Friends’ Delinquent Behavior—Denver Youth Survey

1. Abstract

The Friends’ Delinquent Behavior—Denver Youth Survey (FDB-DYS) is an established psychometric assessment designed to evaluate an adolescent’s or young adult’s self-reported exposure to and affiliation with peer delinquency. Originally developed as part of the longitudinal Denver Youth Survey by researchers at the Institute of Behavioral Science at the University of Colorado Boulder, this 8-item measurement tool quantifies peer involvement across multiple domains of rule-violating behavior, including property damage, interpersonal violence, substance consumption, illicit drug distribution, and weapon carrying. Each item is scored on a 5-point Likert-type frequency/proportion metric ranging from 0 (“None of them”) to 4 (“All of them”). Total scores are calculated by computing an unweighted mean across all answered items, yielding an index ranging from 0.00 to 4.00, where higher scores reflect greater saturation of delinquent peer networks. Across longitudinal waves and independent replications, the instrument demonstrates robust internal consistency (Cronbach’s alpha coefficients routinely between .78 and .88), coherent unidimensional factor structure, and pronounced predictive validity for personal delinquency, violent offenses, and substance misuse. Recognized in the Centers for Disease Control and Prevention (CDC) compendium of violence-related assessment tools, the FDB-DYS represents a benchmark instrument in developmental criminology, clinical adolescent psychology, and public health violence prevention programs.

2. Keywords

peer delinquency, Denver Youth Survey, juvenile offending, differential association, adolescent risk behavior, psychometrics, peer contagion, substance use, weapon carrying, developmental criminology

3. Authors

The Friends’ Delinquent Behavior measure was designed and validated by principal investigators associated with the Denver Youth Survey at the Institute of Behavioral Science, University of Colorado Boulder, in collaboration with national research initiatives funded by the Office of Juvenile Justice and Delinquency Prevention (OJJDP):

• David Huizinga, Ph.D. — Senior Research Associate, Institute of Behavioral Science, University of Colorado Boulder. Primary Investigator on the Denver Youth Survey and renowned scholar in longitudinal developmental epidemiology.
• Finn-Aage Esbensen, Ph.D. — Research Associate, Institute of Behavioral Science, University of Colorado Boulder (later E. Desmond Lee Endowed Professor of Criminology and Criminal Justice at the University of Missouri–St. Louis).
• Delbert S. Elliott, Ph.D. — Distinguished Professor Emeritus of Sociology and Director of the Center for the Study and Prevention of Violence (CSPV), Institute of Behavioral Science, University of Colorado Boulder; principal investigator of the National Youth Survey.

Institutional Origin: Institute of Behavioral Science, University of Colorado Boulder, Campus Box 483, Boulder, CO 80309, USA.

4. Purpose

The primary clinical and empirical objective of the Friends’ Delinquent Behavior—Denver Youth Survey is to provide a standardized, psychometrically rigorous quantification of an adolescent’s social proximity to deviant and law-violating peer networks. In developmental psychopathology, sociology, and criminology, association with delinquent peers represents one of the most potent, proximal, and consistent statistical predictors of onset, escalation, and persistence in antisocial behavior during the transition from early adolescence to emerging adulthood.

Historically, delinquency research struggled with measurement artifacts, particularly projection bias and halo effects, wherein adolescents overestimate their peers’ deviance to justify their own conduct. The FDB-DYS addresses this need by utilizing concrete, behavioral behavioral referents anchored to a distinct temporal window (“During the last year…”) and structured around proportion-based response anchors (from “None of them” to “All of them”). Rather than asking broad subjective impressions (e.g., “Are your friends bad influences?”), the scale targets specific manifestations of externalizing problems ranging from normative adolescent boundary-testing (e.g., alcohol experimentation) to severe index offenses (e.g., weapon carrying and drug distribution).

In clinical and educational settings, the tool aids mental health practitioners, juvenile diversion counselors, and social workers in identifying youth immersed in criminogenic social milieus. Because peer contagion significantly attenuates the efficacy of conventional group-based behavioral therapies (often generating iatrogenic deviancy training effects), screening with the FDB-DYS enables clinicians to structure individualized interventions, implement family-focused modalities such as Multisystemic Therapy (MST) or Functional Family Therapy (FFT), and alter network-level structural incentives.

In longitudinal research, the instrument serves as a critical time-varying covariate, mediator, or moderator in structural equation modeling (SEM). It allows researchers to disentangle peer selection effects (homophily, where deviant youth actively seek out delinquent peers) from peer socialization effects (influence processes, where affiliation with rule-breaking peers initiates or amplifies individual delinquency).

5. Psychological Construct

The FDB-DYS measures perceived peer delinquency, conceptualized as a multi-domain externalizing construct reflecting the density, prevalence, and diversity of unlawful and antisocial behaviors executed by members of an adolescent’s immediate peer cluster. Although scored as a single unidimensional composite for applied indexation, the operationalized construct encompasses four distinct conceptual facets:

1. Vandalism and Property Destruction

Reflected in Item 1 (“Purposely damaged or destroyed property that did not belong to them?”), this dimension measures covert externalizing deviance and non-confrontational property offenses. Property destruction frequently emerges during middle childhood or early adolescence as an initial foray into collective delinquency, often serving as a shared bonding ritual that tests peer loyalty through defiance of community norms.

2. Interpersonal Aggression and Violent Conduct

Captured across Items 2 (“Hit or threatened to hit someone?”), 7 (“Got into a physical fight?”), and 8 (“Been hurt in a fight?”), this facet indexes direct overt physical aggression, threatening communication, and physical victimization within peer conflict. By including both perpetration (hitting, fighting) and consequences (being hurt in a fight), the scale taps into the reciprocal, chaotic physical environments characteristic of violent street peer groups and juvenile fighting cultures.

3. Substance Involvement and Illicit Markets

Spanning Items 3 (“Used alcohol?”), 4 (“Sold drugs?”), and 5 (“Gotten drunk once in a while?”), this dimension captures both recreational substance experimentation and participation in the illegal drug economy. While moderate alcohol use is statistically prevalent during late adolescence, coupling it with public drunkenness and drug trafficking taps into severe, systemic deviance associated with organized illicit trade networks.

4. Lethal Weapon Carrying

Operationalized in Item 6 (“Carried a knife or a gun?”), this item represents high-severity, high-stakes criminogenic conduct. Weapon carrying in adolescent peer groups drastically elevates the lethality of spontaneous adolescent altercations and functions as an indicator of defensive posture, status signaling, and gang-related affiliation.

6. Theoretical Framework

The design and interpretation of the Friends’ Delinquent Behavior inventory are rooted in three foundational criminological and sociological theories:

Differential Association Theory

Formulated by Edwin H. Sutherland (1947), Differential Association Theory posits that criminal behavior is learned through close interpersonal communication within intimate personal groups. According to Sutherland, learning includes both the specific techniques of committing crimes and the directional orientation of motives, drives, rationalizations, and attitudes. When definitions favorable to the violation of law exceed definitions unfavorable to legal violation, an individual chooses delinquency. The FDB-DYS directly captures the social environment providing these excess delinquent definitions.

Social Learning Theory of Crime

Expanding upon Sutherland’s work through the principles of operant conditioning, Ronald L. Akers (1973, 1998) introduced the Social Learning Theory of Crime and Deviance. Akers identified four core mechanisms: differential association, differential reinforcement, imitation (modeling), and cognitive definitions. Delinquent peer groups provide immediate positive reinforcement (e.g., social status, peer approval, laughter, shared material spoils) for externalizing conduct while modeling specific techniques (e.g., how to conceal stolen property or acquire alcohol). The FDB-DYS operationalizes the frequency and proportion of peer models engaging in these target actions.

Interactional Theory and Deviancy Training

The FDB-DYS is also deeply linked to Terence P. Thornberry’s Interactional Theory (1987) and Thomas J. Dishion’s conceptualization of deviancy training. Interactional theory argues that social structural bonds (attachment to parents, commitment to school) weaken over time, facilitating association with delinquent peers in a bidirectional, reciprocal feedback loop. Dishion demonstrated that delinquent peers reinforce rule violations through contingent verbal positive feedback (shared laughter during discussions of rule breaking). The 8 items of the FDB-DYS capture the behavioral manifestations that are continually nurtured and reinforced within these coercive peer interactions.

7. Validity

The psychometric validity of the Friends’ Delinquent Behavior scale has been extensively documented in longitudinal epidemiological samples, urban cohort studies, and independent programmatic evaluations:

Construct and Convergent Validity

Convergent validity is substantiated by strong, statistically significant correlations between the FDB-DYS total score and concurrent self-reported externalizing problems. In the foundational Denver Youth Survey validation samples (N = 1,890), the peer delinquency composite correlated moderately-to-strongly with self-reported general delinquency (r = .42 to .56, p < .001), self-reported violent offending (r = .38 to .49, p < .001), and individual substance abuse indices (r = .44 to .61, p < .001). Furthermore, the scale correlates positively with self-reported gang membership (biserial r = .48, p < .001) and measures of adolescent sensation-seeking and impulsivity, while correlating negatively with academic achievement and conventional parental attachment (r = -.31 to -.43, p < .001).

Predictive and Longitudinal Criterion Validity

Longitudinal structural equation modeling from the Denver Youth Survey demonstrated that baseline scores on the FDB-DYS predicted subsequent self-reported arrests, formal police contact, and chronic offending trajectories over 1-, 2-, and 4-year follow-up intervals, even after strictly controlling for baseline individual delinquency. Youth scoring in the upper quartile of peer delinquency were between 3.5 and 5.2 times more likely to be arrested during subsequent assessment waves than those in the lowest quartile.

Discriminant Validity

Discriminant validity has been demonstrated by showing that the correlation between the FDB-DYS and internalizing symptoms (such as generalized anxiety, somatic complaints, and unipolar depressive symptoms) is consistently modest (r = .10 to .19), confirming that the instrument specifically indexes externalizing social influence rather than diffuse negative emotionality or generalized adolescent distress.

8. Reliability

The psychometric reliability of the FDB-DYS has demonstrated sustained stability across multiple demographic strata and developmental epochs:

Internal Consistency

Across the longitudinal waves of the Denver Youth Survey (ages 12 to 24), the 8-item inventory demonstrated high internal consistency. In the baseline representative urban cohort of 1,890 youth, Cronbach’s alpha coefficients ranged from .78 to .88 across demographic subgroups. Replications in independent urban and rural cohorts yielded comparable internal consistency estimates:

• Overall baseline urban sample: Cronbach’s α = .84
• Male adolescents: Cronbach’s α = .85
• Female adolescents: Cronbach’s α = .82
• Sub-sample across racial/ethnic groups (African American, Hispanic, Caucasian): Cronbach’s α = .80 – .86

Test-Retest Stability

Because the scale is anchored to an extended retrospective reporting window (“During the last year…”), short-term test-retest reliability across 2- to 4-week intervals yielded high stability coefficients (Pearson’s r = .81 to .87; intraclass correlation coefficient [ICC] = .83). Long-term year-over-year autocorrelations across consecutive annual waves ranged between r = .52 and .64, which is theoretically congruent with both developmental continuity in friendship networks and expected normative turnover in adolescent peer associations.

9. Factor Analysis

The dimensional structure of the 8-item scale has been evaluated through both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA):

Exploratory Factor Analysis (EFA)

Principal axis factoring and maximum likelihood EFA conducted on early adolescent cohorts consistently reveal a strong primary dimension. The initial eigenvalue of the first unrotated factor regularly accounts for 45% to 58% of the total variance, with a steep drop-off to the second eigenvalue (e.g., λ₁ = 3.92, λ₂ = 0.88), firmly supporting an essentially unidimensional latent construct of general peer delinquency. Factor loadings across all 8 items are consistently robust, rarely falling below .50:

• Item 1 (Damaged/destroyed property): λ = .58 – .68
• Item 2 (Hit or threatened to hit someone): λ = .66 – .76
• Item 3 (Used alcohol): λ = .52 – .64
• Item 4 (Sold drugs): λ = .55 – .69
• Item 5 (Gotten drunk once in a while): λ = .58 – .71
• Item 6 (Carried a knife or a gun): λ = .60 – .72
• Item 7 (Got into a physical fight): λ = .72 – .82
• Item 8 (Been hurt in a fight): λ = .63 – .74

Confirmatory Factor Analysis (CFA)

Confirmatory factor analytic models testing a single-factor specification have demonstrated satisfactory goodness-of-fit indices across adolescent samples after accounting for minor residual covariance between conceptually proximate item pairs (e.g., between Item 3 [used alcohol] and Item 5 [gotten drunk], or between Item 7 [got into a fight] and Item 8 [hurt in a fight]):

• Comparative Fit Index (CFI) = .962 to .981
• Tucker-Lewis Index (TLI) = .948 to .973
• Root Mean Square Error of Approximation (RMSEA) = .044 (90% CI [.036, .053])
• Standardized Root Mean Square Residual (SRMR) = .031

Multi-group measurement invariance analyses confirm full metric and scalar invariance across biological sex and ethnic subgroups, ensuring that mean-level comparisons reflect true variations in peer delinquency exposure rather than differential item functioning.

10. Instrument / Measurement Tool

• Instrument Name: Friends’ Delinquent Behavior—Denver Youth Survey (FDB-DYS)
• Target Population: Adolescents and young adults (ages 12 to 24)
• Administration Format: Self-administered paper-and-pencil questionnaire, computer-assisted personal interviewing (CAPI), or audio computer-assisted self-interviewing (ACASI)
• Number of Items: 8 items
• Temporal Referent: “During the last year…”
• Response Format: 5-point proportion scale:
  • 0 = None of them
  • 1 = Very few of them
  • 2 = Some of them
  • 3 = Most of them
  • 4 = All of them
• Scoring Protocol: Item responses are scored numerically from 0 to 4. Total score is computed as the unweighted arithmetic mean of all 8 items:
  
  FDB Score = (Sum of Item Responses) / (Total Number of Items Completed)
• Score Range: 0.00 to 4.00 (continuous). A higher score indicates greater association with or saturation of delinquent peer networks. If missing data are present, scores can be averaged across completed items provided that at least 6 of the 8 items (75%) were answered.

11. Permissions & Fee and Test Year

The Friends’ Delinquent Behavior scale was finalized and deployed in 1987 as an integral module of the Youth Interview Schedule for the Denver Youth Survey by the Institute of Behavioral Science at the University of Colorado Boulder. The development of the survey was supported by federal grant funding from the Office of Juvenile Justice and Delinquency Prevention (OJJDP), U.S. Department of Justice.

Licensing and Availability: The assessment is in the public domain as an instrument developed under federal research sponsorship and cataloged within government technical resources, including the Centers for Disease Control and Prevention compendium, Measuring Violence-Related Attitudes, Behaviors, and Influences Among Youths: A Compendium of Assessment Tools (Dahlberg et al., 2005, pp. 212–213). It is free to use for academic, clinical, and non-profit research purposes without payment of royalty fees. Researchers are expected to appropriately cite the Institute of Behavioral Science and the original Denver Youth Survey publications.

12. References

• Akers, R. L. (1998). Social learning and social structure: A general theory of crime and deviance. Northeastern University Press.
• 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. 212–213). Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. https://www.cdc.gov/violenceprevention/pdf/YV_Compendium.pdf
• Dishion, T. J., McCord, J., & Poulin, F. (1999). When interventions harm: Peer groups and problem behavior. American Psychologist, 54(9), 755–764. https://doi.org/10.1037/0003-066X.54.9.755
• Elliott, D. S., Huizinga, D., & Ageton, S. S. (1985). Explaining delinquency and drug use. SAGE Publications.
• Esbensen, F.-A., & Huizinga, D. (1993). Gangs, drugs, and delinquency in a survey of youth. Criminology, 31(4), 565–589. https://doi.org/10.1111/j.1745-9125.1993.tb01142.x
• Huizinga, D., Esbensen, F.-A., & Weiher, A. W. (1991). Are there multiple paths to delinquency? Journal of Criminal Law and Criminology, 82(1), 83–118. https://doi.org/10.2307/1143809
• Institute of Behavioral Science. (1987). Youth Interview Schedule: Denver Youth Survey. University of Colorado Boulder.
• 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

13. Items of the Scale