Distinguishing between early and late onset delinquents: race, income, verbal intelligence and impulsivity.
There have been several attempts to test Moffitt's predictions about the relationship between impulsivity and early onset antisocial behavior. White, Moffitt, Caspi, Bartusch, Needles, and Stouthamer-Loeber (1994) categorize the two approaches to measuring impulsivity as cognitive (using neuropsychological tests) and behavioral (using various observer or self reports of impulsive behavior). They found that behavioral impulsivity is more strongly related to delinquent behavior for males at ages 10 to 13 than is cognitive impulsivity. They did not specifically test Moffitt's hypothesis by examining the impulsivity of late onset delinquents.
Taylor, Iacono, and McGue (2000) found both cognitive and behavioral impulsivity was greater in delinquents with an onset earlier than 11 years old than in late onset delinquents (onset 12 or later) or nondelinquent controls. Silverthorn, Frick, and Reynolds (2001) measured impulsivity using the Antisocial Process Screening Device (Frick & Hare, 2001) and found that adolescent males with an early onset of antisocial behavior showed significantly poorer impulse control than males with a later onset of delinquency. White, Bates and Buyske (2001) found that adolescent males, whose antisocial behavior persisted into adulthood, showed more behavioral impulsivity than did those whose delinquency ended after adolescence. Raine, Moffitt, Caspi, Loeber, Stouthamer-Loeber, and Lynam (2005) used cognitive measures of executive functioning/impulsivity and found no difference in the functioning of adolescents with adolescence-limited antisocial behavior and those with life course persistent antisocial behavior. However, Raine et al. (2005) did find that adolescents with life course persistent antisocial behavior had higher rates of diagnosis with attention deficit/hyperactivity disorder than did adolescents with adolescence limited antisocial behavior.
There have been few direct tests of Moffit's (1993) hypothesis regarding lower verbal intelligence in early onset delinquents. Taylor, Iacono, and McGue (2000) reported lower verbal intelligence in a sample of early onset delinquents, when compared to the later onset delinquency group and a non-delinquent control group. This study included 36 early onset delinquents, 86 late onset delinquents, and 25 nondelinquent controls. Silverthorn, et al. (2001) compared the intelligence of males with early onset delinquent behavior to males with late onset delinquent behavior and found no significant difference. However, the sample sizes were so small (11 early onset and 12 late onset) that statistical power was extremely limited. White, et al. (2001) compared the verbal abilities of adolescence-limited delinquents with the verbal abilities of adolescents whose antisocial behavior persisted through adulthood and found no significant difference. The trend, though non-significant, was for adolescence-limited delinquents to have higher verbal ability than the persistent delinquents. This study included 230 adolescence-limited delinquents and 49 delinquents whose criminal behavior persisted into adulthood. Raine et al. (2005) did find that adolescents with life course persistent antisocial behavior (n=44) had lower verbal intelligence than adolescents with adolescence limited antisocial behavior (n=68).
Moffitt (1993) indicated that, in addition to the effect of limited verbal abilities and impulsivity, negative or disadvantaged environments were important in maintaining and exacerbating behavioral problems and contributing to entrenched antisocial behavior, and identified socioeconomic status as an important predictor of life course persistent antisocial behavior. White et al. (2001) found no differences in socioeconomic status between nondelinquents, adolescence-limited delinquents, and persistent delinquents. Taylor, et al. (2000) did not address socioeconomic status in their analyses. Raine et al. (2005) did not find a difference in rates of poverty between life course persistent and adolescence limited antisocial behavior, but did find that parental neglect was more common in adolescents with life course persistent antisocial behavior.
That black adolescents are more likely to engage in delinquent behavior is well documented (e.g. Hawkins, Laub, & Lauritsen, 1998). The elevated rate of delinquency in blacks is often attributed to environmental conditions such as lower socioeconomic status and urbanization (e.g. Hawkins, Laub, & Lauritsen, 1998; Lynam et al. 1993). Lynam, et al. (1993) found that race had no relationship to delinquent behavior once intellectual ability and socioeconomic status had been controlled for. White et al. (2001) and Taylor et al. (2000) did not address the race of delinquents, because the samples were 90% and 98% white, respectively. Thus, these two studies were then essentially studies of the characteristics of white delinquents and little is known about the verbal intellectual abilities of life-course-persistent and adolescence-limited black delinquents. The Raine et al. (2005) sample was 59% black. However, they reported no analyses of the relationship of race to any of the variables of delinquency, executive functioning, verbal intelligence, or psychosocial or socioeconomic factors.
In summary, there has been some support for Moffitt's (1993) hypotheses regarding importance of impulsivity and low verbal intelligence in the development of life course persistent delinquency. However, the results have been somewhat mixed for each. Thus far, no research has indicated that socioeconomic disadvantages are more common in life course persistent delinquents than they are in adolescence-limited delinquents. No prior studies have examined the relationships among race, verbal intelligence, impulsivity, socioeconomic status and age of onset of delinquency.
The present study attempts to test several of Moffitt's (1993) hypotheses by comparing verbal intellectual ability, family income, impulsivity, and race in early and late onset delinquents. The current study uses early or late age of first arrest to approximate what Moffitt (1993) refers to as life-course persistent delinquents and adolescence-limited delinquents. This is an imperfect operationalization of these two groups. However, early onset of serious behavior problems is a key hallmark of Moffitt's taxonomy.
Operational definitions of early onset of delinquent behavior tend to involve delinquency beginning before age 11. Moffitt (1993) notes that through age 11, only 5% of the male population has shown marked antisocial behavior, but that prevalence increases substantially after that. The Diagnostic and Statistical Manual, fourth edition (DSM-IV; American Psychiatric Association, 1994) defines childhood onset conduct disorder as being a pattern of antisocial behavior emerging by age 10. Taylor, et al. (2001) identified early onset delinquents as those showing antisocial behavior by age 12. Silverthorn et al. (2001) identified adolescents with delinquent behavior at age 12 or earlier as early onset delinquents.
For the purposes of this study, adolescents who were first formally charged with an offense at or before the age of 11 were considered early onset delinquents, while adolescents who were first charged with an offense at age 15 or 16 were considered late onset delinquents. Adolescents whose first formal arrest occurred from age 12 to 14 were excluded from most analyses. This was done to reduce the risk that any adolescents were misclassified. Adolescents arrested by age 11 or earlier are clearly evidencing an early onset of serious behavioral problems. Similarly, it would be unusual, though not impossible, for an adolescent to have the severe and chronic behavior problems of the life course persistent offender and not be arrested prior to age 15. Thus, removing subjects from the sample sacrificed some statistical power in order to gain a better operationalization of Moffitt's taxonomy.
Participants were 826 male adolescents in a South Carolina Department of Juvenile Justice (SCDJJ) evaluation center. The data were collected at the evaluation center, as part of the multidisciplinary evaluation that is conducted for all juveniles who are sent to this center after they are adjudicated delinquent. The researchers utilized coded extant data without identifying information for the analyses. The juveniles who receive an order for the evaluation stay for approximately one month. An evaluation is completed and a recommendation is made to the court regarding disposition/sentencing and any services a juvenile may need. Less than one fourth of these juveniles are committed to a juvenile correctional institution following their release from the evaluation center. Offenses committed by these juveniles cover virtually the entire spectrum of juvenile crime, from status offenses to serious, violent offenses. Three hundred eighty (46.0%) of the subjects were white and 446 (54.0%) were black. Their ages ranged from 11 to 17 (M = 14.89, SD = 1.2). More than 98% of the juveniles evaluated at the evaluation center were either black or white. Therefore, the small number of adolescents of other ethnic backgrounds was excluded from this study so that these very small groups would not distort analyses involving race.
For analyses comparing early and late onset delinquents, two groups of subjects were selected; early onset offenders whose first official arrest occurred at age 11 or earlier, and late onset offenders, whose first arrest occurred at age 15 or later. Adolescents whose first formal criminal charge was between ages 12-14 were excluded from these analyses in order to reduce the frequency of possible misclassifications of early and late onset delinquents. There were 191 early onset offenders with an age of first arrest ranging from 5 to 11 (M =9.79, SD =1.39). There were 151 late onset offenders with an age of first arrest of either 15 or 16 (M = 15.30, SD =.46).
Annual family income was distributed as follows: 11% had incomes in the $0- $4,999 range; 23% had incomes in the $5,000-$9,999 range; 22% had incomes in the $10,000-$14,999 range; 12% had incomes in the $15,000-$19,999 range; 9% had incomes in the $20,000-$24,999 range; and 24% had incomes greater than $25,000.
Adolescents were administered the intelligence tests as a standard part of their psychological evaluation by master's or doctoral level psychologists. MMPI-A testing was conducted in groups of approximately twelve juveniles. A doctoral or master's level psychologist supervised testing. Juveniles were given a booklet from which they could read test items. In addition, a tape recording of the test items was played so that any juvenile who had difficulty reading could listen to the test items. The test was administered in a single two-hour administration with one to three short breaks during testing. MMPI-A scores were screened for validity such that any profiles in which L<70, F<90, K<70, VRIN <80, TRIN <80, or Cannot Say<30 were eliminated as invalid. This screening took place as part of the process of selecting the sample. Therefore, all 826 participants had valid MMPI-A profiles.
Family Income. Family income was based on a parent's report of the total annual family income, with income being recorded in 6 different ranges (0-4,999, 5,000-9,999, 10,000-14,999, 15,000-19,999, 20,00024,999 and 25,000 or greater).
Verbal Ability. Verbal intellectual ability was assessed through administration of a standardized intelligence test. In most cases this was the Verbal IQ of the Wechsler Intelligence Scale for Children, 3rd Edition (WISC-III: Wechsler, 1991). For some adolescents for whom prior test results using the WISC-III were available, only a Kaufman Brief Intelligence Test (K-BIT; Kaufman & Kaufman, 1990) was administered. In these cases an average of the prior WISC-III Verbal IQ and the current K-BIT Vocabulary standard score was used.
Impulsivity. Impulsivity was assessed using the Disinhibition/ Excitatory Potential Structural Summary Score from the MMPI-A. The Structural Summary (Archer & Krishnamurthy, 1994) was developed as a result of a factor analysis of the MMPI-A scale and subscale scores that indicated eight primary factors (Archer, Belevich, and Elkins 1994). Archer (1997) indicates that the Disinhibition/Excitatory Potential dimension reflects "adolescents' propensity to engage in impulsive and poorly controlled actions (p. 264)." In this study the Disinhibition/Excitatory Potential Score was arrived at by calculating the mean T score of all the scales and subscales on that dimension.
32% of the black delinquents were early onset offenders and 13% of white delinquents were early onset offenders. Chi-squares were computed for each race using the proportion of early and late onset offenders in the entire sample as the expected proportions. The distribution of early and late onset offenders was different from the entire sample in both the black ([chi.sup.2]=12.27, df=1, p<.001) and white ([chi.sup.2]=20.87, df=1, p<.001) groups.
Univariate ANOVA indicated that black and white delinquents differed significantly in their verbal intelligence (F(1,824)=215.0, p<.001), in their disinhibition scores (F(1,824)=4.2, p=.04) and in their family income F(1,824)=.55.6, p<.001). Table 1 shows the verbal intelligence, disinhibition scores, and family incomes for black and white delinquents. Given the large differences, particularly in verbal intelligence and family income, between black and white delinquents, as well as the large difference in proportion of adolescents that were early onset delinquents, most of the remaining analyses regarding the differences between early and late onset offenders were conducted separately for each race, in order to avoid having race confound the relationships between verbal intelligence, income, and delinquency.
Univariate ANOVA indicated that among black delinquents the verbal intelligence of early onset delinquents was significantly lower than the verbal intelligence of late onset delinquents (F(1,208)=4.65, p=.03). Univariate ANOVA indicated that among white delinquents the verbal intelligence of early onset delinquents was not significantly lower than the verbal intelligence of late onset delinquents (F(1,130)=1.74, p=.19). Table 2 shows the Verbal IQs of black and white early and late onset delinquents.
Univariate ANOVA indicated that among black delinquents the disinhibition scores of early onset delinquents were not significantly higher than the disinhibition scores of late onset delinquents (F(1,208)=.003, p=.96). A Univariate ANOVA indicated that among white delinquents the disinhibition scores of early onset delinquents were significantly higher than the disinhibition scores of late onset delinquents (F(1,130)=7.37, p<.01). Table 3 shows the disinhibition scores of black and white early and late onset delinquents.
Family incomes were coded based on the lower end of the income range in thousands of dollars (e.g. family income in the $5,000 to $10,000 range was coded as 5). Univariate ANOVA indicated that among black delinquents, the family income of early onset delinquents was significantly lower than the family income of late onset delinquents (F(1,208)=24.10, p<.001). A Univariate ANOVA indicated that among white delinquents, family income of early onset delinquents was significantly lower than the family incomes of late onset delinquents (F(1,130)=6.40, p=.01). Table 4 shows the family incomes of black and white early and late onset delinquents.
In order to examine the relationship between early versus late onset of delinquency and all three of the factors noted above (verbal intelligence, family income, disinhibition), logistic regression was used, in order to predict the categorical variable of early versus late onset from the three continuous variables of income, disinhibition, and verbal intelligence. In order to determine if the variables made independent contributions to the prediction of early onset of delinquency, a forward stepwise procedure was used, adding variables to the equation if p<.05 and removing them from the equation if p>.05. Among black delinquents, family income and verbal intelligence were significantly associated with early onset of delinquency (family income, R=-.26, df= 1, p<.0001; verbal intelligence, R=-.08, df=1, p=.05) and were included in the final model. ([Chi.sup.2] for the full model=25.5, df=2, p<.0001). Disinhibition was not significantly associated with early vs. late onset (R=.00, df=1, p=.73) in black delinquents. Among white delinquents, family income and disinhibition scores were significantly associated with early onset of delinquency (family income, R=-.13, df= 1, p<.03; disinhibition, R=.15, df=1, p=.02) and were included in the final model. ([Chi.sup.2] for the full model=12.3, df=2, p=.002). Verbal intelligence was not significantly associated with early vs. late onset (R=.00, df=1, p=.39) for white delinquents.
In addition, in order to determine if the black and white delinquents differed in the frequency of early onset delinquency, even when family income, verbal intelligence and disinhibition were controlled, a logistic regression analysis was performed, including the categorical variable race as a predictor of early vs. late onset, along with the continuous variables of income, disinhibition, and verbal intelligence. When the variables of family income, verbal intelligence, and disinhibition were included, race was a significant predictor of early onset delinquency (Full Model [Chi.sup.2]=67.2, df=4, R=.14, p=.001)
Because the verbal intelligence and family incomes of black and white delinquents in this study are substantially different, it may be that black and white delinquents were simply so different in family income and verbal intelligence that they were not comparable. Therefore, the delinquents were divided into four groups based on verbal intelligence and family income. Delinquents were divided according to whether their verbal intelligence was more than one standard deviation below the normative average (i.e. 85), or not and thus those in the lower intelligence group could be considered to have a true deficit in verbal intelligence, and whether their family income was more or less than $15,000. These groups were created to examine the rate of early onset offending in groups of black and white delinquents, who were comparable in family income and verbal intelligence. The proportions, which were early onset offenders, were calculated for each group. For each of the four groups, logistic regression was used to predict early onset delinquency from race, thus determining whether the proportion of early onset offenders differed by race in each of these relatively homogenous groups. The results are presented in Table 5. Race did not significantly predict early onset offending in delinquents with higher verbal intelligence and higher family income. However, race was a significant predictor of early onset delinquency in delinquents with lower income and higher verbal intelligence ([Chi.sup.2] =12.3, df=1, p=.0005), in delinquents with higher income and lower verbal intelligence ([Chi.sup.2] =5.5, df=1, p=.02), and in delinquents with lower income and lower verbal intelligence ([Chi.sup.2] =5.6, df=1, p=.02).
The results are generally moderately supportive of Moffitt's hypotheses. Low family income is consistently a predictor of early onset offending. The evidence for the presence of deficits in verbal ability and impulsivity, specifically in early onset delinquents, was more mixed. Low verbal intelligence is associated with early onset offending in black delinquents. This complements Taylor et al. (2000) and Raine et al. (2005), which previously found that low verbal intelligence is associated with early onset antisocial behavior. Of note is that in the current study although the relationship between verbal intelligence and early onset delinquency in white delinquents was not significant, the difference in effect size for black and white delinquents was modest (i.e. .32 vs. .23). It is also of note that among white delinquents the average verbal intelligence of early onset delinquents was nearly in the average range. Therefore, it appears likely that only a small number of white delinquents had what could be considered a verbal "deficit." This limited range may partially account for the small relationship in this sample.
This study also confirmed a relationship between impulsivity and early onset offending in white delinquents, but not in black delinquents. In contrast to the results for verbal intelligence, there was a substantive difference in effect sizes of disinhibition for black and white delinquents, but there was not a large overall difference between the disinhibition scores of black and white delinquents. This is an intriguing racial difference in the correlates of early onset delinquency that should be explored further. As noted above, most studies that have found a relationship between impulsivity and early onset offending (e.g. Taylor, et al., 2000) have been conducted using white delinquents. This may suggest that impulsivity is a more important factor in the development of early onset delinquency in whites than in blacks.
Finally, this study indicated that black adolescents are more frequently early onset offenders, even when family income, verbal intelligence and impulsivity are controlled. The exception is the specific instance of the absence of the risk factors of low verbal intelligence and low family income, in which case the rates of early onset offending are similar for black and white delinquents. While the presence of one or both of these risk factors is associated with an increased rate of early onset delinquency in white delinquents, the effect is significantly stronger for black delinquents. Thus, being black seems to carry an increased risk for early onset delinquency only when combined with other risk factors.
Taken as a whole these results are supportive of Moffitt's hypotheses. However, they also suggest the factors associated with early onset delinquency may be somewhat different in different ethnic groups or subcultures. This is consistent with Moffitt's (1993) hypothesis that neurological deficits interact with the physical and social environment in which children are raised to produce early onset delinquency. Previous research on race differences in delinquency has tended to attribute this to socioeconomic factors, but it appears from the current results this may only be part of the explanation. Other potential explanations may lie in parenting styles and efficacy, family stability, quality of schools and neighborhood environments. Also, Moffitt (1993) characterized adolescence-limited delinquency as being a response to peer models, and it may be that among black children delinquent peer models are available at an earlier age. If this is the explanation, then this would suggest that blacks who are early onset offenders may be less likely to be life-course-persistent delinquents than white early onset offenders, as their early onset of offending would be a result of peer influences, which may decrease in adulthood rather than a result of neurological deficits, which are lifelong. This would need to be evaluated in a longitudinal design that extends into adulthood. Previous longitudinal research into adulthood (e.g. White et al. 2001) has studied primarily white delinquents, and thus, has not addressed issues of race differences in predictors of adult antisocial behavior.
These differences should be explored, in order to gain a better understanding of the factors that contribute to life course persistent delinquency in specific ethnic or subculture groups. Explanations of these differences need to address, not simply absolute differences in the rate of early onset delinquency in black and white delinquents, but why certain risk factors have a greater impact on early delinquency in blacks (e.g. income, verbal intelligence) or whites (e.g. impulsivity). These findings may facilitate more appropriate interventions and service options for juvenile offenders.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders(4th ed.) Washington, DC: Author.
Archer, R.P. (1997). MMPI-A: Assessing adolescent psychopathology (2nd ed.). Mahwah, NJ: Lawrence Erlbaum.
Archer, R.P., Belevich, J.K.S., & Elkins, D.E. (1994). Item-level and scale-level factor structures of the MMPI-A. Journal of Personality Assessment, 62, 332-345.
Archer, R.P. & Krishnamurthy, R. (1994). A structural summary approach for the MMPI-A: Development and empirical correlates. Journal of Personality Assessment, 63, 554-573.
Frick, P.J. & Hare, R.D. (2001). The Antisocial Process Screening Device. Toronto: Multi-Health Systems.
Hawkins, D.F., Laub, J.H., & Lauritsen, J.L. (1998). Race, ethnicity, and serious juvenile offending. In R. Loeber & D. Farrington (Eds.), Serious and violent juvenile offenders: Risk factors and successful interventions (pp.167-196).
Kaufman, A. & Kaufman, N. (1990). Kaufman Brief Intelligence Test. Circle Pines, MN: American Guidance Service.
Lynam, D., Moffitt, T., & Stouthamer-Loeber, M. (1993). Explaining the relation between IQ and delinquency: Class, race, test motivation, school failure, or self-control? Journal of Abnormal Psychology, 102, 187 - 196.
Moffitt T. E. (1993). Adolescent-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100 (4), 674 701.
Raine, A., Moffitt, T.E., Caspi, A.C., Loeber, R., Stouthamer-Loeber, M., & Lynam, D. (2005). Neurocognitive impairments in boys on the life-course persistent antisocial path. Journal of Abnormal Psychology, 114, 38-49.
Silverthorn, P., Frick, P.J., & Reynolds, R. (2001). Timing of onset and correlates of severe conduct problems in adjudicated boys and girls. Journal of Psychopathology and Behavioral Assessment, 23, 171-181.
Taylor, J., Iacono, W. G., & McGue, M. (2000). Evidence for a genetic etiology of early-onset delinquency. Journal of Abnormal Psychology, 109 (4), 634 643.
Wechsler, D. (1991). Wechsler Intelligence Scale for Children (3rd ed.). San Antonio TX: Psychological Corporation.
White, H. R., Bates M. E., & Buyske, S. (2001). Adolescent-limited versus persistent delinquency: Extending Moffitt's hypothesis into adulthood. Journal of Abnormal Psychology, 110 (4), 600 - 609.
White, J. L., Moffitt, T. E., Caspi, A., Bartusch, D. J., Needles, D. J., & Stouthamer-Loeber, M. (1994). Measuring impulsivity and examining its relationship to delinquency. Journal of Abnormal Psychology, 103 (2), 192 205.
Jennifer S. Parker
University of South Carolina Upstate
Todd L Morton
SC Department of Juvenile Justice
TABLE 1 Verbal Intelligence, Disinhibition Scores, and Family Incomes of Delinquents by Race Black n=446 White n=380 M SD M SD d Verbal IQ 79.7 12.0 92.5 13.1 .91 *** Disinhibition 51.9 5.8 51.1 5.5 .14 * Family Income 10.8 8.1 15.2 8.6 .50 *** Note. d computed using standard deviation of the full sample. Family incomes are annual family incomes in thousands of dollars * p<.05. ** p<.01. *** p<.001 TABLE 2 Verbal IQ of Early and Late Onset Delinquents by Race Early Onset Late Onset M SD (n) M SD (n) d Black 78.0 10.4 (143) 81.5 11.4 (67) .32 * White 89.9 13.8 (48) 93.0 13.0 (84) .23 Note. d computed using standard deviation of the full sample, divided by race. * p<.05. ** p<.01. *** p<.001 TABLE 3 MMPI-A Disinhibition Scores of Early and Late Onset Delinquents by Race Early Onset Late Onset M SD (n) M SD (n) d Black 51.6 5.7 (143) 51.5 5.7 (67) .01 White 52.9 6.2 (48) 50.0 5.7 (84) .50 ** Note. d computed using standard deviation of the full sample, divided by race * p<.05. ** p<.01. *** p<.001 TABLE 4 MMPI-A Family Income Scores of Early and Late Onset Delinquents by Race Early Onset Late Onset M SD (n) M SD (n) d Black 9.0 7.4 (143) 14.7 8.7 (67) .69 *** White 12.6 8.1 (48) 16.4 8.5 (84) .45 * Note. Family incomes are annual family incomes in thousands of dollars. d was computed using standard deviation of the full sample * p<.05. ** p<.01. *** p<.001 TABLE 5 Proportion of Early Onset Offenders by Race, Income and Verbal Intelligence Black White % Early (n) % Late (n) Onset Onset Income <15,000 & Verbal IQ<85 36 (215) 20 (55) Income<15,000 & Verbal IQ>=85 37 (78) 15 (108) Income>=15,000 & Verbal IQ<85 30 (95) 13 (48) Income>=15,000 & Verbal IQ>=85 14 (58) 9 (169)
|Printer friendly Cite/link Email Feedback|
|Author:||Parker, Jennifer S.; Morton, Todd L.|
|Publication:||North American Journal of Psychology|
|Date:||Jun 1, 2009|
|Previous Article:||Brief report: factor structure of the Experiences in Close Relationships with gay and lesbian individuals.|
|Next Article:||Effects of peppermint and cinnamon odor administration on simulated driving alertness, mood and workload.|