|Sponsored by the HealthcareTrainingInstitute.org providing Quality Education since 1979|
If you would like additional information on this topic,
Peer-Reviewed Journal Article References:
- Croake, James W., Treating Conduct Disorder in Adolescents; Individual Psychology: The Journal of Adlerian Theory, Research & Practice, Jun86, Vol.42 Issue 2, p270, 4p
- Goldstein, Arnold P. and Barry Glick, "Aggression Replacement Training: A Comprehensive Intervention for Aggressive Youth", Research Press: Illinois, 1987
- Ikomi, Philip A.; Harris-Wyatt, Georgetta; Doucet, Geraldine; Rodney, H. Elaine; Treatment for Juveniles Who Sexually Offend in a Southwestern State; Journal of Child Sexual Abuse, December 2009, Vol 18 Issue 6, p594
- Landes, Sara J. The Case: Treating Jared Through Dialectical Behavior Therapy. Journal of Clinical Psychology. May2013, Vol. 69 Issue 5, p488-489. 2p. DOI: 10.1002/jclp.21984.
- MacKenzie, Robert J. EdD, "Setting Limits in the Classroom", Three Rivers Press: New York, 2003
- Mpofu, Elias and Ralph Crystal, Conduct disorder in children: challenges, and prospective cognitive behavioral treatments, Counselling Psychology Quarterly, Mar2001, Vol. 14 Issue 1, p21-32, 12p
- National Institute for Health and Care Excellence. (2013). Antisocial Behaviour and Conduct Disorders in Children and Young People. The British Psychological Society and The Royal Collage of Psychiatrists.
- Phelan PhD, Thomas and Sarah Jane Schonour, MA, "1-2-3 Magic for Teachers: Effective Classroom Discipline Pre-K through Grade 8", ParentMagic, Inc: Illinois, 2004
- Stevenson, Margaret C.; Sorenson, Katlyn M.; Smith, Amy C.; Sekely, Ady; Dzwairo, Rukudzo A; Effects of defendant and victim race on perceptions of juvenile sex offenders; Behavioral Sciences & the Law, December 2009, Vol 27 Issue 6, p957
- Webster-Stratton, Carolyn and Jamia M. Reid, Treating Conduct Problems and Strengthening Social and Emotional Competence in Young Children: The Dina Dinosaur Treatment Program; Journal of Emotional & Behavioral Disorders, Fall2003, Vol. 11 Issue 3, p130, 14p
The High Costs of Aggression: Public Expenditures Resulting From Conduct Disorder. By: Foster, E. Michael; Jones, Damon E.. American Journal of Public Health, Oct2005, Vol. 95 Issue 10, p1767-1772, 6p, 2 charts, 1 diagram, 1 graph; DOI: 10.2105/AJPH.2004.061424; (AN 18461127)
Objectives. We explored the economic implications of conduct disorder (CD) among adolescents in 4 poor communities in the United States. We examined a range of expenditures related to this disorder across multiple public sectors, including mental health, general health, school, and juvenile justice.
The data for our analyses were collected as part of the Fast Track project, a multicohort, multisite longitudinal study of 1191 children who were at risk for emotional/behavioral problems.( n11-n13) The Fast Track project included an intervention targeted to children identified in kindergarten as "at risk" for such long-term problems; the individuals receiving such treatment are not included here. Rather, our focus was on 664 subjects: ( 1) children who were screened into the high-risk comparison group (n=396) and ( 2) children not identified as being at risk but recruited as part of a "normative" sample (n=268). When analyzed with probability weights, the combination of these 2 groups was representative of children from low-socioeconomic status neighborhoods in the 4 sites.
The Fast Track project was designed to answer 2 sets of questions. The first involved whether and for whom the intervention worked. The second set of questions involved normative development in these poor neighborhoods. Our analyses here fall into this second category.
The project relied on an array of sources for information on participating children and their families. Both the youths and their primary caregiver were present at in-person interviews. Information also was collected from school and court records and from the youths' teachers. Recruitment for Fast Track began in 1990 and continued for 3 years. The annual interviews began in the summer after kindergarten and are ongoing.
Three of the Fast Track sites were urban: Durham, NC, Nashville, Tenn, and Seattle, Wash. The racial makeup across these 3 sites was 49% African American. The remaining 51% were nearly all White (48%) with a small number of Latinos and Asians (3%). The fourth site was in rural Pennsylvania, where 98% of the sample was White. At baseline, approximately 41% of the participating youths at the 4 sites lived in single-parent households, and the overall socioeconomic status was between lower and lower-middle class. In the initial screening process, children were identified as being at high risk for behavior problems (or not) on the basis of assessments by teachers and parents using the Teacher Observation of Classroom Adaptation (revised) and Child Behavior Checklist, respectively. High-risk children then were randomly assigned into treatment or comparison groups. The screening process was repeated across 3 cohorts. More detail on the overall screening process for this project can be found in other primary Fast Track research.( n11-n13)
In years 7 (sixth grade) and 10 (ninth grade) of the project, the Fast Track group assessed the participants with the Diagnostic Interview Schedule for Children to determine the presence of mental disorders.( n14) Of the 664 participants, 638 (96%) provided diagnosis data in at least 1 of those years. Parents were asked whether their children demonstrated symptoms related to CD during the past year. Children exhibiting more than 2 symptoms were diagnosed as having CD. In the year 7 sample, 39 individuals (4.0% of the sample when weighted) were classified as having CD; in the year 10 interview, 31 (3.5%) were classified as having CD. Males in the sample were much more likely to be diagnosed with CD (10.5% males diagnosed vs 2.7% females diagnosed in either year). African Americans were also more likely to be diagnosed with CD than Whites (4.9% vs 2.3%, respectively).
The oversampling of high-risk children (i.e., children scoring positive for problem behaviors on the initial screening) boosted the number of children and youths with a diagnosis of CD. In particular, children scoring positive for such behaviors on the initial high-risk screening were more than twice as likely to have a subsequent diagnosis of CD in either year (10.9% vs 5.1%). Their presence in the weighted sample did not affect the representativeness of the data, but they greatly improved the precision of the cost estimates.
For our study, children with conduct problems were classified into 1 of 3 groups, with the remainder of the sample (488; 81.1%) classified as others. Participants were considered to have CD if they met diagnostic criteria in either year 7 or year 10 (59 cases; 6.2% of the sample). Of the remainder, a second group (78; 8.2%) comprised those meeting diagnostic criteria for oppositional-defiant disorder, "a pattern of negativistic, hostile and defiant behavior."( n7) A final group (40; 4.4%) included those who never met diagnostic criteria but who exhibited elevated levels of problem behaviors (i.e., within 1 criterion of being diagnosed).
As noted, as many as 96% of the sample provided diagnostic information in at least 1 year. Our analyses relied on multiple imputations to correct for any systematic patterns in attrition or other nonresponse. Additional information on the imputation is available from the authors.
We derived information on service use from interviews with participating families and from administrative data. The latter involved a review of school records every summer, which provided information on whether youths repeated a grade or received special education. (These outcomes represent the marginal costs of additional services received. We did not include the costs of regular education because youths with and without conduct problems would experience those costs, canceling them out.)
Information on the use of health and mental health services as well as juvenile justice involvement was provided by parents in the Service Assessment for Children and Adolescents,( n15) including (for the purposes of our study) how many service visits and number of days a youth received service occurred in the past 12 months. Additionally, parents provided annual information on whether or not their child required medications for emotional/behavioral problems.
Data collection for this information began in year 7 of the project and is ongoing. The cohorts differed somewhat in the data they contributed to our analyses. For instance, because service assessments were staggered, cohort 1 did not receive the Service Assessment for Children and Adolescents in year 8. Additionally, cohort 3 had 2 fewer years of data than cohort 1 at the time of the analysis. In order to provide full (entire sample) estimates for the 7-year time period, we created multiple imputations to accommodate missing cases.
Calculation of public costs of behaviors requires information on both the behaviors and the outcomes involved and also relevant per-unit costs. In general, these figures were calculated as the amount that a state or local government paid for the service or treatment. For costs that could not be derived from previous research, we relied on data from follow-up examination of service use within this sample. All figures taken from the literature were converted to 2000 dollars using the Consumer Price Index.
The outcomes included cover several categories, each combining several service types. These categories include general health (emergency department, family doctor, general hospital); inpatient mental health costs (psychiatric hospital, residential treatment center, group home, foster care); outpatient mental health costs (drug and alcohol clinic, day treatment center, mental health center, in-home provider, individual counselor/ therapist); juvenile justice (detention center, arrest costs); and school (school counseling, special education, grade retention).
Parental reports indicated high rates of service use and system involvement. For instance, by year 11 (the last year data were available for all cohorts), roughly 5% of youths had ever received services for emotional/behavioral problems at an inpatient facility, 15% had received outpatient services, and 18% had received special education services. An additional 21% had had contact with the police.
Figure 1 plots the average total costs across years for the conduct disorder symptom groups. The CD group stood out quite strikingly from the other 3 groups, which were clustered together. The differences grew as the children matured; annual costs in year 13 (end of high school) exceeded $14 000 per child for the average youth with CD. This figure was over 6 times that for youths without conduct problems (roughly $2300). The trend over time was significant (P=.05) only for the CD group. (The time trend was assessed through a joint test of the group by year interaction.)
Table 1 provides more detail on these overall levels. Statistical significance refers to the differences among the 4 groups, calculated by regressing the logarithm of the costs of the diagnosis groups. (Significance refers to the joint significance of the 3 dummy variables representing the between-group differences.) The between-group differences were significant at most ages and for most types of costs. The exceptions included physical and mental health services in some years. In the case of the former, especially inpatient mental health services, the lack of significance largely reflected the small number of individuals involved in those systems. When mental health costs were incurred, they were often quite substantial. For example, inpatient and outpatient mental health costs accounted for nearly 70% of the difference between the CD and Other groups.
The other notable cost for which between-group differences were not significant involved juvenile justice at young ages. As we expected, youths were seldom involved in that system before year 9 of the study (eighth grade).
The variation across groups explained only 5% of the total variance (calculated using the log transformation of costs model). One explanation for the variation within groups is comorbidity. Supplemental analyses revealed that comorbid attention problems did not raise expenditures substantially.
Hidden in these averages were substantial within-group differences. Table 2 reports the median and other key percentiles (25th, 75th, and 90th) for total costs. For all groups, mean expenditures were far greater than median expenditures, reflecting the presence of very high cost individuals. To provide a sense of how important a few cases were, the last column provides the percentage of total costs accounted for by the top 10% of each group (i.e., the individuals above the 90th percentile, which is also reported). These individuals accounted for roughly half of all expenditures in all 4 groups.
Supplemental analyses illustrated how high levels of service use accounted for expenditures in these extreme cases. The year 10 data, for example, revealed that among those admitted to inpatient facilities, the median length of stay was 58 days. However, 10% of those admired had a length of stay of 365 days--they were in a residential facility for the entire year.
Figure 2 displays the average composition of public costs across sectors. School expenditures represented a substantial proportion of public expenditures and were influenced by the high costs of special education and retention. In proportional terms, these costs were smaller for the CD group. (In reading this figure, one should be careful not to translate the lower percentages into lower absolute amounts. The CD group still had the highest school costs averaged across years among these groups, as evidenced in Table 1.) To some extent, this difference reflects the fact that youths with CD were more likely to drop out of high school, reducing the use of school services. Figure 2 also shows that juvenile justice expenditures represented approximately 20% of total expenditures for the youths with CD. This far exceeds the highest percentage among the other groups (oppositional defiant disorder; 11%). Interestingly, children who had elevated levels of problem behaviors (but were not diagnosed with CD) required a higher percentage of spending on inpatient mental health assistance than the other groups.
The public costs of behavior problems among children and youths are enormous. When summed across all 7 years (years 7 through 13), expenditures for the CD youths were nearly $70 000 larger than those for the children in the No Disorder group. It is important to note, however, that the latter are hardly risk free. They live in poor neighborhoods and so are at much greater risk of poor outcomes than the average American youth. For that reason, one can think of these findings as capturing the effect of CD with control for a range of risk factors.
Our results are consistent with limited prior research. In a 1999 study of 10 youths referred for mental health services, Knapp et al.( n16) found that yearly social costs exceeded $15 000. The greatest expense fell on the families themselves (just over a third), while the education authority bore an additional third. Significant costs fell on the health, social services, and welfare systems. In a second study, Scott et al.( n17) examined the experiences of 142 youths living in central London. These youths were identified at 10 years of age as either having no conduct problems, conduct problems, or a diagnosis of CD. The authors then considered the costs incurred by various public agencies through 28 years of age. These costs primarily involved the costs of crime as well as educational, health, and social services expenditures. Costs for those with CD were 10 times those without problems and 3.5 times those for youths with conduct problems (but not meeting the criteria for CD). Crime-related costs were the largest for the conduct problems and CD groups (35% and 64%, respectively). Educational costs accounted for the second-highest proportion, representing 31% and 18% of the 2 groups, respectively. These findings suggested a somewhat greater gap for those with CD relative to those with only elevated levels of symptoms.
Unlike Scott et al., we have not extended our analyses beyond the public costs to include social costs. Nonetheless, the expenditures captured here are especially important from a public policy perspective. They highlight the fact that these children and youths are already costing taxpayers a great deal of money. The key policy question is not whether to spend money on these children but rather how to spend it.
Still, this perspective does have limitations. For example, the school costs began to fall in later years as the youths left school. Dropping out may reduce public expenditure but only when gauged from a very narrow, short-term perspective. Clearly, dropping out of school raises the likelihood of future welfare and crime-related expenditures. For that reason, these expenditures represent a lower bound for the potential societal benefits of prevention.
As briefly noted, a surprising finding is that comorbid attention problems did not raise public costs. However, this finding rests on the validity and reliability of the instrumentation used, and the role of other comorbid conditions or profiles deserves additional attention.
The data on which this study was based are unique and have several strengths. They were longitudinal and included an oversample of young children who were at greater risk of developing CD. Because expenditures were so skewed, these children were particularly important in statistical analyses. This study also has other strengths, such as data collection from 4 diverse communities and high follow-up rates. Furthermore, the data included true diagnostic measures of CD rather than simple measures of aggressive symptoms (such as the Child Behavior Checklist).
Nonetheless, this study has several limitations. One limitation of these analyses is that we applied the same per-unit costs to facilities (within a given category) regardless of the diagnosis of the children involved. As a result, our findings may understate the difference between youths with CD and other youths if the former were treated in more expensive or intensive settings. For that reason, our findings are best judged as conservative.
Another limitation is that measures of service use reflected parental self-reports and may have underreported actual service activity. As a result, the costs presented may be underestimates. However, comparisons of system involvement for which we have both parental report and administrative data are somewhat reassuring. For example, a comparison of court records and parental reports of juvenile justice involvement showed 85% agreement. If parents or youths underreported services, we have no reason to believe that one group underreported to a greater degree than another. In that case, the between-group gaps would be smaller in absolute terms but not in percentage terms. (Measures of statistical significance involving between-group differences would not be affected either.)
The high costs of bad behavior are discouraging. These figures, however, offer some reason for hope. In particular, the expenditure gap between children with CD and those with lower, but elevated, levels of behavior problems was still substantial. This gap suggests that the problem behaviors of the high-cost children with CD need not be eliminated entirely to achieve substantial savings; rather, substantial savings could be realized simply by reducing those problems.
Our analysis suggests that public expenditures may be reduced if resources are moved from coping with problem behaviors to preventing them. A necessary condition is that effective programs be developed, and results in this area are encouraging. Furthermore, even among these youths with CD in high-risk neighborhoods, the public costs are still relatively concentrated among a small group. As a result, the cost-effectiveness of an effective intervention could be enhanced by effective targeting, and research suggests that these children can be identified accurately.( n18)
The implementation of such programs, however, represents a public health challenge. Resources will need to be shifted from established uses, such as juvenile detention, into other areas, such as the public mental health system. Such a change in focus and resources likely will require strong leadership from a public health leader with a broad public health perspective.
Factor structure of the Eyberg Child Behavior Inventory: a parent rating scale of Oppositional Defiant Behavior Toward Adults, Inattentive Behavior, and Conduct Problem Behavior. (eng; includes abstract) By Burns GL, Patterson DR, Journal Of Clinical Child Psychology [J Clin Child Psychol], ISSN: 0047-228X, 2000 Dec; Vol. 29 (4), pp. 569-77; PMID: 11126634
Used the Eyberg Child Behavior Inventory (ECBI) to measure disruptive behavior problems in children and adolescents. A controversy exists, however, on the dimensional structure of the ECBI. To evaluate this issue, an exploratory factor analysis was first performed on a sample of 1,263 children and adolescents. This analysis identified 3 meaningful factors (i. e., Oppositional Defiant Behavior Toward Adults, Inattentive Behavior, and Conduct Problem Behavior) and a fourth, poorly defined factor. A confirmatory factor analysis (CFA) evaluated the fit of the 3 meaningful factors in a second sample of 1,264 children and adolescents. The 3-factor model with 2 correlated errors provided a excellent fit. This 3-factor model also provided a significantly better fit than 2- and 1-factor models. Multiple group CFA indicated that the factor pattern, item-factor loadings, factor correlations, and correlated errors were equivalent across the samples. The CFA on sex yielded similar results. Initial normative information is presented for boys (n = 1,322) and girls (n = 1,205) within 4 age ranges (i.e., 25, 6-9, 10-13, 14-17) for the 3 factors. The use of these 3 factors, especially Oppositional Defiant Behavior and Conduct Problem Behavior, should make the ECBI more useful as a screening and outcome measure.
ECBI. The ECBI contains 36 disruptive behavior problems. The parent indicates on a 7-point scale how often each behavior occurs; 1 (never), 2 and 3 (seldom), 4 (sometimes), 5 and 6 (often), and 7 (always). The parent also indicates if the occurrence of the specific behavior is currently a problem by circling "yes" or "no" for each behavior. This results in two summary scores--an intensity score (IS) and a problem score (PS). The IS score represents the total frequency of occurrence of the 36 behaviors (possible range from 36 to 252). The PS represents the total number of the 36 behaviors that are indicated to be problems (possible range from 0 to 36). Table 1 shows the 36 items on the ECBI.
For the pediatric sample, a total of 1,526 ECBIs were completed by parents or guardians in five outpatient pediatric clinics in four northwestern states (Pullman, WA; Seattle, WA; Lewiston, ID; Missoula, MT; and Portland, OR). For the random sample, 300 children were randomly selected on the basis of sex and ethnicity (Asian, African American, and Caucasian) within each grade level for Grades 1 to 12 from the Seattle School District (a total of 3,600 parents were mailed ECBIs). A total of 1,003 completed ECBIs were returned by the parents. This return rate of 28% was similar to a return rate of 29% obtained in a second study in the Seattle School District (Bums et al., 1997). Two of the adolescents were 18 years old and these two ratings were eliminated because they were outside the age range of the ECBI (2-17). This left a total of 1,001 children and adolescents.
The combination of the pediatric and random samples resulted in data on 2,527 children and adolescents. The sample was 52% boys and 48% girls, with an average age of 8.95 years (SD = 4.36, range 2-17). A total of 1,639 (65%) of the children were living with their biological mother and father; 464 (18%) with their mother only; 30 (1%) with their father only; 245 (10%) with their mother and stepfather; 42 (2%) with their father and stepmother; 18 (< 1%) with foster parents; and 89 (4%) with other relatives. In terms of ethnicity, 85% of the children were Caucasian, 5% African American, 4% Asian, 3% American Indian, less than 1% Hispanic, and 4% mixed ethnicity (e.g., 1/2 Caucasian and 1/2 African American). A total of 2,180 (86%) ECBIs were completed by the child's mother, 255 (10%) by the child's father, and 92 (4%) by other relatives or foster parents. The average education of the person who completed the ECBI on the child was 13.91 grades (SD = 2.65). A total of 179 (7%) of the raters had not completed high school; 860 (34%) had obtained a high school degree; 626 (25%) had attended some college; 538 (21%) had obtained a college degree; and 324 (13%) had completed some graduate study. In terms of family income, 335 families (14%) reported a yearly income of less than $10,000; 391 (16%) between $10,000 and $19,999; 557 (22%) between $20,000 and $29,999; and 1,194 (48%) over $30,000. Fifty of the raters did not provide information on family income. In terms of treatment status, 2,335 (92%) of the children were not currently in treatment for learning disabilities or behavioral problems; 86 (3%) were in treatment for learning disabilities; 72 (3%) for behavioral problems; and 34 (1%) for learning and behavioral problems.
Structural Organization of the ECBI
The 2,257 children were randomly separately into two samples, 1,263 in the first and 1,264 in the second. The random assignment was performed so that each sample contained an equal percentage of children from the pediatric clinics and the Seattle School District. The factor analyses were performed on the IS item ratings because the PS item ratings involved a categorical variable (i.e., a "yes" or "no" answer for each item).
Table 4 shows the internal consistency coefficients (Cronbach's alpha), means, standard deviations, and the scores corresponding to the 80th, 90th, and 98th percentiles for the Oppositional Defiant, Inattentive, and Conduct Problem Behavior subscales. These results are presented for boys and girls as well as boys and girls within four age ranges (2-5,6-9,10-13,and 14-17).
For boys (n = 1,322), the correlation between the IS and PS Oppositional Defiant Behavior subscales was .77, .70 for the Inattentive Behavior subscales, and .73 for the Conduct Problem Behavior subscales. For girls (n = 1,205), the correlation between the IS and PS Oppositional Defiant Behavior subscales was .77, .70 for the Inattentive Behavior subscales, and .73 for the Conduct Problem Behavior subscales.
The results from the EFA identified three meaningful factors in the ECBI and a fourth, poorly defined factor that probably represents response bias (Nunnally & Bernstein, 1994; C. Parks, personal communication, May 8, 2000).
The three meaningful factors were Oppositional Behavior Toward Adults, Inattentive Behavior, and Conduct Problem Behavior. A CFA on a second sample indicated that these three factors provided an excellent fit. The CFA also indicated that this model provided a significantly better fit than two- and one-factor models. In addition, multiple-group CFA demonstrated the equivalence of the factor pattern, loadings, correlations, and correlated errors across the samples. Similar results also occurred for sex. Three items, however, had significantly higher loadings for boys than girls (i.e., "destroys toys and other objects," "verbally fights with friends his/her own age," and "physically fights with friends his/her own age"). For boys, thus, verbal aggression, physical aggression, and property destruction had a stronger relation to Conduct Problem Behavior construct than for girls.
The attention deficit and disruptive behavior domain (Quay & Hogan, 1999) involves other factors in addition to these three factors in the ECBI. Within the ECBI item set, however, the Oppositional Defiant Behavior Toward Adults, Inattentive Behavior, and Conduct Problem behavior are the three clinically meaningful dimensions. Given the size of our samples as well as the repeatability of our findings across samples, sex, and IS/PS ratings, the same three factors should occur in other similarly large samples.
Although our normative data on the three subscales do not represent national norms and individuals should carefully consider the characteristics of our sample in their use of these norms, this information on 1,322 boys and 1,205 girls from four states as well as urban and rural settings represents the best available normative information on the ECBI. The normative information also addresses a major limitation of the ECBI. When it is important to identify children who are presenting with "pure" CP [conduct problems], one potential solution [to the problem that the ECBI contains ODD, CD, and ADHD items] might be to score only those items related to either ODD (or ODD and CD), which would facilitate the selection of more homogeneous samples of children (but would preclude the use of existing normative data and cut-off scores). (McMahon & Estes, 1997, p. 153)
Neurobehavioral characteristics of adolescents with behavioral dysregulation disorder. (eng; includes abstract) By Pineda DA, Ardila A, Rosselli M, Puerta IC, Mejia S, Toro MC, The International Journal Of Neuroscience [Int J Neurosci], ISSN: 0020-7454, 2000; Vol. 101 (1-4), pp. 133-55; PMID: 10765995
Twenty-five misbehaved male adolescent students and 25 normal male controls participated in this study. The two samples were matched by age (13 to 17-year-olds), and socioeconomic status (SES). All the participants were selected from the same school in Medellin (Colombia). This school is a private institution which receives 2 to 3 SES children (middle low SES). Six different SES are recognized in Medellin. SES 4 represents the widest middle class, and SES 1 represents the poorest population; SES 5 and 6 are formed by highest socioeconomic level individuals. SES 2 and 3 are represented by families with a monthly income equivalent to about 3 to 4 minimal legal wages (i.e., they have a month income equivalent to about 500 to 1,200 American dollars). Usually, families living in SES areas 2 and 3 have a high school level of education, and frequently some additional technical or clerk training. They work as qualified workers, clerk employees, salespersons, cab-drivers, etc., SES is determined in each city area according to the price of the houses, and the monthly income of the population in that particular area. As in most countries, gangs are more frequent in low SES city areas. Because in Colombia only primary education represents a state obligation, the secondary (high school) education is mainly private. Children attending to specific school most often belong to the same SES. All the children were monolingual native Spanish speakers. Testing was performed in Spanish.
School teachers were instructed, blind to the research objectives, to administer a behavioral questionnaire to the three best, most disciplined, and reliable students; and to the three worst, undisciplined, and unreliable students from each grade (7th, 8th, 9th, 10th and 11th). Teachers used their own direct knowledge and professional criteria in selecting these children. The latter group was referred to as "Behavioral Dysregulation Disorder" (BDD) or "misbehaved children" (experimental group). An abbreviated Spanish version of the Conners Teacher Rating Scale (Conners, 1979) was constructed (see Appendix A). The eight questions comprising the Behavioral Disorder factor and four questions of the Hyperactivity Index of the Conners Teacher Rating Scale (short version) were selected. A scale ranging from zero to three, was used for each question (Goyette, Conners and Ulrich, 1978). The maximum score possible was 36. A cut-off point was established: under 50% of the maximum score (i.e., below 18 points) for the control subjects; and over 30 points (i.e., 80% of the maximum score) for the BDD cases. These cut-off scores were based on a randomized database of 540-subject standardization of the Conners Teachers Rating Scale in Colombia (Pined et al., unpublished). The Alpha reliability coefficient for the full scale was 0.91 in male adolescents (12 to 17-year-old), for the conduct problem subscale it was 0.82, and for mixed conduct problems plus hyperactivity and academic problems subscale (used in this study) was 0.85. Five control subjects with scores between 19 to 22, and five cases with scores between 26 and 30 were removed. As a result, 25 subjects were in each group. None of the participants had a previous diagnosis of ADHD.
The following instruments were administered to all the subjects in both groups:
3.1. Phonological verbal fluency: verbal production of words
in one minute, beginning with the Spanish sounds/f/,/a/and/s/.
3.2. Semantic verbal fluency: animal and fruits produced in
one minute. For the analyses, number of correct words in the
phonological, and semantic conditions were used (Ardila,
Rosselli and Puente, 1994).
These questionnaires were answered either by the
participants' mothers, or by the maternal grandmothers when
the mother was absent for any reason. Quantitative scores
were obtained using a zero-to-three scale for each question.
The higher the score, the greater number of symptoms in the
participant (Appendix B).
6.4. Use of weapons and drug-use. The mother and children
were independently asked to answer some questions about the
use of weapons and drugs (alcohol, cocaine, "bazuco" -base
cocaine paste, inhalants, marijuana, heroine, and
benzodiacepines). A quantitative score was obtained using
a zero-to-three scale for each question. Similarly, a
higher score indicates more frequent weapon contact or
7. A prenatal, neonatal and neurological history questionnaire was answered by the subjects' mothers (Appendix C). The following
questions were included: vaginal bleeding during pregnancy,
abdominal pain and contractions, infections, use of drugs
and medications, maternal alcoholism or smoking, eclampsia,
and seizures during pregnancy. Some questions about the
children's early development were also included: seizures,
developmental milestones, cerebral palsy, language and
motor retardation, meningitis, and encephalitis before the
age of six. Quantitative scores were obtained using a zero
(never) to three (almost always; very frequent) scale.
Testing was performed by specially trained graduate neuropsychology students under the supervision of a professor. Examiners were blind to the purpose of the research.
A database for the statistical software SGPLUS (Statgraphic) 7.0 and the SPSS 8.0 were used. Descriptive procedures were developed. Multivariate analyses of variance (MANOVA) were performed in order to analyze differences between age, school achievement, and groups (cases and controls). Age and school achievement were correlated with scores in the different tests and questionnaires. Several Multiple Regression Stepwise (Criteria Probability of F < .05) analyses were carried out in order to identify variables which were able to predict the variance on the criteria variable.
Table I presents the characteristics of the two samples. No statistically significant age differences were observed between both experimental and group. A statistically significant difference on the Conners Teacher Rating Scale questionnaire-used as the selection criterion, was found. The Full Scale IQ (FSIQ) and Verbal IQ (VIQ) were significantly lower in the experimental group, as compared to controls. Performance IQ (PIQ), however, was similar in both groups. The experimental group also had a statistically significant lower school achievement; almost one school grade differentiated both groups.
When VIQ and FIQ subtests were analyzed, it was observed that the Information, Similarities and Vocabulary subtests presented statistically significant differences between both groups, with lower scores in the BDD group. Hence, lower verbal skills in the experimental group were evident (Tab. II). In the Performance subtests, statistically significant differences were observed only on the Digit-Symbol subtest.
None of the tests used to assess executive functioning showed statistically significant differences between both groups (Tab. III). Scores were rather similar. A significant higher score in the weapons use and drugs-use measure was observed in the case group. The questions about use of weapons and drugs were the only two behavioral variables that significantly differentiated both groups (Tab. IV). No significant differences were observed in the other questionnaire scores.
No significant between group differences in the prenatal, neonatal and neurological history variables were found (Tab. V). Ideomotor dyspraxia, and left/right orientation difficulties were the only two SNS that presented a tendency to differentiate between groups, with a poorer performance in the experimental group (Tab. VI).
Table VII presents the Pearson correlation analysis with the variables that showed statistically significant correlations with the Conners Teacher Rating Scale questionnaire, used as the criterion variable for group selection. All correlations were in the moderate range. VIQ, Information, and Similarities were the variables with higher correlation coefficients. It was observed that several WISC-R variables, including the compound scores named VIQ and FSIQ were significantly related to the school achievement.
A Multiple Regression Stepwise analysis with executive function variables with low Spearman Correlation Coefficients--in order to control the collinear effect, and the Conners teachers based questionnaire as dependent variable was developed. No variable predicted significantly the variance of the criteria variable. Similar analyses found that neither any variable of the neurological antecedents nor the SNS examination were able to predict this variance. Several stepwise procedures with WISC-R variables showed that a model formed by WISC-R VIQ variables Information, Similarities and Vocabulary was able to predict the variance of the Conners teacher based questionnaire in our sample (Ad-R-SQ = 0.20, F-Ch. = 5.270, p < 0.01). No WISC-R FSIQ variable was able to predict the dependent variable. FSIQ, VIQ and PIQ as compound scores were not included in any regression model either.
A Multiple Regression Stepwise procedure with behavioral antecedents variables (Conflictive peer relationships, Attention Deficit history, Impulsive disorder antecedents, Oppositional disorder history, Child Conduct disorder before 12 years old) and current behavior problems (drug-use and weapon use) was developed. Only drug-use was selected in the model (Ad-R-SQ = 0.10, F-Ch. = 6.747, p < .02).
Finally, a model using the predictive variables in each domain (Information, Vocabulary, Similarities, and drug-use) was developed. Two models were selected: (1) WISC-R information (Ad-R-SQ = 0.17, F.Ch. = 10.025, p < .01), and (2) WISC-R Information and drug-use (Ad-R-SQ = 0.26, F.Ch. = 9.478, p = .0001).
Some significant limitations to this study must be pointed out. Subjects in the experimental group were not antisocial or delinquents, or even ADHD individuals. They were just, according to the teachers' opinion, "the most undisciplined and unreliable studens." This is the reason to refer to them as "behavioral dysregulated" (or "misbehaved children"). An association with antisocial or delinquent behavior has not been demonstrated. Furthermore, the sample was too small (only 25 subjects in each group) complicating the statistical analyses and obscuring the conclusions.
Our results demonstrated statistically significant differences in VIQ and FSIQ between the case and control groups. However, PIQ was similar in both groups. An analysis of the scores on each subtest showed that the statistically significant differences were determined by lower scores in the Information, Similarities, and Vocabulary subtests in the experimental group. The scores on the other subtests were similar in both the case and control groups.
Only three of eight cognitive skills evaluated for obtaining the compound quotients (VIQ, PIQ, and FSIQ) were significantly abnormal in the experimental group, and all of these were related to verbal abilities. However, these differences may also be related to cultural and scholastic variables (Rosselli and Ardila, 1997). The school achievement of the case group was significantly lower than in the control group, and several variables of intelligence, including the compound scores as measured by VIQ and FSIQ, were significantly related to school achievement.
The finding of significantly lower school achievement in the BDD group, and the significant correlation between several variables assessing intelligence and the scholastic level, could support the assumption of a poorer academic training in misbehaved adolescents. Several studies have found that low scholastic achievement is related to difficulties in verbal communication abilities, the presence of conduct disorders, a definite tendency to solve interpersonal problems by means of aggression, and with an inability to inhibit impulsive responses (Kazdin and Crowley, 1997; Moffitt, 1993). This verbal deficiency would also produce a lack of categorical comprehension and conceptualization, with difficulties in behavioral auto-regulation using the internal language commands (regulatory language function) (Luria, 1979, 1984).
Conduct disorders have usually been analyzed in the context of the psychopathology. Conduct disorders are characterized by a persistent breaking of the social rules, which are internally coded as linguistic propositions (i.e., syntactic complex sentences) (Kamphaus and Frick, 1996). An interaction between the social rules in a cultural context and their correct comprehension is necessary for controlling one's conduct. Understanding this interaction would have crucial implications in therapeutic approaches to children and adolescents with behavioral problems (Kazdin and Crowley, 1997; Moffit, 1993).
Contrary to that expected, only two child behavioral problems related to adolescent conduct disorder were found: higher scores on the weapon use and drug-use variables. None of the other questionnaires related to the child behavioral problems coded in the DSM-IV (American Psychiatric Association, 1994) (e.g., ADHD, ODD, CCD) were found to differentiate between both groups. This finding may indicate that the onset of the specific misbehavior in our sample was exclusively related to adolescence. Perhaps, they have different behavioral characteristics with regard to young adults with conduct disorder, which have been clearly related to ADHD, ODD and CCD in several studies (Biederman et al., 1995; Gittelman, Mannuzza, Ronald and Bonagura, 1985; Mannuzza et al., 1988, 1989; Satterfield, Swanson, Schell and Lee, 1994). We preferred to refer to our misbehaved experimental group as "behavioral dysregulation disorder" (misbehaved children), instead of "conduct disorder" individuals.
Significant differences between groups on the tests for assessing executive functioning were not found. This result contradicts several research studies, which have found lower scores on the executive function tests in young individuals with conduct and behavioral disorders (e.g., Blake, Pincus and Buckner, 1995; Elliot, 1992; Giancola and Zeichner, 1994; Lapierre, Braun and Hodgins, 1995; Lueger and Gill, 1990; Malloy, Bihrle, Duffy and Cimino, 1992). Three hypotheses could account for this discrepancy: (1) The specific adolescent misbehaving disorder population in our sample had milder and different deficits than the young adult participants included in other studies. (2) It may be conjectured that the cognitive tests we used to measure executive functioning did not assess the domain in a comprehensive manner. Hence, the tests we administered may have been sensitive to some executive functions, but not other executive abilities. For example, it is known that subjects with a frontal behavioral syndrome (particularly, inferomedial) can present normal scores in several multioperational and time controlled tests used to assess executive functions (Cumming, 1985; Damasio, Tranel and Damasio, 1990; Fuster, 1989; Hare, 1979; Hare, Williamsom and Harper, 1988; Luria, 1979, 1984; Luria and Stvetkova, 1981; Stuss and Benson, 1986; Thorpe. Rolls and Maddinson, 1983; Vanderploeg and Haley, 1988). In these cases only those tests assessing social problems can disclose the behavioral deficiencies behind an apparently unaltered cognitive activity (Hall and Kramer, 1995; Hare, Williamson, and Harper, 1988; Luria and Tsvetkova, 1981).
Even though "executive function" tests did not discriminate between both groups, performance was slightly higher in the control group. Hyperactivity-Impulsive, ADD, Oppositional defiant disorder, and conduct disorder scores were slightly higher in the experimental group, albeit differences were not statistically significant. Drug and weapon use, however, did discriminate between groups. This result may point to a significant environmental influence in misbehaved children.
Only two SNS variables (right-left orientation, and ideomotor dyspraxia) presented a tendency to discriminate between the groups. This finding was not completely unexpected. SNS are clearly related with age and school achievement, and tend to disappear spontaneously in adolescence (Ardila and Rosselli, 1996; Lopera, 1997). None of the prenatal, neonatal and neurological history variables showed significant differences between groups, which is in disagreement with other studies (Harris, 1995; Nichols and Tu-Chuan Chen, 1981). Nonetheless, the frequency in both groups was extremely low, and results are not totally reliable.
The final multifactor stepwise regression procedure selected two models from two different domains. The first corresponded to the verbal skills WISC-R Information, which could be related to language learning domain (school achievement) or environmental general knowledge. In the second models a defined behavioral-environmental variable (drug-use) enhanced the predictive capability of the WISC-R Information for predicting misbehavior.
Attention to novelty, fear-anxiety, and age: their effects on conduct problems. (eng; includes abstract) By Eaves RC, Darch C, Williams TO Jr, The Journal Of Genetic Psychology [J Genet Psychol], ISSN: 0022-1325, 2004 Dec; Vol. 165 (4), pp. 425-49; PMID: 15636387
CHILDREN AND ADOLESCENTS WITH EXTREME CONDUCT DISORDER (antisocial behavior) are easily bored, exhibit poorly sustained attention, and tend toward sensation-seeking behavior, according to Quay (1965). Others have supported this proposition (e.g., Borkovec, 1970; DeMyer-Gapin & Scott, 1977; Orris, 1969; Raine, Reynolds, Venables, Mednick, & Farrington, 1998; Siddle, Nicol, & Foggitt, 1973; Skrzypek, 1969; Walker et al., 1991; Whitehill, DeMyer-Gapin, & Scott, 1976). Quay (1977) argued that these characteristics are motivated by a pathological need for stimulation. For example, Skrzypek found that psychopathic delinquents had significantly lower pretest anxiety and significantly higher preference for novelty scores than did neurotic delinquents.
Alternatively, we hypothesized that low attention to novelty would lead to more conduct problems, that high fear-anxiety would lead to more conduct problems, and that age would interact with level of attention to novelty and level of fear-anxiety to influence level of conduct problems. Specifically, we expected older children with lower levels of attention to novelty and higher levels of fear-anxiety to show higher levels of conduct problems. Conversely, we expected older children with higher levels of attention to novelty and lower levels of fear-anxiety to show lower levels of conduct problems.
Data for 368 participants were solicited by 141 examiners, who were enrolled in an assessment course at Auburn University. The participants were students attending local public and private schools in Alabama. The sample from which we drew the participants comprised 569 volunteers. Of the 184 boys and 184 girls, 233 were White, 113 were African American, 6 were Asian, and 2 were Pacific Basin/Aleut (ethnic information for the remaining 14 participants was unavailable). The mean age of the participants was 9 years, 11 months (SD = 2 years, 6 months; range = 6 to 16 years). We estimated socioeconomic status (SES) by using scores that were based on the occupation of the head of household (U.S. Bureau of the Census, 1963). The mean SES score was 65.26 (SD = 26.15; range = 1 to 99), which indicated that the sample was largely from middle-class backgrounds, but the group exhibited considerable variability. The mean IQ for the group, as measured by the Slosson Full-Range Intelligence Test (Algozzine, Eaves, Mann, & Vance, 1993), was 101.38 (SD = 17.73). Of the 368 participants, 295 (80%) were attending regular classes. The remainder were receiving special education services at least part of the school day for the following disabilities: emotional conflicts (n = 37), learning disabilities (n = 17), mild mental retardation (n = 12), and moderate mental retardation (n = 7). Of those participants attending regular classes, 19 had Full-Range IQs of 132 or higher, considered intellectually gifted by some schools. We computed a chi-square test to determine whether or not the older and younger participants differed in terms of the educational services they received. The chi-square statistic was not significant, χ² ( 5,N = 368) = 4.94, p = .67. Because conduct problems are known to occur much more frequently among males, we randomly dropped 179 cases from our sample to equate the number of boys and girls within each of the 8 three-way interaction cells (attention to novelty x fear-anxiety x age). We dropped an additional 22 participants because of administration errors in the test protocols.
Attention to novelty. The Visual Similes Test II (VST II) is a research instrument developed for use with children aged 5 to 21 years to test Eaves' (1995) hypotheses concerning affective arousal (i.e., VST II Affective Form) and cognitive arousal (VST II Cognitive Form). Researchers have studied arousal and have implicated a wide array of phenomena: (a) orienting response (Spinks & Siddle, 1983), (b) defensive response (Mogenson, 1977), (c) curiosity (Berlyne, 1960), (d) habituation and perception (Sokolov, 1963), (e) attention (S. E. Rose, Feldman, Wallace, & Cohen, 1991), (f) motivation (Revelle & Loftus, 1992), (g) vigilance (Kinomura, Larsson, Gulyas, & Roland, 1996), (h) preference for novelty (Thompson, Fagan, & Fulker, 1991), (i) stimulation seeking (Zuckerman, 1979), and (j) persistence (Lindsley, 1958). All vertebrates share these functions. The VST II was developed as an analogue to novelty preference instruments that had been used in early childhood research to predict childhood intelligence (e.g., Fagan & Detterman, 1992; Fagan & McGrath, 1981; Lewis & Brooks-Gunn, 1981; McCall, 1994; McCall & Carriger, 1993; D. H. Rose et al., 1986; and Slater et al., 1985).
In typical early childhood studies, infants are exposed to a photograph until they have gazed at it for a predetermined amount of time. This result is called the familiar stimulus. The infants are then simultaneously exposed to the familiar photograph and a photograph to which they have not been previously exposed (i.e., the novel stimulus). The duration of the infants' gaze at the familiar stimulus and the novel stimulus is then measured until some predetermined criterion duration is reached. The measure, the percentage of novelty preference (%NP), is the time spent gazing at the novel stimulus divided by the time spent looking at the novel and familiar stimuli combined and multiplied by 100. Researchers studying early childhood have shown that infants generally spend more time looking at novel stimuli rather than at familiar stimuli. More important, the %NP predicts IQ during the childhood years. That is, infants with the highest %NPs have higher IQs in childhood than do infants with the lowest %NPs. Three investigations using the VST (Eaves, 1992) or the VST II (Eaves, 1995) have shown that the early childhood results are maintained into childhood and adolescence (Cox, 2000; Eaves & Glen, 1996; P. Williams, 1996).
The VST II Affective Form consists of 30 visual stimuli, and the VST II Cognitive Form consists of 25 visual stimuli. The VST II Affective Form stimuli were designed to elicit limbic responses related to survival (e.g., predation, pain, food, death, disasters). The VST II Cognitive Form stimuli were designed to draw on abstract cortical activity (i.e., they depict stimuli that require an abstract transformation of the information presented, such as January 1 as New Year's Day). The items that are presented in each VST II form are relatively easy to identify in their unoccluded form. To make the items more novel, the researchers use a checkerboard pattern that obscures 50% of each stimulus. The checkerboard occlusions make each item less familiar and make the task more difficult. In addition, the checkerboard pattern makes lucky guesses improbable.
There are separate test booklets, response sheets, standardized testing procedures, and scoring manuals for the VST II Affective Form and for the VST II Cognitive Form. Each VST II booklet contains, on separate pages, the visual stimuli that are to be presented to the examinee. The testers use stop watches to record the amount of time that the examinee spends on each item (i.e., response latency). The average time needed to administer the complete VST II is approximately 25 min.
Unlike most tests, examinees control the pace of the VST II administration. Examinees may freely respond impulsively or with great deliberation. For example, the response "I don't know,", is immediately accepted and the examiner moves on to the next item. The examiner intervenes in the examinee's responding in only three circumstances. First, when a response contains no flaws, but is incomplete, the examiner questions the examinee (e.g., "Tell me more about that"). Second, if 30 s pass without a response from the examinee, the examiner asks, "What do you think it is?" Finally, if 60 s have elapsed and the examinee is not actively responding, the examiner says, "Let's move on to the next item." Initial tests showed that examinees rarely require as much as 60 s to produce a scorable response. The intent of allowing the examinee to control the pace of the administration was to systematically minimize examiner characteristics that would contaminate estimates of the strength of the examinee's arousal. The time devoted to questioning is subtracted from the total time for that particular item. Examiners rarely spent more than 1 or 2 s to ask each question.
Each item has scoring criteria for assigning points (1 or 0) and for questioning a response. For example, for an item showing a child receiving a spanking, the criteria require that the response must mention both a child and the action of being spanked. Typical correct responses are presented: (a) "A woman spanking a little boy" and (b) "A kid getting a whipping." Typical incorrect responses also are provided: (a) "A doctor giving medicine" and (b) "Someone working." Finally, information about questioning an incomplete response is provided: "If the examinee says, 'It's an old woman,' say, 'Okay, tell me, what is she doing?'"
A variety of scores can be computed for the VST II. A raw score of correct responses, correct response latency (CRL), error response latency (ERL), and %NP can be calculated. The CRL is calculated by averaging, across items, the amount of time, in seconds, between the presentation of the stimulus and the receipt of a correct response from the examinee. The ERL is calculated by averaging, across items, the amount of time, in seconds, between the presentation of the stimulus and the receipt of an incorrect response for the examinee. Thus, the CRL and ERL are the mean scores, in seconds, for correct and incorrect responses, respectively. The %NP is the ratio of ERL divided by the total response latency multiplied by 100 (i.e., ERL / [CRL + ERL] x 100). Thus, %NP indicates the relative amount of time that the examinee spent attending to unfamiliar, or novel, stimuli.
The evidence regarding the psychometric properties of the VST II is sparse, but consistently supportive. The split-half reliabilities of the VST II affective and cognitive forms are high. Using yearly samples of children who were between 5 and 15 years (n = 366), the median split-half reliability for response latency was .96 for the VST II Affective Form (range = .92 to .98). For the VST II Cognitive Form, the median split-half reliability for response latency was .94 (range = .85 to .96). The split-half reliabilities were corrected using the Spearman-Brown formula. The items of the VST II Affective Form correlate more highly with the VST II Affective Form total score than with the VST II Cognitive Form total score. The parallel statement also can be made regarding the items of the VST II Cognitive Form. A group of 45 judges who did not know the item categories and purposes of the subscales were asked to rate the extent to which each item depicted affective and cognitive content. Items were presented in a mixed arrangement without occluding checkerboards. Mean ratings of each item were significantly different in the expected direction. That is, items from the affective form were rated as significantly more affective, and items from the cognitive form were rated as significantly more cognitive. An exploratory factor analysis (using miniscales constructed from affective and cognitive items, respectively) indicated a strong, unrotated first factor, presumably indicating that both subscales measure a behavioral dimension of arousal. When the variables were rotated obliquely, two factors emerged that corresponded exactly with the affective and cognitive forms of the VST II (T. O. Williams, Cox, & Eaves, 2000). Finally, T. O. Williams, Eaves, and Cox (2002) conducted a confirmatory factor analysis of the VST II using a representative sample of children who were 10 to 12 years old. The results for the second-order model suggested that it was the best model for the data. Although the statistically significant chi-square value was indicative of a poor fit and thus not supportive, there was considerable improvement in chisquare statistic, χ² ( 34,N = 216) = 88.47, when compared with the one-factor model, χ² ( 35,N = 216) = 409.68, and uncorrelated two-factor model, χ² ( 35,N = 216) = 283.35. The normed fit index, Tucker-Lewis index, and comparative fit index values all exceeded .90, which supported the hypothesized model. The root mean square error of approximation (RMSEA) value of .086 slightly exceeded the accepted RMSEA level of .08. In addition, all measured variables loaded on their respective constructs saliently, and all critical ratios exceeded the minimal accepted value of 1.96.
Fear-anxiety. We used the Personality Problem (PP) subscale of the Behavior Problem Checklist (BPC; Quay & Peterson, 1979) to estimate participants' level of fear-anxiety. We used the PP subscale rather than the anxiety-withdrawal subscale of the Revised Behavior Problem Checklist (RBPC; Quay & Peterson, 1987) because the former subscale has 14 items compared with 11 items on the latter subscale. In general, as the number of scorable items increases, reliability increases. We also attempted to increase subscale reliability by using the entire three-point Likert-type scale rather than the dichotomous scoring procedure recommended in the BPC manual. Participants assigned to the high fear-anxiety group had raw scores greater than 3 on the PP subscale. This resulted in unequal numbers of participants in the two groups (high PP n = 164, low PP n = 204). Although a number of the participants in the high group would not be considered excessively anxious or fearful in an absolute sense, if we had raised the cut-off score to 4, the group sizes would have been even more imbalanced.
The PP subscale has proven to be a useful measure of a broad internalization construct. The items reflect attributes such as social withdrawal, fearfulness-anxiety, tension, timidity, inhibition, and so forth. The BPC manual reported split-half reliability coefficients of .83 and .81 for samples of more than 1,000 and 124 delinquents, respectively. Interrater reliability was low between parents and teachers (r = .32 to .41), but better between parent pairs (r = .67). Finally, the 2-week test-retest reliability reported by Quay and Peterson (1979) was good (r = .74). The validity of the BPC PP subscale is well established. Its factor validity has been replicated many times by different researchers for different clinical samples and across cultural groups (e.g., Hayashi, Toyama, & Quay, 1976; Lessing & Zagorin, 1971; Peterson, 1965; Quay, Morse, & Cutler, 1966). The PP subscale correlates well with the RBPC Anxiety-Withdrawal subscale (Mdn r = .86). Finally, the PP subscale has been used in a variety of investigations of both clinical and nonclinical samples to study relationships between anxiety-withdrawal and delinquent recidivism (e.g., Quay & Love, 1977), physiological reactions such as galvanic skin responses (Borkovec, 1970) and electroencephalograph records (Müller & Shamsie, 1967), and psychological and pharmacological treatment changes (Greenwold & Jones, 1971).
Age. Participants were assigned to the high or low group on the basis of the midrange split of the ages, measured in months. We assigned 184 participants to the high age group and 184 participants to the low group.
Three dependent measures of childhood conduct problems and aggression were available for analysis--BPC conduct-problem scores, RBPC conduct-disorder scores, and RBPC socialized-aggression scores. BPC conduct-problem scores were available for all 368 participants. RBPC conduct-disorder and socialized-aggression scores were available for only 285 of the participants.
The reliabilities of the dependent measures reported in the BPC manual are more than adequate for research purposes. For example, the median conduct-problem internal-consistency reliability for two studies was .90. As reported in the BPC manual (Quay & Peterson, 1979), interrater reliabilities for conduct-problem scores across four studies were lower: Mdn = .71 (range = .23 to .83). The validity of the conduct-problem subscale has been extensively investigated (e.g., Gajar, 1979; Herr, Eaves, & Algozzine, 1977; Speer, 1971; Zold & Speer, 1971). Quay and Peterson (1979) summarized the results, "The studies ... provide evidence for the relationships of the BPC subscales to a wide variety of psychologically relevant variables which are consonant with the psychological meanings of the subscales" (p. 9).
With regard to the conduct-disorder and socialized-aggression subscales of the RBPC, Quay and Peterson (1987) cited internal-consistency reliability estimates for five samples. The reliability coefficients ranged from .92 to .95 with a median of .94 for the conduct-disorder subscale. For the socialized-aggression subscale, reliabilities ranged from .85 to .93 with a median of .89. Interrater reliability estimates were lower. Across seven samples, the median interrater reliability for the conduct-disorder subscale was .71 (range = .13 to .87). Across the same seven samples, the median interrater reliability for the socialized-aggression sub-scale was .59 (range = -.10 to .93). Validity evidence reported in the RBPC manual covered a wide variety of investigations: (a) correlations with the BPC; (b) differentiation of deviant and normal children; (c) a comparison between RBPC categories and those of the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1980); (d) correlations of the RBPC with behavioral observations, peer nominations, intelligence, and achievement; and (e) correlations of the RBPC subscales with other measures of behavior problems. Most of these studies supported the validity of the RBPC conduct-disorder and socialized-aggression subscales.
The BPC, RBPC, VST II (Affective Form, Cognitive Form), and Slosson Full-Range Intelligence Test were part of a larger battery of instruments that the examiners administered to the participants. The additional instruments included the Woodcock-Johnson Reading Mastery Tests-Revised (Woodcock, 1987) and the KeyMath-Revised (Connolly, 1988). Prior to administration, each examiner was assigned an administration sequence, which ensured that all instruments appeared with equal regularity in each administrative position.
The 141 examiners were students studying school psychology and special education and who were enrolled in an Auburn University course designed to enhance their assessment skills. All of the examiners had completed at least one course in assessment, and the first author had instructed all of them in the correct use of each instrument before administering the battery to children. The examiners were told only that the VST II was being administered to obtain item-tryout data. They were not given any reason for estimating item response latencies, and they were unaware of the true item-difficulty sequence. Therefore, there was little opportunity for the examiners to allow experimental biases to affect their scoring of items.
Table 1 presents the means and standard deviations of the research variables for each group. We found two significant differences between groups for the independent variables. On Affective ERL, N = 368, the high and low age groups differed by 2.56 s. This difference was statistically significant, F ( 1,372) = 13.47, p = .0003, and indicates that, on average, older children spent relatively more time on items they answered incorrectly than did younger children. The high and low ERL groups differed in age by 6.59 months, a difference that was statistically significant, F ( 1,372) = 4.46, p = .0353. Thus, the high ERL group was older than was the low ERL group.
We also found three statistically significant differences between groups for the demographic variables. The mean Full-Range IQ of the high ERL group was 4.71 points higher than was the score of the low ERL group, F ( 1,372) = 6.71, p = .0100. The mean Full-Range IQ of the low PP group was 9.08 points higher than was the score of the high PP group, F ( 1,372) = 25.86, p < .0001. Finally, the mean SES of the low PP group was 12.52 points higher than was that of the high PP group, F ( 1,372) = 22.40, p < .0001.
We computed a completely randomized three-factor analysis of variance (ANOVA) for the dependent variable (i.e., conduct problem). The results, displayed in Table 2, indicated a significant main effect for the PP factor and the three-way interaction (ERL x PP x age). For the PP factor, high PP participants (M = 8.37) had higher conduct-problem scores than did low PP participants (M = 2.54).
Scheffé multiple comparisons for the three-way interaction indicated that the conduct-problem mean for the low ERL-high PP-high age group (M = 10.29), the high ERL-high PP-low age group (M = 9.52), and the low ERL-high PP-low age group (M = 8.63) were statistically equivalent, but the means were significantly higher than for the remaining five groups. The conduct-problem mean for the high ERL-high PP-high age group (M = 6.51) was significantly higher than it was for the high ERL-low PP-high age group (M = 3.20), the low ERL-low PP-low age group (M = 2.65), the high ERL-low PP-low age group (M = 2.19), and the low ERL-low PP-high age group (M = 2.10). We did not find other significant differences. A layout of these interaction effects is shown in Table 3.
Conduct Disorder and Socialized Aggression
Because data for 285 participants were available for both conduct disorder and socialized aggression, we computed a multivariate ANOVA using three independent and two-dependent variables. The results are displayed in Table 4. With one exception (i.e., the PP x age interaction), all main effects and interactions were statistically significant. Consequently, we computed separate univariate ANOVAs for the RBPC conduct-disorder and socialized-aggression subscales.
Conduct disorder. The results of the univariate ANOVA for conduct disorders are displayed in Table 5. We found significant main effects for ERL and PP and significant interactions for ERL x PP and ERL x Age. With regard to the main effect for ERL, participants with low ERL scores had a significantly higher conduct disorder mean (M = 8.24) than did participants with high ERL scores (M = 5.95). Participants with high PP scores had a significantly higher conduct-disorder mean (M = 12.27) than did participants with low PP scores (M = 3.11).
For the ERL x PP interaction, Scheffé multiple comparisons indicated that the conduct-disorder mean for the low ERL-high PP group (M = 14.10) was significantly higher than for the other three groups. The conduct-disorder mean for the high ERL-high PP group (M = 10.07) was significantly higher than for the low ERL-low PP group (M = 3.24) and for the high ERL-low PP group (M = 2.96). The difference between the latter two groups was not significant. For the ERL x Age interaction, the mean conduct-disorder scores for the low ERL-high age group (M = 9.34), the low ERL-low age group (M = 7.49), and the high ERL-low age group (M = 6.86) were not significantly different, but all were significantly higher than were the conduct-disorder mean of the high ERL-high age group (M = 5.28). A layout of these differences can be seen in Table 6.
Socialized aggression. The results of the univariate ANOVA for socialized aggression, displayed in Table 7, indicated significant main effects for ERL, PP, and age, as well as significant interactions for ERL x PP, ERL x Age, PP x Age, and ERL x PP x Age. For the main effects, low ERL participants (M = 2.01) had higher socialized-aggression scores than did high ERL participants (M = 1.08); high PP participants (M = 3.01) had higher aggression scores than did low PP participants (M = 0.46); older participants (M = 2.39) had higher aggression scores than did younger participants (M = 0.80). The three-way interaction can be seen in Figure 1.
For the ERL x PP interaction, Scheffé multiple comparisons indicated that the aggression mean for the low ERL-high PP group (M = 3.97) was significantly higher than for the other three groups. The aggression mean for the high ERL-high PP group (M = 1.73) was significantly higher than for the high ERL-low PP group (M = 0.60) and the low ERL-low PP group (M = 0.33). The difference between the latter two groups was not significant. For the ERL x Age interaction, the mean aggression score for the low ERL-high age group (M = 3.74) was significantly higher than for the other three groups. There were no significant differences between aggression means for the remaining three groups: high ERL-high PP (M = 1.29), low ERL-low age (M = 0.81), and high ERL-low age (M = 0.79). For the PP x Age interaction, the mean aggression score for the high PP-high age group (M = 4.17) was significantly higher than for the other three groups. The high PP-low age group (M = 1.78) was not different from the low PP-high age group (M = 0.89), but it was higher than the low PP-low age group (M = 0.07). Finally, the low PP-high age group was not significantly different from the low PP-low age group.
The Scheffé multiple comparisons for the three-way interaction indicated that the aggression mean for the low ERL-high PP-high age group (M = 7.21) was significantly higher than for the remaining seven groups. The aggression means for the high ERL-high PP-low age group (M = 1.95), the low ERL-high PP-low age group (M = 1.68), the high ERL-high PP-high age group (M = 1.59), the high ERL-low PP-high age group (M = 1.05), and the low ERL-low PP-high age group (M = 0.70) were not different from one another. However, the aggression means for the latter five groups were significantly higher than were those of the low ERL-low PP-low age group (M = 0.08) and the high ERL-low PP-low age group (M =0.06). The aggression means for the last two groups were not significantly different from one another. A layout of these differences can be seen in Table 8.
For two out of three dependent measures, we found low attention to novelty, as measured by the VST II Affective Form, to be associated with high conduct problems, irrespective of participants' levels of fear-anxiety and age. Attention to novelty also interacted with fear-anxiety for two of three dependent measures; low ERL combined with high fear-anxiety produced significantly higher conduct-disorder and socialized-aggression scores than any other combination. Conversely, the combination of high attention to novelty and low fear-anxiety produced conduct-disorder and socialized-aggression mean scores that fell into the lowest set of means across groups. Attention to novelty also interacted with age for the same two dependent measures. Older participants with low attention to novelty had conduct-disorder and socialized-aggression scores that were among the highest, and older participants with high attention to novelty had conduct-disorder and socialized-aggression scores that were among the lowest.
Fear-anxiety, as measured by the BPC personality problem subscale, had a direct influence on conduct-problem scores for all three dependent measures. High fear-anxiety was invariably associated with higher conduct problems. Fear-anxiety interacted with attention to novelty for the dependent measures conduct-disorder and aggression scores. Fear-anxiety also interacted with age to influence aggression scores. Older children with high fear-anxiety scores produced the highest aggression scores; older participants with low fear-anxiety produced aggression mean scores that were in the lowest stratum.
The main effect of age was pronounced only for the aggression measure. Given the nature of the items on the socialized-aggression subscale (e.g., drug and alcohol abuse, gang activities, theft), which are rarely seen among younger children, we were not surprised that the older group mean exceeded the younger group mean by a factor of three. The significant interactions involving age suggest that age is of primary importance when it is linked to particular levels of attention to novelty and fear-anxiety.
The null hypothesis of greatest interest stated that the level of conduct problems does not change as a function of the interaction between levels of attention to novelty, fear-anxiety, and age. Our alternative hypothesis anticipated that older children with low levels of attention to novelty and high levels of fear-anxiety would produce high levels of conduct problems. Conversely, we expected older children with high levels of attention to novelty and low levels of fear-anxiety to produce low levels of conduct problems. The null hypothesis was rejected for two of the three dependent variables (i.e., conduct problem and socialized aggression); the F - ratio for the third dependent variable (conduct disorder) approached significance (p = .07). A visual inspection of Figure 1 shows that the older participants with low attention to novelty (ERL) and high fear-anxiety (PP) scores produced the highest aggression scores, whereas older participants with high attention to novelty and low fear-anxiety scores produced aggression scores that were significantly lower. The group reflecting opposite characteristics (i.e., high attention to novelty-low fear-anxiety-low age) consistently had dependent measure mean scores that fell in the lowest stratum. This general pattern was similar for all three analyses.
We have inferred that these results support the conclusions of Kashani et al. (1991), Ialongo et al. (1994,1996), and Zoccolillo (1992), who found that conduct problems and anxiety co-exist in children, and that anxiety and fear do not serve as a protective factor against the development of conduct problems in a significant number of cases. The results are also consistent with Gray's (1987) theory, which conceptualized activation and inhibition, not as opposite extremes of a single continuum, but as two systems that vary independently.
The hypotheses for our investigation were based on three common-sense ideas. With regard to attention to novelty, it makes good adaptive sense to pay close attention to unfamiliar (i.e., novel) environmental stimuli. On the one hand, such stimuli may lead to sustenance in the form of food, water, shelter, sexual gratification, and so forth. On the other hand, they may signify threat of punishment, pain, or even death. In either case, people who ignore novel stimuli do so at their own risk. In evolutionary terms, the fit individual is the one who attempts to classify novel stimuli as good or bad and as reinforcing or punishing. This classification is accomplished by responding to the stimulus in some way and then noting the outcome. For example, if the response leads to the cessation of hunger, the stimulus is good; if the response leads to physical discomfort and pain, the stimulus is bad. Repeated exposures to novel stimuli and their behavioral outcomes gradually lead to the consolidation, in memory, of behavioral routines that have led to satisfactory outcomes on past occasions. In this way, adaptively important novel stimuli become familiar stimuli that are associated with successful behavioral responses. We assumed that individuals who spend little time attempting to classify novel stimuli fail to develop effective adaptive responses to those stimuli. To the observer, it seems as if such individuals fail to learn from experience, often finding themselves in conflict with authority figures because they make the same mistakes over and over. Thus, we expected participants with low attention to novelty scores to exhibit high levels of conduct problems.
When individuals exhibit low attention to novelty, they often misjudge the meaning of environmental events. Reinforcing stimuli may be ignored, neutral stimuli may be actively avoided, and aversive stimuli may be maladaptively approached. When individuals living in a chaotic environment misjudge a neutral environmental stimulus to be aversive, they commonly feel fear. Fear usually evokes one of two responses. The individuals either flee or fight. When flight appears to be impossible, individuals have little recourse but to aggress. Given this practical description, we expected anxiety and fear, together with an interactive influence from low attention to novelty, to be associated with conduct problems in a significant number of participants.
We expected time, as measured by age, to be an important interactive factor in the growth of conduct problems as a function of low attention to novelty and high fear-anxiety. Except in cases when they are genetically determined, low arousal and high fear do not become chronic without a multitude of environmental experiences that underlie their existence. In summary, we considered attention to novelty, fear-anxiety, and age to have good potential to interact in the formation of conduct problems among school-aged children and adolescents. The results of our investigation seem to support this formulation.
The results of our study conformed to the tenets of Eaves' (1993) integrated theory of human behavior, which postulates that, at birth, humans are genetically capable of becoming many different sorts of people. The sort of person individuals actually become depends on the demands placed on them by the environment. By the end of the developmental period, individuals develop those neurologically based routines that most closely conform to the specific demands they encounter regularly in their environments. Although we have no direct evidence that the participants in the low attention to novelty-high fear-anxiety-high age group were leading their lives in harsh, chaotic environments, the results are consistent with such speculations. Because the groups with opposite characteristics on the independent measures invariably had mean scores at opposite extremes of the distributions of the dependent measures, this lends further credence to our alternative hypothesis: Children who live in chaotic environments develop low levels of attention to novelty and high levels of fear and anxiety. In addition, as they grow older, they exhibit higher levels of conduct problems.
- Coordinating Author/Instructor: Tracy Appleton,
Others who bought this Conduct Disorders Course
CEU Continuing Education for
Psychology CEUs, Counselor CEUs, Social Worker CEUs, MFT CEUs