Sex Differences in Criminal Behavior

Abstract

Criminologists have long debated whether the risk factors for criminal behavior differ for males and females. Previous studies have predominantly focused on whether environmental risk factors for criminal behavior vary by gender, with little to no investigation of the impact of genetic sex differences. That is, whether the same genetic risk factors are relevant to offending for males and females and whether genetic risk factors have a stronger effect on criminal behavior for one gender (versus the other). Using data from the Add Health (140 MZ males, 135 MZ females, 124 DZ males, 118 DZ females, and 186 DZ opposite-sex twin pairs), the results from the qualitative and quantitative sex difference analyses revealed that the same genetic factors are influencing criminal behaviors in males and females and that the magnitude of the genetic effects on criminal behavior does not differ across the sexes. The implications of these findings are discussed from a biosocial approach to the study of criminal behavior.

Keywords

Add Health behavioral genetics criminal behavior gender Mx

Introduction

There is a debate within the field of criminology as to whether criminal behavior among males and females can be explained by the same theories, or whether separate theories are needed to properly explain male and female offending (Smith, Cullen, & Latessa, 2009). One body of literature suggests that gender neutral theories are adequate enough to explain offending among both males and females (Smith & Paternoster, 1987; Steffensmeier & Allan, 1996). Another body of research, however, suggests that the specific risk factors for criminal behavior are different for males and females (Mazerolle, 1998), or a specific risk factor may have a stronger effect on criminal behavior for one gender than the other (Daigle, Cullen, & Wright, 2007; Moffitt, Caspi, Rutter, & Silva, 2001). Although these studies focus on the environmental factors that are operating to explain criminal behavior in males and females, another line of research has highlighted the importance of considering genetic factors when studying antisocial behaviors among the sexes (Vaske, Wright, Boisvert, & Beaver, 2011).

Gender neutral research has highlighted the general importance of genetics in explaining variation in criminal behavior for males and females. For instance, a large body of research shows that antisocial and criminal behaviors tend to “run in the family” (Farrington, Barnes & Lambert, 1996; Farrington, Jolliffe, Loeber, Stouthhamer-Loeber & Kalb, 2001; Frisell et al., 2011; Van de Rakt, Nieuwbeerta, & de Graaf, 2008; West & Farrington, 1977). That is, individuals who exhibit antisocial behaviors are more likely to have genetically related family members who also display antisocial/criminal behaviors. In fact, a study by Farrington and his colleagues in 1996 revealed that 6% of families accounted for approximately half of all convictions. Although these results point to the intergenerational transmission of criminal offending, they are not able to decipher why these patterns of results occur. In other words, family members tend to share both their environment and their DNA; therefore, it is not possible to decipher the effects of genetic factors from those of the environment.

More advanced studies using adopted children and/or twins have been conducted to better assess the influence of genetics on antisocial behaviors. Using behavioral genetic methodology, these studies partition the variance in behavior into three components: (a) genetics, (b) shared environment, and (c) nonshared environment. The shared environment refers to the environmental factor that exerts the same effect on siblings residing within the same home (e.g., socioeconomic status, neighborhood conditions). Nonshared environmental factors, on the other hand, are factors that are unique to each individual child living within the same home (e.g., different peer networks, different parental treatment).

A number of gender neutral behavioral genetic studies have revealed that genetic factors explain approximately 50% to 70% of the variance in delinquent or criminal behavior (Christiansen, 1970, 1974, 1977; Cloninger & Gottesman, 1987; Coid, Lewis, & Reveley, 1993; Dalgard & Kringlen, 1976; Krueger, Hicks, & McGue, 2001; Lyons, 1996; Rowe, 1983, 1986; Rushton, 1996; Taylor, McGue, & Iacono, 2000a; Taylor, McGue, Iacono, & Lykken, 2000b). For example, Rowe (1983) assessed a self-reported measure of delinquency from 168 MZ and 97 same-sex DZ twin pairs between the ages of 13 and 18 years old. Their measure of delinquency included 25 items, including trespassing, breaking windows, shoplifting, and starting a fight. Their results revealed that 70% of the variance in the measure of delinquency was attributed to genetic factors with the remaining variance being attributed to the environment. Using data from the Minnesota Twin Registry, Krueger et al. (2001) also examined the genetic and environmental contributions to criminal behaviors in a sample of adults aged 33 years old (N = 397 twin pairs). Participants self-reported their level of involvement in 27 criminal acts, such as theft, illegal drug and alcohol use, and fighting. The results revealed that 52% of the variance in the measure of criminality was due to genetic factors with the remaining 48% being due to nonshared environmental factors. Altogether, it is undeniable that both genetic and environmental factors influence antisocial and criminal behaviors in particular. Although these results point to the importance of genetic and nonshared environmental factors in the explanation of criminal behaviors, these studies did not address whether genetic and environmental factors are equally relevant to criminal behavior for males and females.

Genetic Sex Differences

To date, few studies have empirically evaluated whether genetic and environmental factors are related to antisocial, delinquent, and criminal behaviors for males and females. Furthermore, the few behavioral genetic studies that have been conducted on antisocial behaviors have produced mixed results. On one hand, studies have found that genetic factors are stronger in males than in females for externalizing behaviors (Silberg et al., 1994; van den Oord, Boomsma, & Verhulst, 1994), aggression (Hudziak et al., 2003; Miles & Carey, 1997), and antisocial behavior (Graham & Stevenson, 1985). For example, Silberg and colleagues in 1994 assessed whether there were quantitative sex differences in externalizing behaviors in a sample of 1,263 twin pairs aged 8-11 years old. They found that the genetic effects on externalizing behaviors were significantly higher in males (0.38) compared to females (0.13).

On the other hand, studies have found that the magnitude of the genetic effects are actually stronger in females than in males for measures of aggression (Vierikko, Pulkkinen, Kaprio, Viken, & Rose, 2003), hyperactivity (van den Oord, Verhulst, & Boomsma, 1996), and nonaggressive delinquent behaviors (Eley, Lichtenstein, & Stevenson, 1999). For example, using a sample of 1,651 twin pairs from the Finnish Twin and Family Study (FinnTwin12), Vierikko et al. (2003) found significantly lower heritability estimates for youth’s aggression in males (14%-27%) compared to females (54%-62%). Even still, other studies have found no sex difference in the genetic influences on externalizing behaviors (van den Oord et al., 1996), aggression (Eley et al., 1999), conduct disorders (Slutske et al., 1997), and aggressive delinquent behaviors (Eley et al., 1999).

In addition, very few studies have examined whether there are genetic sex differences in actual measures of delinquency (see Eley et al., 1999; Taylor et al., 2000b). One such study by Taylor and colleagues (2000b) examined sex differences in the genetic and environmental contributions to delinquent behavior. Their results revealed that genetic factors accounted for 55% of the variance in delinquency in males with the remaining 45% attributed to nonshared environmental factors. On the other hand, 8%, 37%, and 55% of the variance in delinquency in females was attributed to genetic, shared, and nonshared environmental factors, respectively. Although it appears that the genetic effects on delinquency are stronger in males, this difference did not reach statistical significance. Another study conducted by Eley et al. (1999) found that the heritability of nonaggressive forms of delinquency were stronger in females compared to males. In regards to aggressive forms of delinquency, Eley et al. (1999) found no significant sex differences in the magnitude of genetic effects across the sexes. Overall, these inconsistent results call for further examination of the role that genetic and environmental factors may play in explaining delinquent and criminal behaviors for males and females across the life course.

Purpose of Current Study

There is no doubt that genetic factors play a role in the etiology of a variety of antisocial behaviors, including criminality (Moffitt, 2005; Rhee & Waldman, 2002). It is less clear, however, whether these genetic influences on delinquent and criminal behaviors differ across the sexes. To date, studies on the sex differences in the genetic effects on antisocial behaviors have produced mixed results (Eley et al., 1999; Hudziak et al., 2003; Silberg et al., 1994; Slutske et al., 1997; van den Oord et al., 1996; Vierikko et al., 2003) and very few studies have examined whether there are genetic sex differences in actual measures of criminal behavior (Eley et al., 1999; Frisell et al., 2011; Taylor et al., 2000a). As such, the current study adds to the current body of literature by examining whether there are sex differences in the genetic contributions to criminal behavior in a sample of adolescents at three waves of data collection. Specifically, the current study examines two questions pertaining to genetic sex differences in criminal behavior: (a) are the same genetic factors influencing criminal behavior in males and females (i.e., qualitative sex differences) and (b) does the magnitude of the genetic effect on criminal behavior differ across the sexes (i.e., quantitative sex differences)?

Method

Sample

Data used in the current study are from the National Longitudinal Study of Adolescent Health (Add Health), a nationally representative study of American adolescents in grades seven through twelve (Udry, 2003). This longitudinal study began in September 1994 when the participants were between the ages of 11 to 19. Using a multistage stratified random sampling technique, 132 consenting middle and high schools were included in the study, which resulted in over 90,000 students completing the in-school questionnaire (Harris et al., 2003). At this stage, respondents were asked if they had any siblings that were also enrolled in grades seven through twelve. If they responded affirmatively, their sibling(s) were then automatically included in the wave I in-home interviews to oversample for siblings, of varying degrees of genetic relatedness, residing within the same home.

Overall, 20,745 youths participated in the in-home interview between April and December 1995. Most of the interviews were conducted in the interviewees’ home and lasted between 1 to 2 hours. A computer assisted personal interviewing (CAPI) technique was used, which involved questions being read to the interviewee and the interviewer entering their responses directly into the computer. For certain sensitive topics, such as criminal involvement, the verbal exchange between interviewer and interviewee was modified to increase the likelihood that the interviewee would respond truthfully. For those types of questions, the interviewer employed the audio computer assisted self-interviewing (ACASI) technique, which allowed participants to both privately listen to prerecorded questions through headphones and enter their responses directly into the computer themselves (Couper, Singer, & Tourangeau, 2003; Ghanem, Hutton, Zenilman, Zimba, & Erbelding, 2005).

Approximately 1 year later, between April 1995 and August 1996, the second wave of data was collected when the participants were between the ages of 12 to 20. A total of 14,738 respondents completed the wave II in-home survey (Harris et al., 2003). Similar to wave I, CAPI and ACASI were both used to interview the respondents about a variety of topics, such as physical and emotional health, peer relationships, family dynamics, school-related activities, employment, substance use, and criminal involvement. The third wave of data collection occurred between July 2001 and April 2002 and included 15,170 of the participants who originally participated in the wave I in-home interview. Once again, the interview lasted approximately an hour and a half and was administered using CAPI and ACASI (Harris et al., 2003). As most of the respondents were now young adults, between the ages of 18 to 26, the questions were adjusted to reflect their new life experiences, while simultaneously trying to obtain information regarding transitional experiences from adolescence to adulthood. Topics such as marriage, employment, education, childrearing, and criminal behavior occurring at this stage of development were asked in the wave III questionnaire.¹

Analytical Sample

Only participants identified as twins were considered in the present study (N = 784 twin pairs). Zygosity was determined using both self-reported questions and DNA analysis. First, in waves I and II, twins self-reported their zygosity using four questions related to similarities in physical characteristics and identity confusion. Monozygotic (MZ) twins were more likely to report that they possess similar physical characteristics and/or experience identity confusion compared to dizygotic (DZ) twins. Next, in wave III, samples of buccal cells were collected for DNA analysis. This procedure allowed researchers to determine the zygosity of same-sex twin pairs whose zygosity remained uncertain from the self-reported questionnaires. To positively identify zygosity for these twins, researchers compared the DNA from twin 1 and twin 2 to determine their match on 12 unlinked short tandem repeats (STR). If a twin pair was identical for five or more genetic markers (error rate of < 0.0004), they were classified as monozygotic twins. On the other hand, if the twin pair differed on one or more genetic markers, they were classified as dizygotic twins. The results from the DNA analysis demonstrated that 92% of the participants had correctly self-identified their zygosity.

Together, these two methods correctly identified 307 MZ and 452 DZ twin pairs, leaving 25 twin pairs with uncertain zygosity (Harris, Halpern, Smolen, & Haberstick, 2006). For the purposes of this study, twins with uncertain zygosity were classified as dizygotic twins. This is a more conservative approach as it increases the likelihood of overestimating the environmental effects on behavior. Of these twin pairs, only those with known sex were included in the analyses. This resulted in a final analytic sample of 1,406 individual twins (N = 703 twin pairs). Specifically, this included 140 MZ male twin pairs (MZm), 135 MZ female twin pairs (MZf), 124 DZ male twin pairs (DZm), 118 DZ female twin pairs (DZf), and 186 DZ opposite-sex twin pairs (DZo; Harris et al., 2006). If a twin pair was of the same sex (i.e., MZm, DZm, MZf, DZf), each twin was randomly assigned as either twin 1 or twin 2. On the other hand, for opposite-sex DZ twins (i.e., DZo), all females were assigned as twin 1 while all males were assigned as twin 2.

Measures

Criminal Behavior

Criminal behavior was assessed at all three waves by asking respondents how often in the past 12 months they had engaged in 14 different criminal activities (see Appendix A). These questions, ranging from nonviolent forms of criminal behaviors (e.g., stealing, selling marijuana, and damaging other’s property) to more violent and serious types of criminal activities (e.g., shot/stab someone, physical assault), were identical in waves I and II. At wave III, however, four questions were replaced to reflect more age-appropriate questions pertaining to criminal involvement by young adults. For example, the question asking respondents whether they ever drove a car without the owner’s permission was replaced with a question asking whether they ever used someone else’s credit/bank card without their permission. Overall, at all three waves, participants were asked to score their criminal behaviors by choosing an answer that best described their level of criminal involvement, which ranged from 0 (never) to 3 (five or more times). These 14 items were then summed together to create the criminal behavior measures at all three waves, with higher scores indicating greater criminal involvement (wave I alpha = .84; wave II alpha = .81; wave III alpha = .73).

As the distributions for the measures of criminal behavior at all three waves were positively skewed, with most values falling toward zero, the measures were log transformed (log(x+1)) prior to being included in the analyses. Before transforming the data, the level of criminal involvement at waves I, II, and III ranged from 0 to 40 (mean = 2.48; SD = 4.22), 0 to 37 (mean = 1.56; SD = 3.14), and 0 to 22 (mean = 0.69; SD = 1.83), respectively. These measures of criminal behaviors derived from the Add Health data have been used in previous research (Haynie, 2001, 2002; Haynie & Osgood, 2005; Haynie, Giordano, Manning, & Longmore, 2005; Pearce & Haynie, 2003).

Analytical Plan

First, intraclass correlations were created for MZ and DZ twins separately to determine whether genetic influences are operating on criminal behaviors in general. If MZ twin pairs have a higher intraclass correlation compared to DZ twin pairs, this indicates that genetic effects are influencing the behavior. Next, intraclass correlations were created for the five zygosity groups (MZm, MZf, DZm, DZf, and DZo) to assess whether genetic effects differ across the sexes. If the pattern of intraclass correlations differs across males and females, this suggests that the genetic effects may be stronger/weaker for one gender compared to the other (i.e., quantitative sex differences). Sex-specific genetic effects on a behavior can also occur when the genetic effects operate in one sex but not the other (i.e., qualitative sex differences). These effects are noticeable when the intraclass correlation of opposite-sex DZ twin pairs is lower than the average correlation for same-sex DZ twin pairs (Neale & Cardon, 1992).

While examining intraclass correlations is a useful way to tentatively determine the type and magnitude of genetic influences on criminal behavior, this technique is limited. A more statistically advanced approach, known as a sex limitation model, is preferred in behavioral genetic research. This SEM approach incorporates variance–covariance matrices between MZ and DZ twin pairs as well as information from all five zygosity groups to more accurately explore whether qualitative and/or quantitative genetic sex differences are operating on criminal behavior. An illustration of the full sex limitation model is presented in Figure 1. This model is simply an extension of the univariate ACE model, where the genetic correlation (r _g) for MZ and DZ twins is 1 and 0.5, respectively, the shared environmental correlation (r_c) is 1 and the nonshared environment (r_e) is uncorrelated between twins.

Figure 1.

Behavioral genetic sex limitation model

The inclusion of same-sex and opposite-sex twin pairs allows for the examination of both qualitative and quantitative genetic sex differences. When testing for qualitative sex differences, r _g is free to vary between 0 and 0.50 for opposite-sex DZ twin pairs. If this estimate is less than 0.50, this implies that different genes are influencing criminal behavior for males and females (Neale & Cardon, 1992). As such, a difference in χ² test that compares the full sex-limitation model to the submodel that fixes r_g to 0.50 for opposite-sex DZ twins is conducted. If fixing r_g to 0.50 significantly reduces the model fit, this implies that there are different genes influencing criminal behavior across the sexes. Conversely, if fixing r_g to 0.50 does not significantly reduce the model fit, this suggests that the same genes are affecting criminal behavior in both males and females (Neale & Cardon, 1992).

Next, to test for quantitative sex differences, the full sex limitation model, which allows the parameter estimates to vary across the sexes, is compared to a submodel that constrains the parameter estimates to be equal (i.e., a_m = a_f) across the sexes (Neale & Cardon, 1992). If this process significantly reduces the fit of the model, then it is suggested that the strength of the genetic effects significantly differs across the sexes. Conversely, if setting the genetic parameters to be equal does not significantly reduce the fit of the model, then it is concluded that the strength of the genetic effects is the same for males and females (Neale & Cardon, 1992).

Results

The intraclass correlations for MZ and DZ twins as well as the five zygosity groups (MZm, MZf, DZm, DZf, and DZo) are presented in Table 1. First, the results show that MZ twins had higher correlations than DZ twins, suggesting that genetic factors are influencing criminal behavior across the three time periods. For males, the MZ and DZ correlations ranged from 0.33 to 0.57 and 0.25 to 0.43, respectively. For females, the MZ and DZ correlations ranged from 0.32 to 0.57 and 0.18 to 0.43, respectively. Again, the higher correlations found in MZ twins compared to DZ twins in both males and females suggest that genetic factors are influencing criminal behavior. Heritability estimates were then calculated for males and females separately [h² = 2(MZr – DZr)], with estimates ranging from 0.00 to 0.54 in males and 0.18 to 0.28 in females (Table 1). These patterns of results suggest that the strength of the genetic effects may differ across the sexes (i.e., quantitative genetic sex differences). Furthermore, an examination of the intraclass correlations for the opposite-sex DZ twin pairs (i.e., 0.24, 0.22, 0.08) reveals that they are smaller than the average correlations for same-sex DZ twin pairs (i.e., 0.37, 0.41, 0.22). This suggests that there may be different genes influencing criminal behaviors in males and females (i.e., qualitative genetic sex differences).

Table 1.

Intraclass Correlations and Crude Heritability Estimates of Criminal Behaviors for Males and Females [h² = 2(MZr – DZr)]

	MZ			DZ				Heritability
Criminal Behavior	All	Males	Females	All	Males	Females	Mixed	All	Males	Females
Wave I	.58***	.57***	.57***	.32***	.30**	.43***	.24**	.52	.54	.28
Wave II	.40***	.33***	.48***	.33***	.43***	.39***	.22**	.14	.00	.18
Wave III	.34***	.34**	.32**	.22***	.25*	.18	.08	.24	.18	.28

p < .05, **p < .01, ***p < .001.

Table 2 displays the results from the genetic analyses on criminal behavior at the three time periods by presenting four models: (1) the full saturated sex limitation ACE model, (2) the qualitative sex differences model, which fixes the genetic correlation between males and females in the DZ twin pairs to 0.50, (3) the quantitative sex differences model, which equates the genetic parameters across males and females, and (4) the restricted AE submodel. Model 1 presents the −2 log likelihood values and the degrees of freedom for the full sex limitation model for criminal involvement at wave I (−2lnl = 1159.36, df = 1,363), wave II (−2lnl = 962.75, df = 1,250), and wave III (−2lnl = −158.94, df = 1095). In model 2, the genetic correlation between males and females is fixed to 0.50 and compared to model 1. The results reveal a nonsignificant reduction in fit compared to model 1 at wave I (Δχ² = 2.58, p = .11), wave II (Δχ² = 0.15, p = 0.70), and wave III (Δχ² = 0.004, p = .95). These results suggest that the genetic influences on criminal behavior are the same in males and females. In model 3, the genetic parameters are equated across gender and the χ² difference tests reveal a nonsignificant reduction in fit at wave I (Δχ² = −0.70, p = 1.00), wave II (Δχ² = 3.91, p = .05), and wave III (Δχ² = 2.65, p = .10). In other words, the magnitude of the genetic effects on criminal behavior appears to be the same for males and females. Model 4 tested the significance of the shared environmental effects on criminal behavior by setting C to zero. The results revealed that the AE submodel is the best fitting model at wave I (Δχ² = 1.53, p = .47) and wave II (Δχ² = 5.78, p = .06). For wave III, however, dropping the shared environmental effects significantly reduced the fit of the model (Δχ² = 6.82, p = .03).

Table 2.

Sex Effects for Criminal Behavior at Waves I, II, and III

		Model fit			Change statistics				Females			Males
Models		–2lnl	DF	vs	Δχ2	Δdf	p	AIC	A	C	E	A	C	E
Wave I
1	Full	1159.36	1363						.33 (.00-.66)	.26 (.00-.53)	.41 (.32-.53)	.44 (.08-.63)	.09 (.00-.40)	.47 (.37-.58)
2	Drop F	1161.94	1364	1	2.58	1.00	.11	.58	.61 (.00-.69)	.00 (.00-.58)	.39 (.30-.57)	.42 (.00-.65)	.11 (.00-.46)	.47 (.35-.60)
3	af = am	1161.24	1365	2	−.70	1.00	1.00	−2.70	.53 (.29-.65)	.09 (.00-.30)	.38 (.30-.49)	.52 (.27-.60)	.00 (.00-.22)	.48 (.39-.59)
4	AE	1162.77	1367	3	1.53	2.00	.47	−2.48	.60 (.51-.67)	—	.41 (.33-.49)	.55 (.47-.62)	—	.45 (.38-.53)
Wave II
1	Full	962.75	1250						.30 (.01-.57)	.21 (.01-.46)	.48 (.38-.62)	.01 (.00-.30)	.35 (.10-.46)	.63 (.52-.75)
2	Drop F	962.90	1251	1	0.15	1.00	.70	−1.85	.34 (.07-.57)	.18 (.00-.39)	.48 (.37-.61)	.00 (.00-.22)	.36 (.15-.46)	.64 (.53-.75)
3	af = am	966.81	1252	2	3.91	1.00	.05	1.91	.20 (.00-.47)	.28 (.05-.49)	.52 (.40-.67)	.18 (.00-.43)	.19 (.00-.39)	.63 (.52-.77)
4	AE	972.56	1254	3	5.78	2.00	.06	1.78	.50 (.40-.59)	—	.50 (.41-.60)	.44 (.35-.51)	—	.56 (.49-.65)
Wave III
1	Full	−158.94	1095						.00 (.00-.41)	.30 (.00-.42)	.70 (.57-.85)	.41 (.00-.56)	.01 (.00-.32)	.58 (.44-.77)
2	Drop F	−158.94	1096	1	.004	1.00	.95	−2.00	.00 (.00-.33)	.30 (.00-.42)	.70 (.57-.85)	.41 (.18-.56)	.01 (.00-.15)	.58 (.44-.76)
3	af = am	−156.30	1097	2	2.65	1.00	.10	.65	.30 (.12-.43)	.00 (.00-.13)	.70 (.57-.85)	.14 (.06-.22)	.20 (.05-.33)	.66 (.54-.80)
4	AE	−149.48	1099	3	6.82	2.00	.03	2.82	.35 (.22-.47)	—	.65 (.53-.78)	.18 (.11-.25)	—	.82 (.75-.89)

Overall, the results revealed that 55%-60% of the variance in criminal behavior at wave I was attributed to genetic effects and that the remaining 40%-55% was due to nonshared environmental effects. At wave II, genetic factors accounted for 44%-50% of the variance and nonshared environmental effects accounted for 50%-56% of the variance in criminal behavior. At wave III, 14%-30%, 0%-20%, and 66%-70% of the variance in criminal behaviors was attributed to genetic, shared, and nonshared environmental effects, respectively. Altogether, the results suggest that the same genetic factors are influencing criminal behaviors in males and females and that the magnitude of the genetic effects on criminal behavior does not differ across gender.

Discussion

It is well documented that males exhibit higher levels of antisocial behaviors compared to females (Achenbach et al., 1990; Farrington, 1986; Junger-Tas, Terlouw, & Klein, 1994; Kessler et al., 1994; Moffitt et al., 2001; Rutter, 2003; Stattin, Magnusson, & Reichel, 1989). However, very little is known about the mechanisms that underlie these sex differences. Many criminologists have focused on the varying effects that environmental factors have on antisocial behaviors for males and females. The current study takes a behavioral genetic approach to this question by examining whether the same genes are influencing criminal behavior in males and females and whether the strength of the genetic effect differs for males and females.

First, the results from the current study, along with many others, demonstrate that the etiology of antisocial behavior depends on both genetic and environmental factors. Specifically, the heritability of criminal behavior reported here ranged from .44 to .60 in adolescence and .14 to .30 in adulthood. Although these analyses provide an overall estimate of the genetic influence on criminal behavior, it is important to remember that the results are not specifying which genes are influencing criminal behaviors. There are likely multiple genes interacting with each other and simultaneously interacting with the environment to influence behaviors. It is important to note that a genetic predisposition to a behavior/trait increases the likelihood that the behavior/trait will occur, but it is in no way a guarantee. For example, intelligence is a highly heritable trait with estimates at approximately .80. If an individual is predisposed for higher intelligence but is not cognitively stimulated or provided the necessary tools to excel, then the likelihood that he or she will reach his or her maximum intellectual potential substantially decreases. In other words, whether or not an individual displays the heritable behavior of interest, such as criminal behavior, will depend on a host of other factors. In addition, it has been suggested that the genetic effects on antisocial behaviors are likely to operate indirectly through levels of self-control (Boisvert et al., in press), impulsivity (Lahey, Moffitt, & Caspi, 2003), hormones (Pompa & Raine, 2006), and/or intelligence (Moffitt, 1993). For example, a recent study by Boisvert et al. (2011) revealed that the covariation between low self-control and criminal behavior appears to be largely due to the same genetic and nonshared environmental factors operating on both phenotypes.

One of the main purposes of the current study was to assess whether the genetic effects on criminal behavior differed for males and females. Some researchers have argued that genetic and environmental effects on criminal behaviors differ across the sexes (Jacobson, Prescott, & Kendler, 2002; Vierikko et al., 2003). From a genetic perspective, it has been suggested that a larger genetic effect is present in males making them at higher risk for criminal behaviors, which could help to explain their higher prevalence of criminal involvement compared to females. Conversely, others have argued that genetic effects should be stronger in females because the genes that are involved in puberty have a stronger impact on adolescent females given their earlier age of pubertal onset (Jacobson et al., 2002). From an environmental perspective, it has been suggested that boys are more susceptible to certain risk factors, especially related to the family (Moffitt et al., 2001; Rutter, Giller, & Hagell, 1998) and are exposed to more risk factors compared to females, particularly peers (Rutter, 2003).

The results from the current study, however, found no significant differences in the type or magnitude of the genetic effect across the sexes. For females, the heritability of delinquent and criminal behavior was .60 in wave I, .50 in wave II, and .35 in wave III. For males, the heritability coefficient was .55 in wave I, .44 in wave II, and .14 in wave III. The analyses revealed that constraining the heritability coefficients to be similar for females and males at the multiple waves did not significantly reduce the model’s fit; thus, suggesting that the heritability coefficients are statistically equivalent. Similar null findings have been reported for other measures of antisocial behaviors, such as externalizing behaviors (van den Oord et al., 1996), aggression (Eley et al., 1999), conduct disorders (Slutske et al., 1997), aggressive delinquent behaviors (Eley et al., 1999), and delinquency (Taylor et al., 2000a). It is quite possible that the small number of twin pairs included in this study may have biased the results by lacking the power to detect significant sex differences in the genetic contributions to criminal behaviors. For example, it is evident from the large confidence intervals reported in the results that a larger number of twin pairs are needed to increase statistical power. Given that other studies have provided evidence to suggest that genetic sex differences are present in antisocial behaviors (Eley et al., 1999; Graham & Stevenson, 1985; Hudziak et al., 2003; Miles & Carey, 1997; Silberg et al., 1994; van den Oord et al., 1994; Vierikko et al., 2003), future studies including a greater number of twin pairs are needed to further examine this research question. It is also important to note that these varying results reported in the literature may be a reflection of the different ways that antisocial behaviors are defined across the disciplines.

Related to the literature on gender, genetics, and criminal behavior, the current study suggests that future research should focus on identifying specific genetic polymorphisms that are relevant to both male and female criminal behavior, or gender neutral polymorphisms. This finding extends upon previous criminological studies that show that male and female criminal behavior is a function of gender neutral environmental factors (Smith & Paternoster, 1987), and it suggests that the genetic risk factors for criminal behavior may also be gender neutral. Therefore, if the genetic risk factors are gender neutral, researchers may focus initially on polymorphisms linked to autosomes and the X chromosome. Such polymorphisms may include a host of factors that influence neuropsychological functioning, such as monoamine oxidase A and dopaminergic polymorphisms (Boisvert & Vaske, 2011). This hypothesis, however, would need to be more thoroughly investigated. Although the current study suggests that the same genetic factors explain variation in criminal behavior, it is important to note that males and females may have the same genetic polymorphism, but the expression level may significantly vary across genders. Differences in expression may be because of gender differences in sex hormones, the number of “active” X chromosomes, or various other factors (Ostrer, 2001). Thus, researchers may better understand the relationship between genetic polymorphisms and criminal behavior among males and females if they search for genetic polymorphisms and explore whether there are gender differences in the expression levels. One potential avenue for exploring the questions of gender differences/similarities in genetic risks and expression levels may be through the use of genome-wide association studies and microarray studies.

Limitations

There are at least three limitations to the current study in addition to the small sample size. First, the use of SEM models applied to twin data involves three basic underlying assumptions: (a) no assortative mating, (b) no gene–environment interactions, and (c) normally distributed data. Assortative mating refers to the nonrandom manner in which individuals choose mates. Essentially, individuals tend to select sexual partners that have similar characteristics as themselves on some physical, cognitive, and/or personality trait(s). Research has shown that assortative mating can occur for antisocial behaviors. In other words, antisocial individuals tend to mate with other antisocial individuals (Caspi & Herbener, 1990; Quinton, Pickles, Maughan, & Rutter, 1993). Therefore, it is possible that our analyses violated the assumption of nonrandom mating. The effects of nonrandom mating lead to increased similarities between DZ twins relative to MZ twins. From a statistical standpoint, this would lead to an underestimation of the genetic effects and overestimation of the shared environmental effects (Neale & Cardon, 1992). However, Maes et al. (1998) have argued that the overall levels of assortative mating are quite low and that the biases associated with them are generally very small. In any event, future studies should use an extended-kinship study design that incorporates information from both twins and parents to explicitly test for the effects of assortative mating (Neale & Cardon, 1992).

The second assumption underlying the use of SEM with twin data is that there are no gene–environment interactions occurring. Gene–environment interactions refer to situations in which an individual’s genes interact with his or her environment to produce a behavior. Research has shown that gene–environment interactions are involved in the prediction of antisocial and criminal behaviors in both juveniles and adults (Caspi et al., 2002; Foley et al., 2004). The current analyses, however, did not specifically model for the effects of gene–environment interactions. It is important to note that all twins included in the current analyses were raised together. As Heath et al. (2002) point out, in these types of models, where twins are raised together and gene–environment interactions are not specifically modeled, the effects of A × E will be confounded with E and the effects of A × C will be confounded with A.

The final assumption when using SEM with twin data is that the data is normally distributed. As mentioned in the methods section, the measures of criminal behavior included in this study where positively skewed, with most values falling toward zero, implying no criminal involvement. Although we log transformed the data prior to conducting the analyses, this did not completely eliminate the skewness of the data. We recommend that future studies not only include a larger sample of twins but also include a more a high-risk sample of twins. Having a higher proportion of respondents involved in criminal activity will help to meet the normality assumption for these types of analyses.

The next limitation involves the use of self-reported measures of criminal involvement. The reliability of self-reported questionnaires has been called into question, especially when the questions are sensitive and/or incriminating. Participants may be more fearful to respond honestly to these types of questions, which leads to an underreporting of these behaviors. Although the audio computer assisted self-interviewing (ACASI) technique was used to increase the reliability of the results, we recommend that future studies use multiple raters, such as peers, partners, and siblings. These additional reporting sources can provide important information about the participant’s behavior from varying perspectives (Achenbach, McConaughy & Howell, 1987; Bartels et al., 2003; Ronald, Happe, & Plomin, 2005; Van der Ende & Verhulst, 2005).

Last, although males exhibit higher levels of antisocial behaviors compared to females and this sex difference is greatest for violent types of offending (Rutter et al., 1998; Smith & Visher, 1980), our small sample size inhibited us from disaggregating our measure of criminal behavior into violent and nonviolent types of behaviors. Specifically, only 250 same-sex female twin pairs (N = 250) are included in the sample, of which only a small number had reported committing at least one violent criminal act. A recent study by Frisell et al. (2011), however, revealed significant gender differences in the heritability of violent crime. These authors also call for larger samples of female offenders to better capture whether genetic sex differences operate differently based on the level of seriousness of the offense.

Despite the limitations presented above, the current study is part of a recent trend in criminology that is taking a multidisciplinary approach to the study of antisocial behavior. Following a biosocial approach, we are moving beyond simply asking whether antisocial and criminal behavior is heritable. That question has been answered by hundreds of researchers from across many disciplines. It is now time to further examine the role that genetics plays in the etiology of criminal behaviors across the sexes and across the life course. Although the current study examined criminal involvement at three waves of data, it did so univariately. The next step is to examine longitudinal models that allow the genetic effects on behavior at time 1 to influence behavior at time 2, and so on and to assess whether those genetic influences differ for males and females. Having a better understanding of how antisocial behaviors develop from early childhood through to adulthood across the sexes has the potential to greatly influence our prevention and intervention efforts.

Footnotes

Appendix A

Acknowledgements

Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health website (). No direct support was received from grant P01-HD31921 for this analysis. This research uses data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research and/or authorship of this article.

Notes

Bios

Danielle Boisvert, PhD (University of Cincinnati), is an Assistant Professor in the College of Criminal Justice at Sam Houston State University. Her research interests include life-course/developmental criminology, biosocial criminology, and behavioral genetics.

Jamie Vaske, PhD, is an Assistant Professor at Western Carolina in the Department of Criminology and Criminal Justice. Her research interests include biosocial criminology, gender and crime, and quantitative methods and analyses.

John Paul Wright is Associate Professor of Criminal Justice in the School of Criminal Justice at the University of Cincinnati. He has published broadly on biological influences and criminal behavior.

Dr. Valerie Knopik is Director of the Division of Behavioral Genetics and an Associate Professor (Research) in the Departments of Psychiatry and Human Behavior and Community Health at Brown University. She received her Ph.D. in Behavioral Genetics from the Institute for Behavioral Genetics and the University of Colorado, Boulder in 2000. She subsequently completed a postdoctoral fellowship in psychiatric genetics and genetic epidemiology at Washington University School of Medicine in St Louis from 2000-2003 and continued on as junior faculty in the Department of Psychiatry at Washington University School of Medicine from 2003-2004. Dr. Knopik joined the faculty at Brown University in 2004.

References

Achenbach

T. M.

McConaughy

S. H.

Howell

C. T.

(1987). Child/adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychological Bulletin, 101, 213-232.

Achenbach

T. M.

Bird

H. R.

Canino

Phares

Gould

M. S.

Rubio-Stipec

(1990). Epidemiological comparisons of Puerto Rican and US mainland children: Parent, teacher, and self-reports. Journal of the American Academy of Child and Adolescent Psychiatry, 29(1), 84-93.

Bartels

Hudziak

J. J.

Van den Oord

E. J. C. G.

van Beijsterveldt

C. E. M.

Rietveld

M. J. H.

Boomsma

D. I.

(2003). Co-occurrence of aggressive behavior and rule-breaking behavior at age 12: Multi-rater analyses. Behavior Genetics, 33, 607-621.

Boisvert

Vaske

(2011). Genes, twin studies, and antisocial behavior. In Somnit

Peterson

S. A.

(Eds.), Biology and politics: The cutting edge. Bingley, UK: Emerald.

Boisvert

Wright

J. P

Knopik

Vaske

(in press). Genetic and environmental overlap between low self-control and delinquency. Journal of Quantitative Criminology.

Caspi

Herbener

E. S.

(1990). Continuity and change: Assortative marriage and the consistency of personality in adulthood. Journal of Personality and Social Psychology, 58, 250-258.

Caspi

McClay

Moffitt

T. E.

Mill

Martin

Craig

I. W.

Taylor

Poulton

(2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851-854.

Christiansen

K. O.

(1970). Crime in a Danish twin population. Acta Geneticae Medicae Et Gemellologiae, 19, 323-326.

Christiansen

K. O.

(1974). Seriousness of criminality and concordance among Danish twins. In Hood

(Ed.), Crime, criminology and public policy (pp. 63-77). London, UK: Heinemann.

10.

Christiansen

K. O.

(1977). A preliminary study of criminality among twins. In Mednick

S. A.

Christiansen

K. O.

(Eds.), Biosocial bases of criminal behavior (pp. 89-108). New York, NY: Gardner Press.

11.

Cloninger

C. R.

Gottesman

I. I.

(1987). Genetic and environmental factors in antisocial behavior disorders. In Mednick

S. A.

Moffitt

T. E.

Stack

S. A.

(Eds.), The causes of crime: New biological approaches (pp. 92-109). New York, NY: Cambridge University Press.

12.

Coid

Lewis

S. W.

Reveley

A. M.

(1993). A twin study of psychosis and criminality. British Journal of Psychiatry, 162(1), 87-92.

13.

Couper

M. P.

Singer

Tourangeau

(2003). Understanding the effects of audio-CASI on self-reports of sensitive behavior. Public Opinion Quarterly, 67, 385-395.

14.

Daigle

L. E.

Cullen

F. T.

Wright

J. P.

(2007). Gender differences in the predictors of juvenile delinquency: Assessing the generality-specificity debate. Youth Violence and Juvenile Justice, 5, 254-286.

15.

Dalgard

O. S.

Kringlen

(1976). A Norwegian twin study of criminality. British Journal of Criminology, 16, 213-232.

16.

Eley

T. C.

Lichtenstein

Stevenson

(1999). Sex differences in the etiology of aggressive and nonaggressive antisocial behavior: Results from two twin studies. Child Development, 70, 155-168.

17.

Farrington

D. P.

(1986). Age and crime. In Tonry

Norris

(Eds.), Crime and justice: A review of the research (pp. 189-250). Chicago, IL: University of Chicago Press.

18.

Farrington

D. P.

Barnes

G. C.

Lambert

(1996). The concentration of offending in families. Legal and Criminological Psychology, 1, 47-63.

19.

Farrington

D. P.

Jolliffe

Loeber

Stouthamer-Loeber

Kalb

(2001). The concentration of offenders in families, and family criminality in the prediction of boys’ delinquency. Journal of Adolescence, 24, 579-596.

20.

Foley

D. L.

Eaves

L. J.

Wormley

Silberg

J. L.

Maes

H. H.

Kuhn

Riley

(2004). Childhood adversity, monoamine oxidase: A genotype, and risk for conduct disorder. Archives of General Psychiatry, 21, 738-744.

21.

Frisell

Lichtenstein

Langstrom

(2011). Violent crime runs in families: A total population study of 12.5 million individuals. Psychological Medicine, 41(1), 97-105.

22.

Ghanem

K. G.

Hutton

H. E.

Zenilman

J. M.

Zimba

Erbelding

E. J.

(2005). Audio computer assisted self interview and face to face interview modes in assessing response bias among STD clinic patients. Sexually Transmitted Infections, 81, 421-425.

23.

Graham

Stevenson

(1985). A twin study of genetic influences on behavioral deviance. Journal of the American Academy of Child Psychiatry, 24(1), 33-41.

24.

Harris

K. M.

Halpern

C. T.

Smolen

Haberstick

B. C.

(2006). The national longitudinal study of adolescent health (Add Health) twin data. Twin Research and Human Genetics, 9, 988-997.

25.

Harris

K. M.

Florey

Tabor

Bearman

P. S.

Jones

Udry

J. R.

(2003). The national longitudinal study of adolescent health: Research design. Retrieved from http://www.cpc.unc.edu/projects/addhealth/design

26.

Haynie

D. L.

(2001). Delinquent peers revisited: Does network structure matter. American Journal of Sociology, 106, 1013-1057.

27.

Haynie

D. L.

(2002). Friendship networks and delinquency: The relative nature of peer delinquency. Journal of Quantitative Criminology, 18, 99-134.

28.

Haynie

D. L.

Giordano

P. C.

Manning

W. D.

Longmore

M. A.

(2005). Adolescent romantic relationships and delinquent involvement. Criminology, 43, 177-210.

29.

Haynie

D. L.

Osgood

D. W.

(2005). Reconsidering peers and delinquency: How do peers matter. Social Forces, 84, 1109-1130.

30.

Heath

A. C.

Madden

P. A. F.

Bucholz

K. K.

Nelson

E. C.

Todorov

A. A.

Price

R. K.

. . . Martin

N. G.

(2002). Genetic and environmental risks of dependence on alcohol, tobacco and other drugs. In Plomin

Craig

DeFries

McGuffin

(Eds.), Behavioral genetics in the postgenomic era. Washington, DC: American Psychological Association.

31.

Hudziak

J. J.

van Beijsterveldt

C. E. M.

Bartels

Rietveld

M. J. H.

Rettew

D. C.

Derks

E. M.

Boomsma

D. I.

(2003). Individual differences in aggression: Genetic analyses by age, gender, and informant in 3-, 7-, and 10-year-old Dutch twins. Behavior Genetics, 33, 575-589.

32.

Jacobson

K. C.

Prescott

C. A.

Kendler

K. S.

(2002). Sex differences in the genetic and environmental influences on the development of antisocial behavior. Developmental Psychopathology, 14, 395-416.

33.

Junger-Tas

Terlouw

G. J.

Klein

M. W.

(1994). Delinquent behavior among young people in the western world: First results of the international self-report delinquency study. Amsterdam, The Netherlands: Kugler.

34.

Kessler

R. C.

McGonagle

K. A.

Zhao

Nelson

C. B.

Hughes

Eshleman

Kendler

K. S.

(1994). Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: Results from the national comorbidity survey. Archives of General Psychiatry, 51(1), 8-19.

35.

Krueger

R. F.

Hicks

B. M.

McGue

(2001). Altruism and antisocial behavior: Independent tendencies, unique personality correlates, distinct etiologies. Psychological Science, 12, 397-402.

36.

Lahey

Moffitt

T. E.

Caspi

(Eds.). (2003). Causes of conduct disorder and juvenile delinquency. New York, NY: Guilford Press.

37.

Lyons

M. J.

(1996). A twin study of self-reported criminal behaviour. In Ciba Foundation (Ed.), Genetics of criminal and antisocial behavior (pp. 61-69). Chichester, UK: Wiley.

38.

Maes

H. H. M.

Neale

M. C.

Kendler

Hewitt

J. K.

Silberg

J. L.

Foley

D. L.

Meyer

J. M.

Rutter

Simonoff

Pickles

Eaves

L. J.

(1998). Assortative mating for major psychiatric diagnosis in two population-based samples. Psychological Medicine, 28, 1389 -1401.

39.

Mazerolle

(1998). Gender, general strain, and delinquency: An empirical evaluation. Justice Quarterly, 15, 65-91.

40.

Mears

D. P.

Ploeger

Warr

(1998). Explaining the gender gap in delinquency: Peer influence and moral evaluations of behavior. Journal of Research in Crime and Delinquency, 35, 251-266.

41.

Miles

D. R.

Carey

(1997). Genetic and environmental architecture of human aggression. Journal of Personality and Social Psychology, 72, 207-217.

42.

Moffitt

T. E.

(1993). The neuropsychology of conduct disorder. Development and Psychopathology, 5, 135-151.

43.

Moffitt

T. E.

Caspi

Rutter

Silva

P. A.

(2001). Sex differences in antisocial behavior: Conduct disorder, delinquency, and violence in the Dunedin Longitudinal Study. Cambridge, UK: Cambridge University Press.

44.

Moffitt

T. E.

(2005). Genetic and environmental influences on antisocial behaviors: Evidence from behavioral-genetic research. Advances in Genetics, 55, 41-104.

45.

Neale

M. C.

Cardon

L. R.

(1992). Methodology for genetic studies of twins and families. Dordrecht, The Netherlands: Kluwer Academic.

46.

Ostrer

(2001). Invited review: Sex-based differences in gene expression. Journal of Applied Physiology, 91, 2384-2388.

47.

Pearce

L. D.

Haynie

D. L.

(2003). Intergenerational religious dynamics and adolescent delinquency. Social Forces, 82, 1553-1572.

48.

Popma

Raine

(2006). Will future forensic assessment be neurobiologic? Child and Adolescent Psychiatric Clinics of North America, 15, 429-444.

49.

Quinton

Pickles

Maughan

Rutter

(1993). Partners, peers, and pathways: Assortative pairing and continuities in conduct disorder. Development and Psychopathology, 5, 763-783.

50.

Rhee

S. H.

Waldman

I. D.

(2002). Genetic and environmental influences on antisocial behavior: A meta-analysis of twin and adoption studies. Psychological Bulletin, 128, 490-529.

51.

Ronald

Happe

Plomin

(2005). The genetic relationship between individual differences in social and nonsocial behaviours characteristic of autism. Developmental Science, 8, 444-458.

52.

Rowe

D. C.

(1983). Biometrical genetic models of self-reported delinquent behavior: A twin study. Behavior Genetics, 13, 473-489.

53.

Rowe

D. C.

(1986). Genetic and environmental components of antisocial behavior: A study of 265 twin pairs. Criminology, 24, 513-532.

54.

Rushton

J. P.

(1996). Self-report delinquency and violence in adult twins. Psychiatric Genetics, 6, 87-90.

55.

Rutter

(2003). Crucial paths from risk indicator to causal mechanism. In Lahey

Moffitt

T. E.

Caspi

(Eds.), Causes of conduct disorder and juvenile delinquency (pp. 3-24). New York, NY: Guilford Press.

56.

Rutter

Giller

Hagell

(1998). Antisocial behavior by young people. Cambridge, UK: Cambridge University Press.

57.

Silberg

J. L.

Erickson

M. T.

Meyer

J. M.

Eaves

L. J.

Rutter

Hewitt

J. K.

(1994). The application of structural equation modeling to maternal ratings of twins. Journal of Consulting and Clinical Psychology, 62, 510-521.

58.

Slutske

W. S.

Heath

A. C.

Dinwiddie

S. H.

Madden

P. A. F.

Bucholz

K. K.

Dunne

M. P.

. . . Martin

N. G.

(1997). Modeling genetic and environmental influences in the etiology of conduct disorder: A study of 2,682 adult twin pairs. Journal of Abnormal Psychology, 106, 266-279.

59.

Smith

Cullen

F. T.

Latessa

E. J.

(2009). Can 14,737 women be wrong? A meta-analysis of the LSI-R and recidivism for female offenders. Criminology & Public Policy, 8, 183-208.

60.

Smith

D. A.

Paternoster

(1987). The gender gap in theories of deviance: Issues and evidence. Journal of Research in Crime and Delinquency, 24, 140-172.

61.

Smith

D. A.

Visher

C. A.

(1980). Sex and involvement in deviance/crime: A quantitative review of the empirical literature. American Sociological Review, 45, 691-701.

62.

Stattin

Magnusson

Reichel

(1989). Criminal activity at different ages: A study based on a Swedish longitudinal research population. British Journal of Criminology, 29, 368-385.

63.

Steffensmeier

Allan

(1996). Gender and crime: Toward a gendered theory of female offending. Annual Review of Sociology, 22, 459-487.

64.

Taylor

McGue

Iacono

W. G.

(2000a). Sex differences, assortative mating, and cultural transmission effects on adolescent delinquency: A twin family study. Journal of Child Psychology and Psychiatry, 41, 433-440.

65.

Taylor

McGue

Iacono

W. G.

Lykken

D. T.

(2000b). A behavioral genetic analysis of the relationship between the socialization scale and self-reported delinquency. Journal of Personality, 68(1), 29-50.

66.

Udry

J. R.

(2003). The national longitudinal study of adolescent health (add health), waves I and II, 1994 -1996; wave III, 2001-2002. Chapel Hill, NC: Carolina Population Center, University of North Carolina at Chapel Hill.

67.

van den Oord

Verhulst

F. C.

Boomsma

D. I.

(1996). A genetic study of maternal and paternal ratings of problem behaviors in 3-year-old twins. Journal of Abnormal Psychology, 105, 349-357.

68.

van den Oord

E. J. C. G.

Boomsma

D. I.

Verhulst

F. C.

(1994). A study of problem behaviors in 10-to 15-year-old biologically related and unrelated international adoptees. Behavior Genetics, 24, 193-205.

69.

van der Ende

Verhulst

F. C.

(2005). Informant, gender and age differences in ratings of adolescent problem behaviour. European Child and Adolescent Psychiatry, 14, 117-126.

70.

Van de Rakt

Nieuwbeerta

de Graaf

N. D.

(2008). Like father, like son: The relationships between conviction trajectories of fathers and their sons and daughters. British Journal of Criminology, 48, 538-556.

71.

Vaske

Wright

J. P.

Boisvert

Beaver

K. M.

(2011). Gender, genetic risk, and criminal behavior. Psychiatry Research, 185, 376-381.

72.

Vierikko

Pulkkinen

Kaprio

Viken

Rose

R. J.

(2003). Sex differences in genetic and environmental effects on aggression. Aggressive Behavior, 29(1), 55-68.

73.

West

D. J.

Farrington

D. P.

(1977). The delinquent way of life: Third report of the Cambridge study in delinquent development. London, UK: Heinemann Educational Books.