Comprehensive Misconduct Inventory (CMI – 50)

1. Abstract

The Comprehensive Misconduct Inventory (CMI), commonly cataloged in psychometric literature as the CMI-50 and its extended 58-item iteration (CMI-58), is a multidimensional self-report assessment developed by Delroy L. Paulhus and colleagues at the University of British Columbia. Designed to capture a wide continuum of non-normative, delinquent, antisocial, and deviant behaviors within non-forensic, community, and collegiate populations, the inventory bridges the gap between everyday subclinical transgressions and severe criminal actions. The scale systematically measures transgressive behaviors across two temporal frames: retrospective lifetime delinquency during high school years (Items 1–43) and concurrent transgressive behavior occurring during an average month in adulthood (Items 44–58).

Empirical factoring typically yields five to seven primary behavioral clusters: Soft Drug Abuse (SDA), Hard Drug Abuse (HDA), Minor Criminality (MC), Serious Criminality (SC), Driving Misbehavior (DM), Bullying/Harassing (BH), and Anti-Authority Misbehavior (AA). Methodologically, the scale employs an open-ended or bounded behavioral frequency count response format. Because raw frequency counts across behaviors with radically disparate base rates (such as speeding versus armed assault) generate severe positive skew, the psychometric protocol mandates standardizing item raw scores (via z-score transformations) prior to computing unit-weighted composite subscales and composite indices (e.g., Overall Misbehavior, General Substance Abuse, General Criminality). Psychometric evaluations demonstrate moderate to high internal consistency across subscales (Cronbach’s α ranging from .68 to .89), solid temporal stability, and robust construct and criterion validity. In particular, CMI scores correlate strongly with subclinical psychopathy measured by the Self-Report Psychopathy Scale (SRP-II / SRP-4), making it a gold-standard criterion variable in studies investigating the Dark Triad of personality.

2. Keywords

Comprehensive Misconduct Inventory, CMI-50, CMI-58, Delroy L. Paulhus, Self-Report Psychopathy, antisocial behavior, deviance, Dark Triad, behavioral frequency, delinquency, driving misconduct, substance abuse.

3. Authors

The primary architect and intellectual developer of the Comprehensive Misconduct Inventory is Delroy L. Paulhus, Ph.D., Emeritus Professor of Psychology at the University of British Columbia (UBC), Vancouver, Canada. Dr. Paulhus is internationally recognized for his seminal contributions to personality assessment, social desirability responding, self-deception, and the conceptualization of subclinical dark personalities (notably co-originating the Dark Triad framework alongside Kevin M. Williams). His laboratory, the Paulhus Personality Laboratory at UBC, has developed numerous standardized measures widely cited across differential, personality, and social psychology.

Key collaborators in the psychometric development, empirical validation, and conference dissemination of the CMI include:

• Kevin M. Williams, Ph.D. — Former doctoral researcher at the University of British Columbia, whose collaborative work with Paulhus and Hare established the structural linkages between four-facet psychopathy models and behavioral misconduct inventories.
• Robert D. Hare, Ph.D. — Emeritus Professor of Psychology at the University of British Columbia, developer of the Psychopathy Checklist-Revised (PCL-R) and the Youth Version (PCL:YV), who co-authored the empirical foundations demonstrating the predictive utility of psychopathy measures on CMI self-reported delinquency.
• A. McAndrew, T. Learn, and Peter D. Harms, Ph.D. — Research affiliates who contributed to the initial empirical presentations of the CMI in connection with antisocial behavior patterns and entertainment preferences at the American Psychological Association (APA).

Inquiries regarding academic research use and laboratory archives are typically directed to the Department of Psychology, University of British Columbia, 2136 West Mall, Vancouver, BC, Canada V6T 1Z4.

4. Purpose

The primary purpose of the Comprehensive Misconduct Inventory (CMI) is to provide an objective, nuanced, and psychometrically robust self-report measure of antisocial and non-normative behavior suitable for non-institutionalized, subclinical populations. Traditional forensic assessments, such as the Psychopathy Checklist-Revised (PCL-R), were designed predominantly for incarcerated offenders or forensic psychiatric inpatients. When applied to university students or normal community adults, institutional delinquency measures suffer from severe floor effects, as the base rates of overt violent felony, institutional infraction, and parole revocation are nearly zero in non-clinical cohorts. Conversely, common adolescent delinquency scales frequently focus solely on minor high school truant behaviors, failing to measure adult transgressions, vehicular aggression, cyber-harassment, or serious covert illegal acts.

The CMI resolves this measurement dilemma by compiling a wide 50-to-58-item spectrum of deviant acts ranging from universally acknowledged mild indiscretions (e.g., sneaking into a movie theater without paying, swearing at adults, copying a classmate's homework, parking illegally) to moderate offenses (e.g., underage drinking, driving while impaired, shoplifting, workplace or academic plagiarism) to extreme, felony-level infractions (e.g., commercial drug trafficking, armed assault, auto theft, break-and-enter, forced sexual contact). The theoretical and empirical utility of the CMI extends across multiple domains:

• Validation Criterion for Dark Personality Traits: The CMI serves as an external, objective behavioral benchmark to evaluate whether personality constructs—specifically the Dark Triad (subclinical psychopathy, narcissism, and Machiavellianism), trait sadism, and low agreeableness—translate into real-world disruptive actions rather than merely reflecting inflated self-perception or cognitive attitudes.
• Differentiation of Temporal Trajectories: By separating adolescent/high school deviance (Items 1–43) from adult/current misconduct (Items 44–58), the tool allows developmental criminologists and psychologists to operationalize the distinction between adolescence-limited delinquency and life-course persistent antisocial behavior.
• Multifaceted Deviance Profiling: Rather than treating antisociality as a monolithic entity, the inventory yields differentiated scores across vehicular recklessness, peer bullying, authority defiance, illicit drug usage, and property/violent crime.
• Forensic and Industrial Screening Research: In organizational psychology, variants of the CMI are employed in simulated integrity testing to study counterproductive work behaviors (CWB) and rule-breaking propensities.

5. Psychological Construct

The psychological construct underlying the Comprehensive Misconduct Inventory is multidimensional antisocial deviance. In psychometric and behavioral science, deviance represents actions that violate established legal codes, institutional regulations, or normative moral standards within a given socio-cultural context. The CMI operationalizes this broad construct into seven distinct, intercorrelated behavioral dimensions, which can also be aggregated into higher-order domains:

1. Soft Drug Abuse (SDA)

This dimension quantifies the recreational consumption and misuse of socially prevalent, lower-tier psychoactive substances. Measured items evaluate frequency of cannabis/hashish consumption, underage procurement of alcoholic beverages, heavy binge drinking (“getting totally drunk” or consuming multiple alcoholic beverages in an average sitting), and the misuse of mild prescription sedatives or analgesics. SDA captures high sensation-seeking and normative peer-group experimentation that may not necessarily signify deep criminal entrenchment.

2. Hard Drug Abuse (HDA)

In contrast to SDA, Hard Drug Abuse reflects engagement with high-potency, heavily criminalized illicit substances characterized by severe addictive potential and neurochemical toxicity. Items track the historical and current ingestion of cocaine, heroin, hallucinogens (LSD, mescaline, peyote), methylenedioxymethamphetamine (MDMA/“Ecstasy”), and non-prescribed amphetamines or barbiturates (“uppers and downers”). Elevating on HDA denotes severe behavioral disinhibition, elevated risk-taking, and active participation in illicit drug networks.

3. Minor Criminality (MC)

Minor Criminality assesses covert and overt property and rule infractions that violate criminal or municipal codes without involving direct physical violence against persons. Representative items include retail shoplifting, theft of personal items at school or social gatherings, buying or selling stolen goods, illegal transit evasion, academic plagiarism, cheating on examinations, and misrepresenting one's age. MC captures pervasive opportunistic dishonesty, rule-bending, and low conscientiousness.

4. Serious Criminality (SC)

Serious Criminality captures severe, high-stakes offenses that threaten the physical safety of others, violate fundamental societal taboos, or incur substantial criminal liability. The subscale includes purposeful destruction of property (vandalism), burglary (breaking into structures to steal or destroy), armed assault, gang affiliation, commercial distribution of cannabis or hard drugs, grand theft auto (joyriding without permission), weapons possession (carrying a concealed weapon), and sexual coercion (using physical force to compel sexual contact). SC is the primary behavioral marker of the callous-unemotional and chronically antisocial core of psychopathy.

5. Driving Misbehavior (DM)

This subscale measures dangerous, aggressive, and non-compliant behaviors behind the wheel of a motor vehicle. Behaviors include driving under the influence of alcohol or narcotics, street racing, driving 20 miles per hour over the legal speed limit, aggressive tailgating, honking out of anger, verbal abuse toward other motorists, and chronic illegal parking. A dedicated exposure item (Item 58: frequency of driving a motor vehicle) allows researchers to statistically control for driving opportunity, isolating genuine vehicular hostility from base mileage exposure.

6. Bullying and Harassing (BH)

Bullying/Harassing reflects interpersonal malice, physical intimidation, social cruelty, and emotional degradation directed toward peers, acquaintances, or vulnerable individuals. Items capture unprovoked physical aggression (shoving, slapping, kicking), inflicting physical injury, targeted telephonic or electronic harassment, verbally cruel remarks, and public ridicule of individuals with vulnerabilities or differences. This subscale strongly reflects everyday sadism and interpersonal antagonism.

7. Anti-Authority Misbehavior (AA)

Anti-Authority Misbehavior focuses on deliberate non-compliance, defiance, and hostility directed specifically toward authority figures and structural institutions. Behaviors include verbal insubordination toward parents, educators, retail personnel, and law enforcement; running away from home; school suspensions; unauthorized nocturnal exits (sneaking out at night); and engaging in disruptive acts specifically designed to provoke distress or anger in figures of authority.

6. Theoretical Framework

The Comprehensive Misconduct Inventory is theoretically rooted in several converging paradigms within personality, developmental criminology, and psychopathology:

The Dual-Domain and Four-Facet Models of Psychopathy

The foundational theoretical framework for the CMI is Robert Hare’s and Delroy Paulhus’s structural modeling of psychopathy. Traditional formulations, such as the two-factor model of the Psychopathy Checklist-Revised (PCL-R) and Self-Report Psychopathy Scale (SRP-II), bifurcate the condition into Factor 1 (Interpersonal/Affective: superficial charm, grandiosity, pathological lying, lack of empathy, callousness) and Factor 2 (Social Deviance/Antisocial: impulsivity, irresponsibility, early behavioral problems, adolescent delinquency, adult criminal versatility).

Subsequent psychometric work expanded this into a four-facet architecture: Interpersonal Manipulation, Callous Affect, Erratic Lifestyle, and Antisocial Behavior (Williams, Paulhus, & Hare, 2007). The CMI was explicitly leveraged to provide an expansive, unconstrained behavioral metric capable of testing whether the affective-interpersonal core of psychopathy uniquely predicts covert, overt, and aggressive delinquency, or whether such misconduct is merely the consequence of erratic impulsivity. Empirical studies utilizing the CMI established that Factor 2 (Erratic Lifestyle / Antisocial Facet) displays the strongest raw correlation with CMI total and subscale scores, but the Callous Affect facet is uniquely essential for predicting interpersonal cruelty (Bullying/Harassing) and Serious Criminality.

Moffitt’s Developmental Taxonomy of Antisocial Behavior

The temporal architecture of the CMI (differentiating high school actions from current adult frequencies) directly aligns with Terrie Moffitt’s developmental taxonomy. Moffitt posits two distinct etiologies of offending:

1. Adolescence-Limited (AL) Delinquency: Highly prevalent, peer-influenced antisocial behavior emerging during puberty as an attempt to bridge the maturity gap, which typically desists upon entering adult social roles.
2. Life-Course-Persistent (LCP) Delinquency: A continuous trajectory originating in early neurodevelopmental deficits, characterized by early conduct problems, continuous antisocial manifestations across settings, and persistent criminal conduct throughout adulthood.

By comparing an individual’s High School Misbehavior Index (Items 1–43) against their Current Misbehavior Index (Items 44–57), psychometricians can empirically differentiate between normative adolescent experimentation and persistent, trait-driven antisocial lifestyles.

Gottfredson and Hirschi’s General Theory of Crime

The CMI also operationalizes Gottfredson and Hirschi’s (1990) General Theory of Crime, which posits that all forms of delinquency, criminal acts, and analogous non-criminal behaviors (such as reckless driving, substance abuse, and interpersonal conflict) share a single underlying psychological cause: low self-control. The inclusion of diverse non-felonious, impulsive behaviors alongside statutory crimes tests whether varied manifestations of risk-seeking and immediate gratification cohere into an overarching latent deviance factor.

7. Validity

Extensive empirical investigations across university cohorts, web-based community samples, and non-forensic male and female adults have substantiated the construct, convergent, discriminant, and predictive validity of the CMI.

Construct and Factorial Validity

Construct validity is evidenced by the clear multidimensional separation of distinct classes of antisocial activity. Oblique factor analytic solutions confirm that while subscales share significant latent variance (reflecting a general antisocial orientation), behavioral manifestations partition distinctly into drug abuse, vehicular recklessness, peer aggression, and institutional criminality. Confirmatory structural equation modeling shows that a hierarchical model—with a higher-order General Deviance factor subsuming the primary behavioral factors—fits the data adequately (CFI > .90, RMSEA ≤ .06).

Convergent Validity

The convergent validity of the CMI has been established via robust correlations with established self-report personality and delinquency inventories:

• Self-Report Psychopathy Scale (SRP-II / SRP-4): CMI subscales demonstrate substantial convergent correlations with the SRP total score (typically r = .50 to .65, p < .001). Serious Criminality, Anti-Authority Misbehavior, and Bullying exhibit the strongest associations with the Antisocial and Callous Affect subscales of the SRP.
• Levenson Self-Report Psychopathy Scale (LSRP): Positive correlations are consistently observed between the CMI total score and both Primary Psychopathy (callousness/manipulation; r ≈ .38–.48) and Secondary Psychopathy (impulsivity/disinhibition; r ≈ .52–.62).
• The Dark Triad (Dirty Dozen / SD3): In multivariate analyses, subclinical psychopathy consistently emerges as the strongest predictor across all CMI subscales (β ranging from .35 to .55), followed by Machiavellianism (β ≈ .15–.25 for covert fraud and theft). Narcissism correlates primarily with driving aggression and status-seeking anti-authority displays.
• Five-Factor Model (FFM): The CMI correlates strongly and negatively with Agreeableness (r = -.40 to -.55) and Conscientiousness (r = -.35 to -.50), while exhibiting moderate positive correlations with Neuroticism/Hostility and the Excitement-Seeking facet of Extraversion.

Discriminant and Incremental Validity

The CMI demonstrates discriminant validity against internalizing psychopathology, such as generalized anxiety, major depressive symptoms, and social phobia (correlations typically non-significant or near zero, |r| < .12). Furthermore, hierarchal regression analyses reveal that CMI scores provide incremental validity in predicting institutional disciplinary sanctions, laboratory-induced cheating, and real-world moving violations over and above standard broad-bandwidth personality measures such as the NEO-PI-R or the HEXACO inventory.

8. Reliability

The Comprehensive Misconduct Inventory exhibits sound psychometric reliability despite the substantial statistical challenges inherent in behavioral count data, such as high positive skewness and variable base rates.

Internal Consistency

Because the raw items are measured on an open frequency or bounded ordinal count scale, researchers must standardize items prior to computing reliability metrics. Internal consistency estimates across published university and community samples yield the following representative coefficients:

• Overall Misbehavior Index: Cronbach’s α ranges between .88 and .93, reflecting exceptionally high internal consistency for an omnibus behavioral deviance composite.
• Minor Criminality (MC; 10 items): Cronbach’s α = .78 – .84.
• Serious Criminality (SC; 10 items): Cronbach’s α = .76 – .82.
• Soft Drug Abuse (SDA; 8 items): Cronbach’s α = .82 – .88.
• Hard Drug Abuse (HDA; 6 items): Cronbach’s α = .79 – .86.
• Driving Misbehavior (DM; 7 items): Cronbach’s α = .74 – .81.
• Bullying / Harassing (BH; 8 items): Cronbach’s α = .75 – .83.
• Anti-Authority Misbehavior (AA; 8 items): Cronbach’s α = .72 – .79.

Temporal Stability (Test-Retest Reliability)

Because the first 43 items assess historical events occurring during a completed developmental window (high school years), retrospective scores demonstrate high temporal stability over intervals ranging from 6 weeks to 6 months (test-retest intraclass correlation coefficients [ICC] ≥ .85). For current monthly misconduct items (Items 44–57), test-retest reliability across a 4-to-8-week window remains robust (ICC = .70 – .78), reflecting persistent behavioral style while accommodating normal situational fluctuations in monthly opportunity.

9. Factor Analysis

The structural composition of the CMI was originally established through exploratory factor analysis (EFA) and subsequent confirmatory factor analysis (CFA) conducted on large non-forensic student samples at the University of British Columbia.

Exploratory Factor Analysis (EFA)

Initial factor extractions using Principal Axis Factoring and Maximum Likelihood extraction with oblique (Promax and Direct Oblimin) rotations revealed that orthogonal models are theoretically and empirically indefensible, as underlying antisocial behaviors naturally correlate with one another. Factor analyses consistently reveal either a 5-cluster or 7-subscale architecture:

• The 5-Cluster Model: Aggregates Soft and Hard Drugs into a single general Substance Abuse factor, and merges Minor and Serious Criminality into an omnibus Criminal Behavior factor, alongside Bullying, Driving Misconduct, and Anti-Authority Aggression.
• The 7-Factor Model (Williams, Paulhus, & Hare): Retains the finer distinction between Minor and Serious Criminality, and separates Soft from Hard Drug Abuse. Primary factor loadings for salient items generally exceed .40, with primary loadings on targeted factors ranging from .45 to .78, and minimal cross-loadings (< .25) on non-targeted factors.

Confirmatory Factor Analysis (CFA)

In CFA evaluations, a correlated seven-factor model exhibits superior fit over a single-factor unidimensional model. Standard goodness-of-fit indices across non-forensic adult samples demonstrate satisfactory fit:

• Comparative Fit Index (CFI) ≥ .91
• Tucker-Lewis Index (TLI) ≥ .89
• Root Mean Square Error of Approximation (RMSEA) ≤ .055 (90% CI [.048, .061])
• Standardized Root Mean Square Residual (SRMR) ≤ .062

Higher-order modeling establishes that a general “Omnibus Misconduct” superordinate construct accounts for approximately 45–55% of the common variance among the first-order factors, supporting the psychometric practice of reporting both individual subscale profiles and an aggregated general deviance index.

10. Instrument / Measurement Tool

• Instrument Designation: Comprehensive Misconduct Inventory (CMI-50 / CMI-58).
• Author: Delroy L. Paulhus, Ph.D.
• Target Population: Adolescents (older high school cohorts), university students, and adult community samples; non-institutionalized, subclinical populations.
• Administration Format: Self-administered paper-and-pencil inventory or secure computerized web assessment.
• Total Item Count: 58 items in the complete inventory (Items 1–43 retrospective high school delinquency; Items 44–57 current adult monthly frequency; Item 58 baseline driving volume control).
• Temporal Timeframes:
  • Items 1–43: Retrospective frequency: “During your high school years, how many times did you…”
  • Items 44–57: Current frequency: “These days, in an AVERAGE MONTH, how often do you do the following…”
  • Item 58: Current driving volume: “These days, in an AVERAGE MONTH, how often do you: drive a motor vehicle.”
• Response Metrics & Scoring Transformation:
  • Respondents write numerical frequency counts (or select discrete frequency bands: 0, 1–2, 3–5, 6–10, 11+ times).
  • Mandatory Standardization: Because raw frequencies have disparate base rates, raw scores must be converted to standard z-scores before aggregation.
  • Subscale Computation: Calculate the mean of the standardized (z-scored) items corresponding to each subscale.
• Subscale Composition:
  • Soft Drug Abuse (SDA): 8 items (12, 15, 18, 35, 37, 44, 45, 49)
  • Hard Drug Abuse (HDA): 6 items (16, 36, 38, 39, 40, 41)
  • Minor Criminality (MC): 10 items (3, 5, 8, 11, 21, 22, 27, 32, 34, 42)
  • Serious Criminality (SC): 10 items (1, 2, 4, 6, 9, 10, 14, 19, 24, 33)
  • Driving Misbehavior (DM): 7 items (17, 23, 25, 31, 46, 53, 55)
  • Bullying / Harassing (BH): 8 items (7, 20, 28, 30, 47, 48, 50, 51)
  • Anti-Authority Misbehavior (AA): 8 items (13, 26, 29, 43, 52, 54, 56, 57)
• Composite Indices:
  • Overall Misbehavior Index: Mean of all standardized items (Items 1–57).
  • General Substance Abuse Index: Mean of SDA and HDA standardized items.
  • General Criminality Index: Mean of MC and SC standardized items.
  • High School Misbehavior Index: Mean of standardized items 1–43.
  • Current Misbehavior Index: Mean of standardized items 44–57.
  • Driving Control Adjustment: Item 58 is utilized as a covariate in regression or ANCOVA models to statistically control for driving opportunity on the DM factor.

11. Permissions & Fee and Test Year

The Comprehensive Misconduct Inventory was developed and formalized between 2001 and 2007 through research presentations and empirical publications emerging from the Paulhus Personality Laboratory at the University of British Columbia (e.g., Williams, McAndrew, Learn, Harms, & Paulhus, 2001; Williams, Paulhus, & Hare, 2007). The instrument was placed in the public academic research domain for scientific, educational, and non-commercial investigative inquiries.

There are no licensing fees or royalty payments required for recognized academic researchers, graduate students, or university-affiliated scholars utilizing the CMI for empirical non-commercial investigations. However, formal attribution and citation of the primary developers (Paulhus, Williams, and Hare) is strictly mandatory in all scientific manuscripts, conference presentations, and technical reports. Parties seeking commercial licensing, organizational personnel screening, or integration into proprietary psychometric platforms must secure explicit written permission from the copyright holder, Delroy L. Paulhus.

12. References

• Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford University Press.
• Hare, R. D. (2003). The Hare Psychopathy Checklist—Revised (PCL-R) (2nd ed.). Multi-Health Systems. https://doi.org/10.1037/t03478-000
• Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100(4), 674–701. https://doi.org/10.1037/0033-295X.100.4.674
• Paulhus, D. L., & Williams, K. M. (2002). The Dark Triad of personality: Narcissism, Machiavellianism, and psychopathy. Journal of Research in Personality, 36(6), 556–563. https://doi.org/10.1016/S0092-6566(02)00505-6
• Williams, K., McAndrew, A., Learn, T., Harms, P., & Paulhus, D. L. (2001, August). Dark personalities: Antisocial behavior and entertainment preferences [Paper presentation]. Annual Meeting of the American Psychological Association, San Francisco, CA, United States.
• Williams, K. M., Paulhus, D. L., & Hare, R. D. (2007). Capturing the four-facet structure of psychopathy in nonforensic samples: The Self-Report Psychopathy scale. Journal of Personality Assessment, 88(2), 118–129. https://doi.org/10.1080/00223890701267940

13. Items of the Scale