درباره رابطه درگیری های گذشته تا درگیری های آینده در جرم و بزهکاری: تجزیه و تحلیل ژنتیکی رفتاری

Abstract Purpose Criminologists have devoted much attention to identifying the factors that drive stability in antisocial behavior. This body of research has, however, overlooked the contributions of behavior genetic research. This study sought to blend behavior genetics with the different perspectives used by criminologists to explain stability. Methods Employing a behavioral genetic research design, the current study analyzed the correlation between adolescent and adulthood crime (a 13 year time span was covered between the two time points) among a sample of sibling pairs drawn from the National Longitudinal Study of Adolescent Health (Add Health). Results The findings revealed that genetic factors accounted for nearly all of the stability in offending behavior from adolescence to adulthood. Environmental factors (particularly, of the nonshared variety) accounted for the majority of the changes in offending. Conclusions The implications of these results for criminological research and theory are discussed.

Introduction Biosocial criminology is an emerging paradigm that holds considerable promise for increasing scholars’ understanding of the origins of antisocial behavior (DeLisi and Piquero, 2011 and Piquero, 2011). Broadly, biosocial criminology seeks to blend biological, genetic, sociological, and environmental explanations of human behavior into a single analytic focus. This body of literature has shown that genetic and biological factors are significant influences on the development of maladaptive traits (Moffitt, 2005 and Raine, 1993). Results from four recent meta-analyses, for example, suggest that about 50 percent of the variance in antisocial behaviors is attributable to genetic factors (Ferguson, 2010, Mason and Frick, 1994, Miles and Carey, 1997 and Rhee and Waldman, 2002). The remaining 50 percent is divided among shared environmental influences (i.e., environments that operate to make siblings more similar to one another) and nonshared environmental influences (i.e., environments that operate to make siblings different from one another). Despite the vast literature, biosocial criminology lacks a unified theoretical framework. To date, there is no single theory that incorporates all of the findings from biosocial research into a succinct set of propositions and theoretical axioms. This does not mean that theorists have not proffered biosocial theories. To be sure, there are a number of theories that incorporate biosocial arguments into their original hypotheses (Barnes, Beaver, & Boutwell, 2011). Notable examples are the theories set forth by Ellis (2005), Moffitt (1993), and Robinson (2004; Robinson & Beaver, 2010). Given the wealth of criminological theorizing (Lilly, Cullen, & Ball, 2011), however, some scholars have argued that extant theories should be revamped to incorporate evidence from biosocial research (Rowe & Osgood, 1984). Along these lines, Walsh (2002) showed that biosocial inquiry may allow researchers to fill in some of the gaps left by contemporary criminological research. One remaining gap, for example, concerns the identification of the various factors underlying stability in antisocial behavior over the life course. Indeed, a great deal of theorizing (Gottfredson and Hirschi, 1990, Sampson and Laub, 1993 and Wilson and Herrnstein, 1985) and empirical work (see below) has been extended to explain the well-known finding that past behavior is one of the best predictors of future behavior (Robins, 1966). The state dependence argument suggests that past criminality increases the probability of future offending due to the effects/outcomes of past behavior (e.g., cumulative continuity). Sampson and Laub (1993) argued that prior involvement in crime and delinquency causes future involvement in crime and delinquency because of opportunities that are lost as a consequence of past behavior. In short, opportunities for a prosocial lifestyle are knifed-off due to earlier delinquent activity; the result being that future delinquency becomes more likely. Population heterogeneity, conversely, argues that an individual's unique propensity toward offending (i.e., their level of criminality) accounts for behavioral stability. Gottfredson and Hirschi's (1990) theory of low self-control offers a primary example of the population heterogeneity perspective. These authors explained that persons will differ in their level of delinquent behavior as a result of their different levels of self-control. Individuals who have lower levels of self-control will be more likely to offend as compared to individuals who have higher levels of self-control. According to Gottfredson and Hirschi (1990), levels of self-control remain relatively stable over time (at least after adolescence; Hay & Forrest, 2006) and, therefore, account for the correlation between past and future involvement in delinquency. Despite numerous empirical tests, scholars remain divided in their interpretations of the influences that drive behavioral stability (Nagin & Paternoster, 2000). As is discussed shortly, this body of research has overlooked the possible contributions from biosocial research. Behavior genetics is one area of biosocial criminology that offers a unique opportunity to analyze the genetic and environmental influences on behavioral stability. The goal of this paper, therefore, is to analyze stability in antisocial behavior through the lens of behavior genetics. The following sections review the relevant literature bearing on the stability of antisocial behavior. Attention is first given to evidence gleaned from criminological studies. Second, relevant findings produced by behavior genetic research are presented.

Findings Because the current analysis was motivated to unpack the stability of delinquent behavior over time, it was first important to observe the correlation between wave 1 delinquency and wave 4 criminal behaviors. The zero-order correlation between wave 1 and wave 4 criminal behavior was .21 and was statistically significant (p < .05, two-tailed test). The correlation indicated that respondents showed a moderate degree of stability in criminal behavior from adolescence (wave 1) to adulthood (wave 4). At the same time, however, a good deal of change occurred between the two observation periods—as evidenced by the correlation coefficient being below unity. Table 1 presents the parameter estimates from the ACE models that analyzed the wave 1 and the wave 4 criminal behavior measures separately. The table is split into two sections with the top section presenting the estimates for wave 1 delinquency and the bottom section presenting estimates for wave 4 criminal behavior. Looking first at the parameter estimates for the wave 1 delinquency measure, the table reveals that genetic influences explained a significant portion of the variance (h2 = .45). As for environmental influences, the shared environment did not have a significant impact on wave 1 delinquency (c2 = .00). Finally, the nonshared environment explained the majority of the variance in wave 1 delinquency (e2 = .55). Taken together, these findings indicate that genetic and nonshared environmental factors are important influences on wave 1 delinquency. Table 1. Univariate ACE Model Parameter Estimates Genetic Factors (A) Shared Environment (C) Nonshared Environment (E) Wave 1 Delinquency .45 .00 .55 (.39-.51) (.00-.00) (.49-.61) Wave 4 Criminal Behavior .36 .00 .64 (.23-.49) (.00-.00) (.51-.77) Note: 95% confidence interval in parentheses. Table options Turning to the ACE model results for the wave 4 criminal behavior variable (bottom portion of Table 1) we see a pattern of results that is similar to the wave 1 results. Specifically, genetic (h2 = .36) and nonshared environmental factors (e2 = .64) combined to explain all of the variance in wave 4 offending. The shared environment did not explain any of the variance in wave 4 criminal behavior (c2 = .00). Thus far, the results have indicated that genetic and nonshared environmental factors underlie delinquent behavior reported at wave 1 and criminal behavior reported at wave 4. These results are an important first step in establishing the relationship between wave 1 and wave 4 criminal behavior, but they do not reveal whether genetic and environmental factors underlie stability and changes in behavior over time. As was previously noted, there was a moderate degree of stability in delinquency from adolescence to adulthood, but there was also a high degree of change. The degree to which stability and changes in offending are influenced by genetic and environmental influences has yet to be determined. Table 2 presents parameter estimates gleaned from the bivariate Cholesky models. Parameter estimates for the stability in criminal behavior from wave 1 to wave 4 are presented separately from parameter estimates for changes in criminal behavior from wave 1 to wave 4. Upon observation of Table 2, one point is immediately obvious: genetic factors accounted for nearly all of the variance in the stability in offending. Specifically, genetic factors accounted for 97 percent of the stability in criminal behavior between wave 1 and wave 4.2 This means that criminal behavior that remained stable from wave 1 to wave 4 was almost completely due to genetic factors that influenced criminal behavior at both time points. In other words, the genetic factors that influenced wave 1 delinquency were also operating on wave 4 criminality.3 Table 2. Bivariate Cholesky Model Parameter Estimates for Stability and Change in Criminal Behavior from Wave 1 to Wave 4 Genetic Factors (A) Shared Environment (C) Nonshared Environment (E) Factors Accounting for Stability .97 .00 .03 (.65-1.00) (.00-.00) (.00-.34) Factors Accounting for Change .36 .00 .64 (.24-.49) (.00-.02) (.51-.76) Note: 95% confidence interval in parentheses. Table options Table 2 also reveals an interesting pattern of findings concerning the factors that account for changes in behavior over time. Specifically, genetic influences explained a significant portion of the changes in criminal behavior (h2 = .36). At the same time, nonshared environmental influences accounted for the majority of changes in criminal behavior (e2 = .64). 4 The final step of the analysis examined whether cross-sibling differences in wave 1 delinquency predicted wave 4 criminal behavior after controlling for genetic influences. The results from this analysis are presented in Table 3. After controlling for sources of population heterogeneity (i.e., genetic factors), cross-sibling differences in wave 1 delinquency were positively and significantly related to wave 4 criminal behavior. In other words, after controlling for genetic influences on wave 4 criminal behavior, the sibling who exhibited more delinquent behavior at wave 1 tended to report more criminal behavior at wave 4. This result is consistent with the Cholesky model results which revealed that the nonshared environment accounted for a portion (albeit, a small portion) of the stability in delinquent/criminal behavior. Table 3. Results from the DF Model Wave 4 Criminal Behavior b SE Genetic Factors .25* .06 Nonshared Environment - Wave 1 Delinquency .02* .01 *p < .05, two-tailed.