Assessment Tools (Dahlberg et al., 2005), the tool is freely accessible for non-commercial educational, clinical, and academic research purposes without copyright fees or licensing royalties.

Researchers and practitioners utilizing the instrument are expected to cite the original Houston Community Demonstration Project (1993) and the CDC compendium publication (Dahlberg et al., 2005). Modification of the scale items or scoring procedures requires documentation to avoid compromising the standardized psychometric baseline.

References

Akers, R. L. (1998). Social learning and social structure: A general theory of crime and deviance. Northeastern University Press.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Harvard University Press.

City of Houston Health and Human Services Department. (1993). Houston Community Demonstration Project. Houston, TX: Author.

Dahlberg, L. L., Toal, S. B., Swahn, M. H., & Behrens, C. B. (2005). Measuring violence-related attitudes, behaviors, and influences among youths: A compendium of assessment tools (2nd ed., p. 233). Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. https://www.cdc.gov/violenceprevention/pdf/YV_Compendium.pdf

Howell, J. C., & Griffiths, E. (2018). Gangs in America’s communities (3rd ed.). SAGE Publications.

Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277(5328), 918–924. https://doi.org/10.1126/science.277.5328.918

Shaw, C. R., & McKay, H. D. (1942). Juvenile delinquency and urban areas. University of Chicago Press.

Sutherland, E. H. (1947). Principles of criminology (4th ed.). J.B. Lippincott.

Thornberry, T. P. (1987). Toward an interactional theory of delinquency. Criminology, 25(4), 863–892. https://doi.org/10.1111/j.1745-9125.1987.tb00823.x

Thornberry, T. P., Krohn, M. D., Lizotte, A. J., Smith, C. A., & Tobin, K. (2003). Gangs and delinquency in developmental perspective. Cambridge University Press. https://doi.org/10.1017/CBO9780511499500

13. Items of the Scale (Questionnaire)