Verbal Ability and Persistent Offending: A Race-specific Test of Moffitt’s Theory

Abstract

Theoretical questions linger over the applicability of the verbal ability model to African-Americans and the social control theory hypothesis that educational failure mediates the effect of verbal ability on offending patterns. Accordingly, this paper investigates whether verbal ability distinguishes between offending groups within the context of Moffitt’s developmental taxonomy. Questions are addressed with longitudinal data spanning childhood through young-adulthood from an ongoing national panel, and multinomial and hierarchical Poisson models (overdispersed). In multinomial models, low verbal ability predicts membership in a life-course-persistent-oriented group relative to an adolescent-limited-oriented group. Hierarchical models indicate that verbal ability is associated with arrest outcomes among White and African-American subjects, with effects consistently operating through educational attainment (high school dropout). The results support Moffitt’s hypothesis that verbal deficits distinguish adolescent-limited- and life-course-persistent-oriented groups within race, as well as the social control model of verbal ability.

Keywords:

• Moffitt
• adolescent limited
• life course persistent
• developmental
• verbal ability

Funding

This paper was supported by grant [#5 R03 DA15717-02] from the National Institute of Drug Abuse (NIH). The views contained here reflect those of the authors.

Notes

¹ Some IQ–delinquency scholars posit that, within offender populations, the effect of verbal ability may shift when outcome variables distinguish overt (violent) from covert (property) offenses. For instance, Walsh (1987) reports the familiar relationship between low verbal ability and violence in his study of male probationers. In contrast, probationers with the highest verbal abilities engaged in more serious property offending (also see Barker et al., 2011).

² Later onset delinquency is more typically associated with deception/property offending (i.e. covert) than would be the case for early-onset delinquency, with the latter reflecting a more general pattern of serious offending that includes covert and overt (i.e. violent) behavior. For a more detailed discussion of these issues see Le Blanc and Loeber (1993), Loeber and Hay (1994), Moffitt (2006), and Patterson, DeBaryshe, and Ramsey (1989).

³ In the original statement of the theory, Moffitt anticipated a third typology comprised of abstainers who refrained from delinquency altogether because of their social exclusion. As this typology is not part of the offending population, we do not elaborate on them further.

⁴ The subsample analyzed in this study is not unusual or unrepresentative in ways that bias the analysis in favor of Moffitt’s theory. Comparison of study variables in wave 1 between the sample analyzed here (analysis sample) and the White and African-American males that left the sample by wave 14 due to attrition (attrition sample) reveals no differences (p < .05) in ever or frequency of arrest, family income, test motivation, and high school completion. Small but significant differences were found in peer drug use (attrition sample slightly greater) and verbal ability (attrition sample slightly lower). All else equal, the analysis may therefore slightly underestimate the size of verbal ability effects due to those differences. Relative to the entire NLSY97 sample minus White and African-American males (exclusion sample), the analysis sample, as expected, is almost twice as likely to be arrested with over twice the frequency of arrest. The exclusion sample is also slightly less affluent and exhibits slightly lower verbal ability, but greater peer drug use and high school dropout with no difference in test motivation.

⁵ Differential validity of delinquency data by race is an unresolved issue (Piquero, Schubert, & Brame, 2014). Previous research on the validity of self-reports examines criterion validity by examining the association between self-reported delinquency and either official or self-reported official delinquency. Two prominent studies are illustrative. Hindelang et al. (1981) report that Black male self-reports exhibit lower validity relative to White males. Twenty-five years later, Farrington, Loeber, Stouthamer-Loeber, Van Kamman, and Schmidt (1996) examined the issue with data drawn from the Pittsburgh Youth Study. In contrast to Hindelang et al.’s findings, Farrington et al. report that black adolescents with police records for criminal delinquency and property and violent index offenses are more likely to report being picked up by the police than similarly charged whites. They conclude “ethnic differences in official delinquency … were not attributable to differential … ethnic validity of measures of delinquent behavior” (p. 511).

⁶ For a detailed description of the ASVAB administration in NLSY97 see Appendix 10 of the NLSY97 codebook supplement.

⁷ Final ability estimates of the word knowledge and paragraph comprehension subtests are used to calculate the verbal ability measure rather than raw scores because computer-adaptive testing was used. The method entails tailoring the difficulty of questions based on each respondent’s correct or incorrect responses to previous questions. Thus, respondents did not answer the same number of questions and the questions asked of each were of varying difficulty, both of which confound use of raw scores. The final ability estimates, created by the Department of Defense using item response theory, are appropriate for comparing verbal ability across respondents.

⁸ Use of the bottom quartile to indicate risk is common in the developmental/life course criminology literature, as well as more general research on the risk factors associated with antisocial behavior (see Farrington & Loeber, 2000).

⁹ When the strict definition of the LCP group is used (i.e. n = 49) the pattern of results is nearly identical. The key difference is that the coefficient for low verbal ability reduces with a slight increase in standard error, which changes statistical significance from p < .05 to p < .10. Other coefficients are virtually identical.

¹⁰ For example, at level-1 we model: η_ij = log(λ_ij), where λ_ij is the event rate reflecting the frequency of delinquency and η_ij is the log of the event rate. Note that while λ_ij is constrained to be non-negative, log(λ_ij) can take on any value. The predicted log event rate can be converted to an event rate by generating λ_ij = exponential{η_ij}.

¹¹ The models in Table 4 are estimated with HLM.

¹² The measures of educational attainment used in the analysis were contrasted with two alternative school measures in the models—grade point average and educational expectations. The alternative measures are less consistently associated with the arrest outcome and do not mediate the entire effect of verbal ability on arrests.

¹³ Unlike African-American and White subjects, there is substantial missing data on the verbal ability measure among Hispanics in the NLSY97 which, unfortunately, limits our ability to analyze that group.

Additional information

Notes on contributors

Paul E. Bellair

Paul Bellair (PhD, University at Albany, State University of New York) is a professor in the Department of Sociology at The Ohio State University. His current research tests theoretical explanations for race differences in violence and recidivism, tests theory pertaining to the effect of verbal ability on delinquency, and examines the social networks and employment experiences of prisoners.

Thomas L. McNulty

Thomas McNulty is an associate professor of Sociology at the University of Georgia. His specialty areas include criminology, urban sociology, and research methods. His most recent work focuses on developing and testing multilevel theoretical models of racial and ethnic differences in crime/violence, with emphasis on the role of individual differences within the context of family, school, and neighborhood environments.

Alex R. Piquero

Alex Piquero is Ashbel Smith professor of Criminology at the University of Texas at Dallas, adjunct professor, Key Centre for Ethics, Law, Justice, and Governance, Griffith University Australia, faculty affiliate, Center for Violence and Injury Prevention George Warren Brown School of Social Work Washington University in St. Louis, and from 2009–2013 he was Co-Editor, Journal of Quantitative Criminology. His research interests include criminal careers, criminological theory, and quantitative research methods. He has received several research, teaching, and mentoring awards and is a fellow of both the American Society of Criminology and the Academy of Criminal Justice Sciences.