This article provides one of the first reports from a major longitudinal twin study of childhood aggression and antisocial behavior among a large ethnically diverse sample of twins. In this study, we focused on phenotypic and genetic analyses of antisocial behavior measures during a first wave of assessment at ages 9 to 10, when twins are on the cusp of adolescence. Instead of relying on information from one source (i.e., teacher or parent ratings of child behavior problems), we obtained ratings from 3 informants. The purpose of this article was to evaluate rater effects on the genetic and environmental influences on a shared view of antisocial behavior, using a composite measure based on a variety of types of aggressive and antisocial behavior.
In the present study we relied on composite measures of antisocial behavior created from 18 different subscales. Within each rater, subscales of reactive, proactive, and relational aggression; childhood psychopathy factors; and delinquent behavior measures (including conduct disorder symptoms) were all moderately to highly correlated with each other. These correlations (nearly uniform within each rater), as well as the results from our principal-components analyses, suggested the presence of a general antisocial or “deviance” factor underlying the various subscales provided by each rater. This general factor may be comparable to an overall externalizing factor that has been proposed by others (Achenbach & Edelbrock, 1981
; Krueger et al., 2002
) and reflects the wide range of behaviors exhibited by these preadolescent children. Although still somewhat negatively skewed, this general deviance factor well characterized the “shades of gray” in individual differences for antisocial behavior in this large sample and has proven useful in examining relationships with various biological (Jacobson, Zumberge, Lozano, Raine, & Baker, 2005
) and social risk factors (Sanchez, Baker, & Raine, 2005
) in this sample. Our continuous ASB factor may therefore reflect a wider spectrum of ASB than what is captured when relying on symptom counts (e.g., in Burt, McGue, Krueger, & Iacono, 2005
) or on extreme forms of disruptive behavior, substance use, or criminal offending.
Our analyses revealed that although mean levels of ASB differed for boys and girls, the sources of individual differences in ASB were similar across gender. One of the most important findings from this study is that a shared view of antisocial behavior is strongly genetically influenced, with little or no effect of shared sibling environment. Although our analyses revealed a moderate genetic basis to individual views of antisocial behavior, with heritabilities ranging from .40 to .50 for individual composites from child, teacher, and caregiver, the estimated heritability of the underlying shared view of antisocial behavior from the common pathways model was nearly 1.0. This latent factor may reflect constellations of stable personality traits (e.g., disinhibition, lack of constraint) that may influence antisocial behavior across many contexts (Krueger et al., 2002
). This highly heritable common factor representing the shared view across multiple informants could therefore prove especially useful in future investigations of specific genetic associations, or quantitative trait loci, in human aggression and antisocial behavior.
In this ethnically diverse sample of twins, heritability estimates within each rater are comparable to estimates from previous studies, which have been based primarily on Caucasian and European samples. Genetic influences for caregiver reports in our study are somewhat higher than those reported for young children in these reviews but are comparable to other recent twin studies of younger schoolchildren (Arsenault et al., 2003
). The somewhat higher heritability in the present study may be due in part to our use of a general composite measure of antisocial behavior, which may be more reliable than individual subscales typically used. We have, in fact, found the pattern of genetic and environmental influences to be more variable when examining specific subscales (Raine, Baker, & Liu 2006
; Ward, 2004
Interrater agreement among caregiver, teacher, and child reports of aggression and antisocial behavior in the present study is comparable to that of other studies (Achenbach et al., 1987
; Youngstrom, Loeber, & Stouthamer-Loeber, 2000
); agreement is lowest between the child and either the caregiver or teacher and somewhat higher between caregiver and teacher ratings. This suggests that although there is clearly a significant degree of overlap among raters, each individual viewpoint is influenced by unique factors. Of particular importance was the identification of significant rater variance for both caregiver and teacher reports. Although we are considering these caregiver and teacher rater effects to be biases due to having the same rater report on both twins, other explanations for these “correlated errors” are possible. Specifically, among caregiver reports, it is not possible to disentangle rater effects from true effects of shared environmental influences (Hewitt et al., 1992
). For example, family-level variables such as parental discipline may influence levels of antisocial behavior for twins in the same family. This shared environmental effect would also account for the correlated view in our model. Regardless of the specific source of the correlated view among caregivers, it is noteworthy that these effects accounted for a relatively modest (albeit statistically significant) proportion of variation (15%) in caregiver reports.
In contrast, for teacher reports, we were able to differentiate rater effects from true shared environmental effects, because although virtually all twins attended the same school, less than half of twin pairs were in the same classroom at school. This allowed us to disentangle shared environmental influences, which would affect the similarity of all twin pairs, regardless of classroom, from rater effects, which would only increase similarity among twins who were rated by the same teacher. In this study, rater effects accounted for more than one fourth (28.1%) of the overall variation in teacher reports. This indicates that the twins in the same classroom are rated more similarly than twins in different classrooms. Although we speculate that this is due to rater bias on the part of the teacher, it is theoretically possible that twins in the same classroom may in fact have a greater shared environment than those in separate classrooms (i.e., a direct classroom effect on behavior). To investigate this possibility, we examined post hoc whether caregiver or child ratings were also more similar if twins were in the same classroom at school, using the same dummy code for shared classroom that we used to evaluate the teacher rater effects (as described earlier). The results of these post hoc analyses indicate that being in the same classroom at school had virtually no effect on twin similarity of antisocial behavior as rated by either caregivers or the twins themselves. Being in the same classroom at school, therefore, does not lead to increased twin similarity in ASB based on either the caregiver's or the child's own view. Thus, our findings suggest that reports from teachers may be more heavily influenced by rater bias effects than are ratings from other reporters, leading to a spurious effect of shared environment when teacher reports are examined alone. However, in the absence of direct observational data, we cannot rule out the possibility that twins in the same classroom behave more similarly while at school. Nevertheless, if this is the case, it is important to note that these “classroom effects” are situational specific and do not affect similarity of behavior in other contexts.
In addition to these specific rater effects, shared environmental factors accounted for a significant 20% of the variation in teacher reports. These influences may reflect the effects of school context on ASB. For example, Rowe and colleagues conducted a behavioral genetic analysis within a hierarchical linear modeling framework and found that aggregate levels of parental warmth moderated both mean level of aggression as well as the overall impact of genetic and environmental influences on individual differences in aggression, with higher shared environmental influences on aggression found among twins and siblings from schools with lower average levels of parental warmth (Rowe, Almeida, & Jacobson, 1999
). Their findings suggest that environmental context, measured at the school level, not only moderates mean levels of ASB but may also alter the sources of individual differences in ASB.
It is important to note that there was no indication that caregivers or teachers moderated their views of twin similarity on the basis of the twins' zygosity. If caregivers or teachers were more likely to rate MZ twins more similarly than DZ twins, this would result in higher within-rater correlations for MZ twins than for DZ twins and would be interpreted in our model as specific genetic influences. The lack of specific genetic influence on either care-giver or teacher reports indicates that the rater effects we discovered did not upwardly bias estimates of heritability, nor can they be related to any characteristic of the child that is genetically influenced.
A different pattern emerged for the child's own report of ASB, with at least two important findings. First, the child's view contributes less weight to the shared view of ASB. Although the latent factor representing the shared view accounts for between 30% and 45% of the variation in caregiver reports and teacher reports, it accounts for only 17.6% of the variation in child reports. It is possible that children at this age are less reliable reporters of ASB. Such an interpretation is consistent with the fact that the 6-month test–retest correlations are somewhat lower for child reports than for caregiver reports (see ) and with the higher estimate of specific nonshared environmental influence on child reports, as nonshared environmental effects include measurement error. On the other hand, the other notable finding is that child reports are the only reports that show evidence for statistically significant specific genetic influence, which may reflect genetic influence on ASB which occurs outside the radar of parent or teacher perception. If this is the case, our results may indicate that reports of ASB from children are more comprehensive and, therefore, more accurate than caregiver or teacher reports. Alternatively, the significant genetic influence on child reports may simply reflect some sort of response bias, which is correlated with genetically influenced personality traits, such as social desirability or overall honesty. Future analysis using cotwin reports of ASB may help untangle genetically influenced rater bias effects from “real” genetic influence on child reports of ASB.
Finally, our results are consistent with the idea that the greater shared environmental influence found among childhood ASB relative to adult ASB may actually be an artifact of rater bias, as studies of children often rely on caregiver or teacher reports. In this study, when young children are asked to report on their own ASB, there is no evidence for significant shared environmental influences, nor do shared environmental influences account for variation in the shared view of ASB. This is consistent with other studies that have found that much of the shared environmental variation in parent reports can be attributed to rater effects (e.g., Hewitt et al., 1992
; although see Bartels et al., 2003
, for contradictory results) and is consistent with the meta-analysis by Rhee and Waldman (2002)
, which found that shared environmental influences on ASB were higher for parental reports than for child self-reports.
The present study should be viewed in the context of several potential limitations. First, we have chosen to “lump” rather than “split” various types of ASB in these analyses. The magnitude and nature of genetic and environmental influences in ASB may very well vary across different types of ASB (e.g., Tackett, Krueger, Iacono, & McGue, 2005
), a possibility we have examined in a separate article, which does suggest some distinction between aggressive–psychopathic behavior and nonaggressive delinquency, based on the underlying genetic and environmental architecture (Jacobson, Baker, & Raine, 2007
). Second, the age of the sample resulted in relatively low average rates of ASB, which may limit the generalizability of these findings. We note, however, that these children and their families appeared to be a representative sample of this urban community, as their ethnic distribution and socioeconomic levels were comparable to those of the local population. Children also exhibited a wide range of behaviors, including some serious conduct problems. Third, results concerning the teacher reports should be viewed with some caution, given the somewhat modest teacher participation rate (60%). Still, teacher participation was not influenced by sex or zygosity of the twins, suggesting that our results are not an artifact of unequal participation of teachers. Moreover, we found no evidence for systematic bias due to this lower response, as caregiver and child-rated ASB did not differ for those whose teachers did and did not participate (results are available upon request). Finally, we have yet to identify the source of the strong genetic effect on the shared view of ASB found in this sample. Ongoing analyses are attempting to address this by examining the genetic covariation of the ASB measures with putative biological endophenotypes, including psychophysiological, neurocognitive, and personality measures (e.g., Baker, Isen, Bezdjian, & Raine, 2005
; Jacobson et al., 2005
). Additional waves of assessment are also ongoing (Wave 2, ages 11–12) with future waves planned and funded through age 17. The Wave 1 assessments of antisocial behavior described in the present article therefore provide an important basis for investigating genetic and environmental influences on the emergence of antisocial behavior in American youth throughout the course of adolescence.