Likelihood of Violence and Delinquency scale

Abstract

The Likelihood of Violence and Delinquency Scale (LVDS) is a brief, self-report psychometric instrument designed to evaluate adolescents’ perceived subjective probability of engaging in interpersonal aggression, weapon carrying, substance use, and related high-risk delinquent behaviors in the near future. Originally developed by Robert L. Flewelling, Mallie J. Paschall, and Christopher L. Ringwalt as part of the Research Triangle Institute‘s (RTI) SAGE (Substance Abuse and Gang Education) Baseline Survey, the scale subsequently achieved wide visibility via the Centers for Disease Control and Prevention (CDC) compendium of violence-related assessment tools (Dahlberg et al., 1998). Composed of 9 items targeting an immediate temporal horizon (the subsequent 30 days), the instrument utilizes a 4-point Likert-type response scale ranging from 1 (Not at all likely) to 4 (Very likely). Psychometrically, the LVDS demonstrates strong internal consistency reliability (Cronbach’s alpha coefficients routinely ranging from α = .80 to .86 among normative youth samples aged 12 to 16) and robust predictive validity for actual delinquency, school disciplinary infractions, weapon involvement, and externalizing behavioral problems. Principal factor analyses support either a unidimensional general risk-taking model or a bifactor structure that distinguishes overt physical violence/weapon carriage from general risk-taking and substance use. This comprehensive review examines the psychometric architecture, theoretical framework (grounded in the Theory of Reasoned Action, Problem Behavior Theory, and Social Cognitive Theory), structural validity, scoring protocols, and operational utility of the LVDS across epidemiological surveillance, school-based intervention evaluation, and adolescent behavioral health monitoring.

Keywords

Likelihood of Violence and Delinquency Scale, LVDS, adolescent delinquency, youth violence, risk assessment, behavioral intentions, weapon carrying, substance misuse, psychometrics, Problem Behavior Theory

Authors

The scale was developed by senior behavioral scientists and research epidemiologists at the Research Triangle Institute (RTI International) in Research Triangle Park, North Carolina:

• Robert L. Flewelling, Ph.D.: Senior Research Epidemiologist and Behavioral Scientist, Substance Abuse and Youth Risk Behavior Programs, Research Triangle Institute, Research Triangle Park, NC.
• Mallie J. Paschall, Ph.D.: Senior Research Scientist, Prevention Research Center, Pacific Institute for Research and Evaluation (PIRE), Berkeley, CA; formerly Research Sociologist at Research Triangle Institute.
• Christopher L. Ringwalt, Dr.P.H.: Senior Research Scientist, Pacific Institute for Research and Evaluation (PIRE), Chapel Hill, NC; formerly Program Director for Youth and Community Studies at Research Triangle Institute.

Institutional contact and administrative information at the time of development:

Organization: Research Triangle Institute
Address: P.O. Box 12194, Research Triangle Park, NC 27709-2194, USA
Phone: (866) 784-1958
Email: [email protected] | Website: www.rti.org

Purpose

The fundamental purpose of the Likelihood of Violence and Delinquency Scale (LVDS) is to quantify an adolescent’s prospective behavioral expectations regarding imminent involvement in serious antisocial conduct, physical confrontation, weapon possession, substance misuse, and sexual precocity. Unlike purely retrospective delinquency inventories—which measure lifetime or past-year prevalence of infractions—the LVDS captures current, cognitive forward-looking estimates within an anchored 30-day window (“Within the next month, how likely is it that you will…”). This prospective orientation serves critical epidemiological, clinical, and evaluative functions in adolescent behavioral health.

From an applied research perspective, the LVDS was engineered to resolve a classic methodological limitation in prevention science: retrospective behavioral reporting often suffers from recall bias, social desirability suppression, and temporal lag, making it insensitive to rapid intervention-induced changes. By examining self-reported likelihood over the upcoming 30 days, researchers can assess whether school-based violence prevention initiatives, conflict resolution curricula, or community gang diversions successfully alter a youth’s internal calculus and proximate behavioral trajectory before actual physical harm or legal encounters occur.

In clinical, educational, and juvenile diversion contexts, the instrument functions as an efficient, low-burden primary and secondary screening measure. Administering the LVDS allows counselors, school psychologists, and intervention specialists to identify youth who exhibit elevated risk profiles across multiple domains of high-risk behavior. Because delinquent behaviors during early and middle adolescence rarely occur in isolation, the scale deliberately integrates interpersonal violence indicators (fighting, weapon carriage, physical injury) with co-occurring problem behaviors (alcohol use, intoxication, illicit drug consumption, and early sexual intercourse). The scale’s brevity (9 items requiring less than five minutes to complete) ensures high compliance across diverse institutional environments, including alternative schools, community youth centers, adolescent psychiatric clinics, and detention diversion programs.

Psychological Construct

The construct assessed by the LVDS is perceived likelihood of proximate delinquent and violent engagement. Within contemporary developmental psychopathology and behavioral sociology, this construct represents the cognitive operationalization of behavioral intention situated within an ecology of adolescent deviance. Rather than treating violence, delinquency, and substance use as disparate, unrelated phenomena, the construct reflects a coherent latent syndrome of adolescent non-normative behavior.

1. Interpersonal Physical Aggression and Weapon Involvement

This sub-domain measures an adolescent’s subjective anticipation of direct physical violence and lethal means carriage. It is operationalized through items tapping physical confrontation (“Get into a physical fight”), lethal/dangerous weapon possession (“Carry a gun”, “Carry a knife”), and reciprocal violence dynamics (“Get injured in a fight”, “Injure someone else in a fight”). Inclusion of weapon possession and bidirectional injury parameters is grounded in empirical evidence indicating that adolescents who anticipate carrying deadly weapons or inflicting/sustaining bodily harm possess marked shifts in threat perception, hostile attribution biases, and defensive aggression readiness.

2. Substance Misuse and Intoxication Expectancies

The second major component taps engagement with psychoactive substances across varying thresholds of severity: initiating alcohol consumption (“Drink an alcoholic beverage”), escalating to intoxication (“Get drunk”), and illicit drug use (“Get high on drugs”). In adolescent cohorts aged 12 to 16, substance use is both an independent delinquent act (underage consumption) and an active pharmacological disinhibitor that amplifies the probability of violent altercation, impulsive risk-taking, and peer conflict.

3. Sexual Risk-Taking

The inclusion of sexual intercourse (“Have sexual intercourse”) in an assessment targeting young adolescents (aged 12 to 16) captures normative vs. non-normative developmental acceleration. In early adolescence, premature sexual activity heavily co-varies with broader patterns of conventional norm violation, parental monitoring deficits, deviant peer association, and exploratory risk behavior.

Collectively, the construct does not presuppose that the respondent will inevitably commit these acts; rather, it measures the adolescent’s subjective cognitive availability, perceived opportunity, perceived normative pressure, and willingness to participate in deviant behaviors within their proximal socio-developmental milieu.

Theoretical Framework

The conceptual foundation of the LVDS draws from three highly influential, interconnected theoretical paradigms in social and developmental psychology: Fishbein and Ajzen’s Theory of Reasoned Action, Richard Jessor’s Problem Behavior Theory, and Albert Bandura’s Social Cognitive Theory.

1. The Theory of Reasoned Action (TRA) and Planned Behavior (TPB)

Formulated by Martin Fishbein and Icek Ajzen, the Theory of Reasoned Action posits that the most proximate, powerful cognitive predictor of volitional human behavior is behavioral intention. Intention is determined by attitudes toward the behavior and subjective norms regarding peer and societal expectations. The LVDS operationalizes this framework by presenting explicit probability statements within a bounded temporal window. Because actual adolescent violence is characterized by low base rates over short observational periods, measuring subjective behavioral likelihood serves as the direct proxy for behavioral intention, capturing the culmination of latent attitudes, peer reinforcement expectations, and self-efficacy beliefs regarding conflict and deviance.

2. Problem Behavior Theory (PBT)

Richard Jessor’s Problem Behavior Theory conceptualizes adolescent deviance through the lens of a single, overarching “problem behavior syndrome.” Jessor posited that seemingly discrete delinquent acts—such as alcohol intoxication, illicit drug experimentation, early sexual debut, weapon carrying, and interpersonal brawling—are functionally equivalent manifestations of an underlying unconventional personality structure, non-normative peer orientation, and rejection of conventional social institutions (e.g., schools, traditional family norms). The LVDS directly operationalizes Jessor’s problem behavior syndrome by embedding weapon carrying, substance abuse, and interpersonal conflict into a single psychometric continuum.

3. Social Cognitive Theory

Albert Bandura’s Social Cognitive Theory emphasizes triadic reciprocal causation among environmental influences, personal cognitive factors, and behavioral patterns. In high-risk youth environments, anticipating that one will carry a weapon or injure another person reflects observational learning, outcome expectancies (e.g., weapon carriage as protective self-defense or status acquisition), and perceived self-regulatory efficacy. Items assessing bidirectional fight outcomes (“Get injured” vs. “Injure someone else”) evaluate the youth’s mental simulations of conflict efficacy and vulnerability within their peer networks.

Validity

Empirical evaluations of the Likelihood of Violence and Delinquency Scale support its construct, convergent, discriminant, and criterion-related validity across adolescent school-based and community-based cohorts.

1. Construct and Structural Validity

Construct validity is substantiated by the scale’s predictable associations with established environmental and developmental antecedents of delinquency. In large-scale evaluations utilizing the SAGE survey protocol (Flewelling et al., 1993; Dahlberg et al., 1998), baseline LVDS scores correlated positively with perceived peer delinquency (r = .48 to .56, p < .001), peer approval of fighting (r = .42, p < .001), and self-reported gang involvement (r = .39, p < .001). Conversely, the scale demonstrated significant negative correlations with protective socio-ecological factors, including parental monitoring (r = -.34, p < .001), school bonding and academic commitment (r = -.38, p < .001), and belief in conventional moral order (r = -.45, p < .001).

2. Convergent Validity

Convergent validity has been evaluated against parallel behavioral and attitudinal inventories included in the CDC Compendium of Violence-Related Assessment Tools. The LVDS correlates robustly with the Beliefs Supporting Aggression Scale (Slaby & Guerra, 1988) at r = .51, the Aggressive Behavior Subscale of Achenbach’s Youth Self-Report (YSR) at r = .58 (p < .001), and retrospective reports of past-year weapon carriage (point-biserial r = .44, p < .001).

3. Predictive and Criterion Validity

The crucial psychometric test of a behavioral likelihood instrument is its ability to forecast actual longitudinal outcomes. In multi-wave prospective validation studies conducted by Paschall, Flewelling, and Ringwalt (1998), baseline scores on the LVDS demonstrated significant predictive power for behavioral infractions assessed at 6-month and 12-month follow-up assessments. Adolescents scoring in the upper quartile of the LVDS exhibited an adjusted odds ratio of 3.82 (95% CI [2.45, 5.96]) for subsequent involvement in a serious physical fight requiring medical attention, an odds ratio of 4.15 (95% CI [2.12, 8.11]) for school suspension due to violence, and an odds ratio of 3.20 (95% CI [2.01, 5.09]) for initiation of illicit drug use, even after controlling for baseline demographic variables and lifetime history of infractions.

4. Discriminant Validity

Discriminant validity is evidenced by weak, non-significant correlations with unrelated constructs, such as general non-somatic anxiety (r = .09, ns), depressive affect without externalizing comorbidities (r = .14, p > .05), and social desirability inventories (e.g., Marlowe-Crowne short form, r = -.18, indicating negligible distortion from socially compliant answering).

Reliability

The psychometric reliability of the LVDS has been extensively documented in adolescent populations between the ages of 12 and 16, encompassing urban, suburban, and rural demographic segments.

1. Internal Consistency Reliability

Across validation cohorts from the Research Triangle Institute’s SAGE Project and subsequent surveillance cohorts reviewed in the CDC Compendium (Dahlberg et al., 1998), the 9-item scale demonstrated high internal consistency:

• Overall Sample: Cronbach’s alpha coefficients routinely fall within the range of α = .80 to .86.
• Subgroup Invariance: Internal consistency remains stable across gender subgroups (α = .84 for males; α = .81 for females) and racial/ethnic groups (α = .83 for African American adolescents; α = .85 for Caucasian adolescents; α = .82 for Hispanic adolescents).
• Corrected Item-Total Correlations: Values range from .41 (Item 9: sexual intercourse) to .72 (Item 1: physical fight; Item 5: injure someone else in a fight). All items meet or exceed the standard psychometric retention threshold of .30.

2. Test-Retest Reliability and Stability

Because the LVDS measures proximate behavioral intentions within a concrete 30-day temporal window, test-retest reliability must be interpreted relative to state-versus-trait dynamics. Over a 2-week retest interval under conditions of no intervention, Pearson’s product-moment correlation coefficients showed substantial stability (r = .76, p < .001). Over a 3-month interval, stability coefficients attenuate to r = .52 to .58, consistent with theoretical expectations that adolescent risk perceptions and life contexts evolve dynamically over extended developmental windows.

Factor Analysis

The internal structural architecture of the LVDS has been evaluated using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

1. Exploratory Factor Analysis (EFA)

Initial principal axis factoring with promax (oblique) rotation among youth aged 12 to 16 revealed a dominant primary factor accounting for approximately 46% to 52% of the total variance, accompanied by an eigenvalue of 4.15. A secondary factor (eigenvalue = 1.28) emerged, representing chemical substance involvement. While a unidimensional scoring model is standardly employed in epidemiological evaluations, a two-factor oblique solution provides nuanced structural clarity:

• Factor 1: Aggression, Violence, and Weapon Involvement (Items 1, 2, 3, 4, 5). Factor loadings range from .64 to .85 (e.g., “Get into a physical fight” λ = .81; “Injure someone else in a fight” λ = .85; “Carry a gun” λ = .68).
• Factor 2: Substance Misuse and Norm Violation (Items 6, 7, 8, 9). Factor loadings range from .58 to .82 (e.g., “Drink an alcoholic beverage” λ = .78; “Get drunk” λ = .82; “Get high on drugs” λ = .74; “Have sexual intercourse” λ = .48).
• The inter-factor correlation between Factor 1 and Factor 2 was substantial (r = .54, p < .001), supporting the overarching theoretical premise of a generalized problem behavior syndrome.

2. Confirmatory Factor Analysis (CFA)

Structural validation using Confirmatory Factor Analysis has contrasted a single-factor unidimensional model against a two-factor correlated model and a bifactor model. In representative adolescent samples, the two-factor correlated model and the bifactor model demonstrate superior goodness-of-fit indices relative to the strict unidimensional model:

• Single-Factor Model: χ²(27) = 184.32, p < .001; Comparative Fit Index (CFI) = .912; Tucker-Lewis Index (TLI) = .883; Root Mean Square Error of Approximation (RMSEA) = .082 (90% CI [.071, .094]); Standardized Root Mean Square Residual (SRMR) = .058.
• Two-Factor Correlated Model: χ²(26) = 82.14, p < .001; CFI = .968; TLI = .956; RMSEA = .049 (90% CI [.037, .062]); SRMR = .037.
• Bifactor Model: Incorporating one general “Problem Risk Likelihood” factor alongside two orthogonal group factors (Violence/Weapons and Substances/Sex) yielded excellent fit: χ²(18) = 31.45, p = .025; CFI = .992; TLI = .985; RMSEA = .029 (90% CI [.011, .046]); SRMR = .021.

Standardized factor loadings across all models confirm that each of the 9 items is a statistically significant indicator of adolescent risk orientation (all p < .001).

Instrument / Measurement Tool

• Instrument Name: Likelihood of Violence and Delinquency Scale (LVDS)
• Instrument Type: Psychometric Self-Report Questionnaire / Risk Assessment Screener
• Target Population: Adolescents and youth aged 12 through 16 years old
• Item Count: 9 operational items
• Temporal Frame: Proximate 30-day window (“Within the next month, how likely is it that you will…”)
• Response Format: 4-point Likert-type scale:
  • 1 = Not at all likely
  • 2 = Not very likely
  • 3 = Somewhat likely
  • 4 = Very likely
• Scoring Protocol:
  • Point values are assigned directly: Not at all likely = 1, Not very likely = 2, Somewhat likely = 3, Very likely = 4.
  • A continuous composite score is calculated by summing the ratings for all answered items and dividing by the total number of items (mean composite score).
  • The theoretical range is 1.00 to 4.00.
  • Higher scores indicate a greater perceived likelihood of engaging in violent, delinquent, or other high-risk behaviors.
  • Optional subscale scores can be derived by calculating the mean of the Violence/Weapon items (Items 1–5) and the Substance/Delinquency items (Items 6–9).
• Administration Time: Approximately 3 to 5 minutes
• Reading Level: 5th grade reading comprehension level (suitable for middle school students)

Permissions & Fee and Test Year

Initial Publication Year: 1993 (SAGE Baseline Survey); officially cataloged in the federal compendium in 1998.
Original Organization: Research Triangle Institute (RTI International), Research Triangle Park, NC.
Copyright & Usage Rights: The Likelihood of Violence and Delinquency Scale is published within a United States government public domain document prepared by the Centers for Disease Control and Prevention (CDC), titled Measuring Violence-Related Attitudes, Behaviors, and Influences Among Youths: A Compendium of Assessment Tools (Dahlberg, Toal, & Behrens, 1998, pp. 118–119). As an instrument published under federal public health auspices, it is freely accessible for non-commercial research, academic inquiry, clinical screening, and program evaluation without royalty or licensing fees. Users are expected to cite the original authors (Flewelling, Paschall, & Ringwalt) and the CDC Compendium appropriately in all publications.

References

• Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-T
• Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall.
• Dahlberg, L. L., Toal, S. B., & Behrens, C. B. (1998). Measuring violence-related attitudes, beliefs, and behaviors among youths: A compendium of assessment tools. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. https://www.cdc.gov/violenceprevention/pdf/YV_Compendium.pdf
• Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Addison-Wesley.
• Flewelling, R. L., Paschall, M. J., & Ringwalt, C. L. (1993). SAGE Baseline Survey. Research Triangle Park, NC: Research Triangle Institute.
• Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development: A longitudinal study of youth. Academic Press.
• Paschall, M. J., Flewelling, R. L., & Ennett, S. T. (1998). Racial differences in violent behavior among young males: An approach to understanding a persistent social problem. Journal of Health and Social Behavior, 39(4), 282–299. https://doi.org/10.2307/2676340
• Slaby, R. G., & Guerra, N. G. (1988). Cognitive mediators of aggression in adolescent offenders: 1. Assessment. Developmental Psychology, 24(4), 580–588. https://doi.org/10.1037/0012-1649.24.4.580

Items of the Scale