The double-edged sword



Download 107.55 Kb.
Page1/2
Date conversion13.05.2016
Size107.55 Kb.
  1   2



The Double-Edged Sword of Genetic Accounts of Criminality:

Causal attributions from genetic ascriptions affect legal decision-making

Benjamin Y. Cheung and Steven J. Heine

University of British Columbia

(in press) Personality and Social Psychology Bulletin



Abstract

Much debate exists surrounding the applicability of genetic information in the courtroom, making the psychological processes underlying how people consider this information important to explore. This paper addresses how people think about different kinds of causal explanations in legal decision-making contexts. Three studies involving a total of 600 Mechanical Turk and university participants found that genetic, versus environmental, explanations of criminal behavior lead people to view the applicability of various defense claims differently, perceive the perpetrator’s mental state differently, and draw different causal attributions. Moreover, meditation and path analyses highlight the double-edged nature of genetic attributions – they simultaneously reduce people’s perception of the perpetrator’s sense of control while increasing people’s tendencies to attribute cause to internal factors, and to expect the perpetrator to reoffend. These countervailing relations, in turn, predict sentencing in opposite directions, although no overall differences in sentencing or ultimate verdicts were found.



Keywords: genetic essentialism, attribution theory, legal decision-making, psychological essentialism

The Double-Edged Sword of Genetic Accounts of Criminality:

Causal attributions from genetic ascriptions affect legal decision-making

On an episode of Law and Order: Special Victims Unit (Green & Platt, 2007), a sex-addicted teenager was charged with rape; but his attorney claimed that the teenager had a history of sexual impulsivity due to a congenital condition that minimized his ability to exert self-control. The teenager then plea-bargained his sentence down to admission into a treatment facility without jail time. He subsequently was raped by the facility’s janitor with a history of sexual violence, whose own defense was, “I couldn’t help it. My brain is wired wrong – just like that kid, right?”



Making Sense of Genetic Evidence in the Courtroom

This Law and Order episode illustrates an important question that recent psychological research has been trying to answer (e.g., Dar-Nimrod & Heine, 2011; Monterosso, Royzman, & Schwartz, 2005). Why are people so strongly influenced by attributions to biological causes? This is particularly important in legal contexts, with courtrooms witnessing the use of biological explanations to mitigate responsibility for various crimes. Indeed, formal biological theories of crime have existed for over a century (e.g. Lombroso, 1876/2007). More recently, innovations in genomics research have provided the public with the perceived prospect of decoding the genetic aspect of criminality (Caspi, et al., 2002; Friedland, 1998). While a single gene for criminality likely does not exist (Alper, 1998), candidate genes have been identified that significantly predict various crime-related behaviors including alcoholism (Pandey, Roy, Zhang, & Xu, 2004), cocaine addiction (Bilbaoa, et al., 2008), and violence (Brunner, Nelen, Breakefield, Ropers, & van Oost, 1993). Evidence of such genetic predispositions has been used as mitigating factors in criminal defenses since the 1990s (Denno, 2006). As more candidate genes will likely be discovered that are associated with a wider array of behaviors, it is important to understand how people make sense of such kinds of genetic causes. Do people view genetic evidence as more mitigating of criminal responsibility than they view other kinds of evidence? Legal scholars have long debated the implications of genetics for legal decision-making (Hoffmann & Rothenberg, 2007; Johnson, 1998; Jones, 2003). The current debate mostly relates to whether genetic predispositions objectively impact volition and free will, and whether this impact should affect how the judicial system punishes offenders. One issue that this debate misses is how people actually perceive genetic evidence.



Impact of genetic attributions on cognitions

People frequently perceive genetic information in biased ways. Nelkin and Lindee (1995) contended that our cultural environment is saturated with simplistic and fatalistic messages about genes, imbuing genes with undue inferential power. Dar-Nimrod and Heine (2011) further proposed that people engage in a specific set of biased and fatalistic cognitions, called genetic essentialist biases, when they encounter genetic explanations of behavior. For instance, encountering genetic attributions leads people to view the associated phenotype as necessarily following from the genotype. Although there are monogenic phenomena that follow this deterministic relationship (e.g., autosomal dominant genetic disorders such as Huntington’s Disease), it is inappropriate to generalize from these cases as the vast majority of genetic effects are far more complex, involving the interaction of numerous genes, whose expression is governed by environmental experiences and other epigenetic influences (e.g. Jablonka & Lamb, 2006). In these far more numerous cases, assuming immutability in the genotype-phenotype relation would be inaccurate. As an example of the biased effects of genetic attributions, people who learn about genetic influences of mental illnesses come to associate the illnesses with less controllability and more persistence, while simultaneously leading to more sympathetic views of the afflicted (Lebowitz, Ahn, & Nolen-Hoeksema, 2013; Phelan, 2005; Walker & Read, 2002).



Implications for Legal Decision-Making. A critical feature of genetic essentialist biases is their link to a perceived lack of control (e.g., Dar-Nimrod, Heine, Cheung, & Schaller, 2011). This has important implications for legal decision-making due to the preponderance of free will and volition assumed by the criminal justice system (Coffey, 1993). Given that the presumption of self-control is paramount for judging responsibility, genetic attributions of criminal behavior should accordingly lead to weaker perceptions of criminal responsibility, and also to more lenient punishments. Despite much discussion within legal circles regarding whether such perceptions should occur as a result of genetic explanations (Berryessa & Cho, 2013; Johnson, 1998; Rose, 2000), few studies have examined whether they do occur. Anecdotal evidence from court cases around the world demonstrates the impact of genetic evidence on judges and jurors. For example, partly on account of evidence based on Caspi et al.’s (2002) behavioral genetics work, two independent Italian courts reduced the sentences of two different defendants charged with murder – a reduction of one year in one case, and a reduction from life imprisonment to twenty years in another case (Forzano, et al., 2010; Owens, 2011). Beyond anecdotal evidence, studies find that reducing beliefs in free will and increasing perception of a neural basis of human behavior both lead to reduced retributive punishment (Shariff, et al., 2014). More germane to our research, state trial judges issued shorter sentences to a hypothetical perpetrator with a biomechanical explanation for their psychopathy, perceiving the biomechanical cause as a mitigating factor (although the researchers did not examine whether other kinds of explanations would be viewed as equally mitigating; Aspinwall, Brown, & Tabery, 2012). Moreover, these judges also found biomechanical explanations to be aggravating, suggesting a higher likelihood of reoffending (i.e. higher recidivism). How the general public negotiates these countervailing expectations have not been well-understood.

Other studies found that biological explanations decreased perceptions of culpability for hypothetical criminals relative to experiential explanations (Dar-Nimrod et al., 2011; Monterosso et al., 2005). These studies contribute to the growing debate about whether biological explanations of crime may influence people’s judgments in ways that other explanations do not (see Appelbaum & Scurich, 2013; Raad & Appelbaum, 2015), and such differences may have important implications for legal decision-making.



Overview of studies

Our three studies strive to identify the degree to which genetic explanations affect various aspects of legal decision-making relative to other kinds of explanations. This paper examines a broader set of legal decisions compared with past research that are either explicitly expressed (e.g. preferred verdict) or internally held (e.g. attributions of the perpetrator), and seeks to identify mediators that underlie these effects. Importantly, as Aspinwall et al.’s (2012) results suggest, genetic attributions may possess a double-edged nature. Thus, we hypothesize that genetic explanations for criminals, relative to environmental explanations, will lead to both mitigating perceptions such as more lenient perceptions of culpability, and aggravating perceptions such as greater expected recidivism. These studies include various novel measures rarely used in past research on essentialist biases in decision-making, such as preferred verdicts, mitigating defenses, and expected recidivism. We are particularly interested in comparisons between genetic and environmental explanations as this allows us to determine whether genetic explanations differ from other types of explanations. Our three studies culminated in an aggregate dataset with an appropriate sample size to conduct a path analysis (Preacher & Coffman, 2006), directly examining the impact of the potential simultaneously double-edged nature of genetic attributions on punitiveness.



Study 1

Method

Participants. In our initial study we anticipated that we would have medium effect sizes. With a target power level of 0.80, an α-level of 0.05, and a total of three groups and one covariate (described below), we required approximately 130 participants (Faul, Erdfelder, Lang, & Buchner, 2007). We subsequently recruited 132 American participants (40% males; Mage = 35.48, SD = 12.71) from Amazon’s Mechanical Turk (MTurk). The sample consisted of 76% Euro-Americans, 9% East Asians, 11% “Others”, 4% unspecified (due to the small number of participants who were not of European descent, we could not test for interactions involving culture). Furthermore, because political orientation is associated with legal decision-making (see Cochran, Boots, & Chamlin, 2006), participants also indicated this using a 5-point scale (1 = “Very liberal”, 5 = “Very conservative”; M = 2.59; SD = 0.97). Participants received $0.25 for participating in the study.

Materials. The study consisted of the following materials, in order:

Experimental manipulation. The manipulation consisted of one of three randomly assigned vignettes about a college student, Patrick, who fatally stabbed another person after an altercation. The three vignettes differed only in terms of the explanation given for his behavior: the Genetic vignette described Patrick’s (fictitious) genetic predisposition for responding to provocations with violence; the Environmental vignette described Patrick’s rearing environment that predisposed him to violence; and the Control vignette offered no explanation. Importantly, the magnitude of the purported effect of each explanation was identical; both causes were described as leading to a fourfold increase in the likelihood of violence.

Perceptions of culpability. To measure the extent to which people perceived Patrick’s culpability for his crime, we asked the following questions:

Defense claims. Participants rated the applicability of three defense claims (Insanity; Diminished Capacity; Intoxication) for Patrick on a 7-point scale (0=Not at all applicable; 6=perfectly applicable). Each defense claim was accompanied by corresponding definitions based on legal texts (American Law Institute, 1962; Brody, Acker, & Logan, 2001; Padfield, 2008). Insanity and Diminished Capacity were included as they constitute a full and partial defense, respectively (Baum, 2011), allowing us to examine people’s perception of the applicability of different degrees of defense claims. Intoxication was included to ensure that participants would not simply deem any defense claim to be more or less applicable (intoxication was unrelated to information from the vignette). We expected that the Insanity and Diminished Capacity defenses would be deemed more applicable in the Genetic condition than the other conditions, but not for the Intoxication defense.

Verdicts. Participants made the same judgement regarding four verdicts (First degree murder; Second degree murder; Manslaughter; Not guilty) as they did with the defense claims discussed above, using the same 7-point scale. Each verdict was presented with an accompanying definition based on those that appear in the United States Code (18 U.S.C. § 1111-1112).

Sentencing. The participants read that Patrick had been convicted of manslaughter, and they were to assign an appropriate prison sentence using an 11-point scale (1=5 years, 11=>50 years). Based on previous studies (e.g. Aspinwall et al., 2012; Dar-Nimrod et al., 2011), we expected that participants in the Genetic condition would assign a shorter prison sentence compared to participants in the other conditions.

Perpetrator-relevant perceptions. Finally, participants indicated how criminally responsible they felt that Patrick was on an 11-point scale (0=not at all criminally responsible, 10=completely criminally responsible). Participants also used a 7-point scale to rate the degree to which Patrick: (a) had conscious control over his actions (0=not at all, 6=complete control); (b) intended to kill the victim (0=not at all, 6=full intention); (c) had knowledge that his actions would lead to the victim’s death (0=not at all, 6=full knowledge); and (d) would reoffend if he were released back into the public (0=not at all likely, 6=completely likely). Participants should perceive less criminal responsibility, less conscious control, less intention to kill, and greater recidivism in the Genetic condition compared to the other conditions. In contrast, the perpetrator’s knowledge that his actions would kill the victim was not expected to differ across conditions as the kind of explanation should be irrelevant to knowledge. Moreover, based on previous findings, particularly regarding the perceived double-edged nature of biological explanations of human behavior (e.g. Dar-Nimrod et al., 2011), we also predicted that perceptions of conscious control and recidivism on the part of the perpetrator should mediate differences between groups in terms of prescribed prison sentences.

Results

We performed one-way analyses of covariance (ANCOVAs) using Conditions (Genetic; Environmental; Control) as the independent variable, with political orientation as the covariate (conservativism may be associated with overall heightened punitiveness). Adjusted means were compared using LSD multiple comparisons. All adjusted means and effect sizes from all pairwise comparisons can be found in the online supplemental materials (OSMs).



Defense claims. There was a main effect of Condition on the applicability of the insanity defense, F(2,123)=6.57, p=.002, partial η2=0.10, CI.95=[0.01,0.19] (see Figure 1A). This defense was preferred more in the Genetic condition (M=2.35, SE=0.28) than in the Environmental (M=1.11, SE=0.29), p=.003, d=0.67, CI.95=[0.23,1.11] and Control conditions (M=1.10, SE=0.28), p=.002, d=0.68, CI.95=[0.25,1.11]. The latter two were not significantly different, p=.972.

There was a main effect of Condition on the diminished capacity defense, F(2,123)=12.40, p<.001, partial η2=0.17, CI.95=[0.06,0.27]. This defense was preferred more in the Genetic condition (M=3.64, SE=0.30) than in the Environmental (M=1.96, SE=0.32), p<.001, d=0.84, CI.95=[0.39,1.28] and Control conditions (M=1.64, SE=0.31), p<.001, d=1.00, CI.95=[0.55, 1.44]. The latter two were not significantly different, p=.469.

There was a main effect of Condition on the intoxication defense, F(2,123)=12.68, p<.001, partial η2=0.17, CI.95=[0.06,0.28]. It was significantly preferred more in the Control condition (M=1.19, SE=0.17) than in the Genetic (M=0.10, SE=0.16), p<.001, d=-1.01, CI.95=[-1.45,-0.56] and Environmental conditions (M=0.26, SE=0.17), p<.001, d=-0.86, CI.95=[-1.30,-0.40]. The latter two were not significantly different, p=.494. We suspect that because the Control condition offered no explanations, participants assumed intoxication was involved.

Verdicts. There was no effect of Condition on perceived appropriateness of any verdicts, Fs<2.50, ps>.100, partial η2s<0.04.

Sentencing. There were no effect of Condition, F(2,123)=0.25, p=.780, partial η2s=0.00, CI.95=[0.00,0.04].

Perpetrator-relevant perceptions. There was a significant effect of Condition on perceived control, F(2,123)=3.14, p=.047, partial η2=0.05, CI.95=[0.00,0.13]. It was marginally lower in the Genetic condition (M=4.19, SE=0.21) than in the Environmental condition (M=4.76, SE=0.22), p=.065, d=-0.41, CI.95=[-0.84,0.03], but significantly lower than the Control condition (M=4.90, SE=0.21), p=.019, d=-0.51, CI.95=[-0.93,-0.08]. The latter two were not significantly different, p=.647.

There was a significant effect of Condition on recidivism, F(2,123)=5.70, p=.004, partial η2=0.08, CI.95=[0.01,0.18]. It was significantly higher in the Genetic condition (M=4.47, SE=0.24) than in the Control condition (M=3.32, SE=0.24), p=.001, d=0.72, CI.95=[0.28,1.15], but not compared to the Environmental condition (M=4.04, SE=0.25), p=.221. The latter two differed significantly, p=.043, d=0.45, CI.95=[0.01,0.88].

There were no effects of Condition on perceived criminal responsibility, intent to kill, and knowledge that actions would have killed, Fs<2.00, ps>.300, partial η2s<0.02.

Mediation analysis1. Despite sentencing not being affected by Condition, such results may mask informative mechanisms for theory building and confirmation (see Rucker, Preacher, Tormala, & Petty, 2011; Zhao, Lynch, & Chen, 2010), especially when there are a priori expectations of underlying mechanisms. Furthermore, the traditional Baron and Kenny (1986) requirement for a significant direct effect for mediation leads to an underpowered approach for testing mediation (Edwards & Lambert, 2007; MacKinnon, Fairchild, & Fritz, 2007); therefore, we focused on the indirect effects in determining mediation. We performed a mediation analysis in accordance with Dar-Nimrod et al. (2011), testing for an indirect effect of different conditions on sentencing through perceived control. Figure 2A shows a marginal tendency for participants in the Environmental condition to perceive Patrick as having more conscious control over his behaviour than participants in the Genetic condition, =0.40, t(123)=1.86, p=.065, CI.95=[-0.03,0.83]. This, in turn, predicts a lengthier sentence, =0.27, t(122)=3.09, p=.002, CI.95=[0.09,0.45]. The indirect effect is in the predicted direction, =0.11, CI.95=[-0.01,0.29]. Compared to the non-mediated direct effect, =0.05, t(123)=0.244, p=.808, CI.95=[-0.41,0.52], this mediated direct effect remains non-significant, =-0.05, t(122)=-0.26, p=.794, CI.95=[-0.50,0.39].

Another mediation analysis tested the double-edged nature of genes, revealing that the Genetic condition was not significantly associated with perceived recidivism compared to the Environmental condition, =-0.26, t(123)=-1.23, p=.221, CI.95=[-0.67,0.14], but perceived recidivism predicted lengthier prison sentences =0.31, t(122)=3.58, p<.001, CI.95=[0.15,0.47]. The indirect effect is in the predicted direction, =-0.08, CI.95=[-0.23,0.04]. The mediated direct effect remains non-significant, =0.13, t(122)=0.65, p=.518, CI.95=[-0.30,0.56], but was nominally larger than the non-mediated direct effect, potentially suggesting a suppression effect (see Figure 2C). We investigate this further in subsequent studies.



Discussion

Study 1 found evidence that exposure to genetic explanations of criminal behavior impacted various aspects of legal decision-making – such explanations led to higher perceived applicability of both the Insanity and Diminished Capacity defenses (but not the Intoxication defense), compared with environmental explanations. This is the first demonstration of genetic ascriptions’ effect on people’s endorsement of various mitigating accounts for a crime. This suggests that people view genes are responsible for one’s actions by liberating them from their own responsibility. Moreover, participants in the Genetic condition tended to differ from those in the Control condition, which tended to be similar to those in the Environmental condition.

Genetic explanations also led people to ascribe less control to the perpetrator than people’s default and environmental explanations, mirroring previous research (Dar-Nimrod et al., 2011). As predicted, perceptions of the perpetrator’s knowledge of the consequences of his actions did not differ, as the genetic explanation affected neither the perpetrator’s intelligence nor knowledge. Unexpectedly, analyses did not reveal a significant effect of condition on perceived intentions. It may be that all participants perceived the perpetrator as having the same degree of intention to kill, but the genetic predisposition rendered the perpetrator less able to control his impulses. We explore this latter possibility in Study 3.

Counter to our expectations, explanations of criminal behavior did not affect perceived criminal responsibility, verdicts, and sentencing, despite affecting perceived control. We had expected that such perceptions would lead to more lenient sentences and preferences for less serious verdicts. Subsequent mediation analyses revealed that, as expected, there are marginal indirect effects of behavioral explanations on sentencing through perceived conscious control and perceived recidivism, but in opposite directions. Specifically, genetic ascriptions lead to nominally lower perceptions of perceived conscious control of one’s behavior, predicting shorter prison sentences. While genetic ascriptions did not lead to significantly higher perceived recidivism, it predicted lengthier prison sentences, resulting in a marginal indirect effect. These two mediators work in opposition to each other, and provide empirical evidence for the double-edged perceptions given by justices in Aspinwall et al.’s study (2012). We further investigate this in subsequent studies.



Study 2

We conducted Study 2 to try to replicate, and test the robustness of, Study 1’s results, and to explore additional effects of genetic explanations. In particular, we explored the relevance of attribution theory (Abramson, Seligman, & Teasdale, 1978; Weiner, et al., 1971) to how genetic explanations affect legal decision-making.

Attribution theory assumes four dimensions relevant to how people understand the causes of behavior: causal locus, causal stability, causal control, and causal specificity (Abramson, Seligman, & Teasdale, 1978; Anderson, Krull, & Weiner, 1996; Weiner, 1985). Causal locus refers to whether the behavioral cause is internal or external to the person. Causal stability is whether the behavioral cause would always exert its influence on the outcome behavior. Causal control refers to whether one has control over the cause of the behavior. Finally, causal specificity is whether the behavioral cause only affects a specific domain rather than being domain-general. Given theoretical expectations about people’s tendencies to essentialize genes (Nelkin & Lindee, 1995), genetic causes should be perceived as more internal, stable, specific, and less controllable, than environmental causes.

Researchers have previously applied attribution theory to legal decision-making. For example, more internal attributions of a crime led participants to view the criminal behavior as being more stable across time, which was associated with greater expected recidivism (Carroll & Payne, 1977), leading to lengthier prison sentences (Tam, Shu, Ng, & Tong, 2013). Note that this is the opposite direction of what past research has found with perceived biological causes of criminality (e.g., Aspinwall et al., 2012), despite genetic causes being seen as more internal than environmental causes (Dar-Nimrod & Heine, 2011). This suggests that attribution theory may be key for understanding the double-edged nature of genetic ascriptions.



Methods

Participants. We increased our sample size, subsequently recruiting 165 undergraduate students from the University of British Columbia (23% males; Mage=21.48, SD=4.89). The sample consisted of 38% Euro-Canadians, 33% East Asians, and 28% “Others”, although no culture-condition interactions were significant, leading us to collapse across ethnic groups. Participants’ political orientation was assessed using the same measure as Study 1 (M=2.47; SD=0.98). Participants received course credit for their participation. As in Study 1, we conducted one-way ANCOVAs with political orientation as the covariate.

Materials. The study used the same measures as in Study 12, with the addition of the Attributional Style Questionnaire (Peterson, et al., 1982) that was adapted for our specific vignettes. Furthermore, the original questionnaire only included questions for causal locus, stability, and specificity (for ease of interpretation, the causal specificity question was reverse-coded so that higher scores refer to greater causal specificity). We thus added a question regarding the causal control to reflect the four dimensions of Weiner’s attribution theory [“In the future, whenever Patrick is provoked in a fashion similar to what occurred in the scenario, will Patrick be able to control how much the cause of his behavior will influence him?”]. We also asked participants about their belief in the general malleability of the causal factor to which they were assigned (causal malleability) [“Is the effect of the cause of Patrick’s behavior something that can be changed or corrected (i.e. such that the cause of Patrick’s behavior will no longer affect Patrick’s behavior)?”]. All questions were on a 1-7 scale, with higher scores indicating greater internal, stable, specific, controllable, and malleable attributions, separately. In addition, based on work suggesting that internal causes are associated with lengthier prison sentences (e.g. Tam, Shu, Ng, & Tong, 2013), causal locus should mediate group differences in the length of prescribed prison sentences.

Results

Defense claims. There was a main effect of Condition on the applicability of the insanity defense, F(2,160)=8.11, p<.001, partial η2=0.09, CI.95=[0.02,0.18] (see Figure 1B). It was endorsed more in the Genetic condition (M=3.76, SE=0.26) than in the Environmental (M=2.31, SE=0.26), p<.001, d=0.76, CI.95=[0.37,1.14], and Control conditions (M=2.85, SE=0.26), p=.013, d=0.48, CI.95=[0.10,0.85]. The latter two were not significantly different, p=.144.

There was also a main effect of Condition for the diminished capacity defense, F(2,160)=3.69, p=.027, partial η2=0.04, CI.95=[0.00,0.11]. It was preferred more in the Genetic condition (M=2.59, SE=0.26) than in the Environmental condition (M=1.61, SE=0.26), p=.008, d=0.51, CI.95=[0.13,0.89], but not in the Control condition (M=2.22, SE=0.26), p=.307. The latter two were marginally different, p=.099, d=-0.32, CI.95=[-0.70,0.06].

There was a main effect of Condition for the intoxication defense, F(2,159)=7.43, p<.001, partial η2=0.08, CI.95=[0.02,0.17]. It was significantly endorsed more in the Control condition (M=1.88, SE=0.20) than in the Genetic (M=1.05, SE=0.20), p=.004, d=-0.56, CI.95=[-0.94,-0.17] and Environmental conditions (M=0.83, SE=0.20), p<.001, d=-0.70, CI.95=[-1.09,-0.31]. The latter two were not significantly different, p=.450.

Verdicts. There was no effect of Condition, Fs<1.20, ps>.300, partial η2s<0.01.

Sentencing. There were no effect of Condition, F(2,160)=0.44, p=.657, partial η2s=0.00, CI.95=[0.00,0.04].

Perpetrator-relevant perceptions. There was a significant effect of Condition on perceived control, F(2,160)=6.95, p=.001, partial η2=0.08, CI.95=[0.01,0.16]. It was significantly lower in the Genetic condition (M=3.46, SE=0.18) than in the Environmental (M=4.39, SE=0.18), p<.001, d=-0.69, CI.95=[-1.07,-0.30], and Control conditions (M=4.11, SE=0.18), p=.012, d=-0.48, CI.95=[-0.86,-0.10]. The latter two were not significantly different, p=.282.

Condition significantly affected perceived intent to kill, F(2,160)=3.26, p=.041, partial η2=0.04, CI.95=[0.00,0.10]. It was significantly lower in the Genetic condition (M=2.46, SE=0.20) than in the Environmental (M=3.04, SE=0.21), p=.045, d=-0.38, CI.95=[-0.76,-0.01], and Control conditions (M=3.14, SE=0.21), p=.020, d=-0.45, CI.95=[-0.83,-0.07]. The latter two were not significantly different, p=.739.

There were no effects of Condition on perceived criminal responsibility and knowledge that actions would have killed, Fs<2.00, ps>.200, partial η2s<0.02.

Causal attributions. Condition significantly affected causal locus, F(2,160)=5.05, p=.007, partial η2=0.06, CI.95=[0.00,0.13]. It is significantly more internal (higher) in the Genetic condition (M=5.39, SE=0.16) than in the Environmental (M=4.77, SE=0.17), p=.008, d=0.51, CI.95=[0.13,0.89], and Control conditions (M=4.74, SE=0.17), p=.005, d=0.54, CI.95=[0.16,0.92]. The latter two were not significantly different, p=.879.

Condition significantly affected causal stability, F(2,160)=6.28, p=.002, partial η2=0.07, CI.95=[0.01,0.15]. It is significantly more stable (higher) in the Genetic condition (M=5.66, SE=0.13) than in the Control condition (M=5.00, SE=0.13), p=.001, d=0.67, CI.95=[0.28,1.05], but not in the Environmental condition (M=5.41, SE=0.13), p=.187. The latter two differed significantly from each other, p=.032, d=0.42, CI.95=[0.04,0.80].

Condition significantly affected causal specificity, F(2,160)=6.53, p=.002, partial η2=0.07, CI.95=[0.01,0.16]. Participants saw greater causal specificity in the Genetics condition (M=2.47, SE=0.15) than in the Environmental condition, (M=1.86, SE=0.16), p=.007, d=0.52, CI.95=[0.14,0.90], but not in the Control condition (M=2.63, SE=0.16), p=.484. The latter two differed significantly from each other, p=.001, d=-0.66, CI.95=[-1.04,-0.27].

Interestingly, condition did not affect causal control, F(2,160)=0.82, p=.441, partial η2=0.01, CI.95=[0.00,0.05], suggesting that people did not think that the effect of the difference causes were more/less controllable in any given situation.

Condition significantly affected causal malleability, F(2,160)=6.49, p=.002, partial η2=0.07, CI.95=[0.01,0.15]. Participants saw less causal malleability in the Genetics condition (M=4.37, SE=0.18) than in the Environmental condition, (M=5.24, SE=0.18), p=.001, d=-0.64, CI.95=[-1.02,-0.25], but not in the Control condition (M=4.50, SE=0.18), p=.621. The latter two differed significantly from each other, p=.005, d=0.50, CI.95=[0.16,0.93].

Mediation analyses. Similar to Study 1, we examined the indirect effect of behavioral explanations on sentencing through perceived conscious control, focusing on the Genetic-Environment contrast. The results replicated Study 1, such that participants in the environmental condition ascribed greater conscious control to Patrick, =0.67, t(160)=3.62, p<.001, CI.95=[0.30,1.03], which, in turn, predicted a lengthier prison sentence, =0.22, t(159)=2.79, p=.006, CI.95=[0.01,0.42]. This indirect effect was significant, =0.14, CI.95=[0.01,0.32], although the mediated direct effect remains non-significant, =0.02, t(159)=0.09, p=.929, CI.95=[-0.37,0.42] from the non-mediated direct effect, =0.17, t(160)=0.87, p=.383, CI.95=[-0.23,0.56] (see Figure 2A).

Another mediation analysis also revealed the predicted indirect effect of behavioral explanations on sentencing through internal attributions. The results suggested that participants in the Environmental condition perceived the cause of Patrick’s behavior to be less internal than participants in the Genetic condition, =-0.50, t(160)=-2.66, p=.008, CI.95=[-0.84,-0.16]. On the other hand, more internal causal attributions were associated with lengthier sentences, =0.17, t(159)=2.13, p=.035, CI.95=[0.01,0.34]. The indirect effect was significant, =-0.08, CI.95=[-0.21,-0.00]. The mediated direct effect remained non-significant, =0.25, t(159)=1.30, p=.194, CI.95=[-0.13,0.64], although it increased from the non-mediated direct effect, =0.17, suggesting a suppression effect (see Figure 2B).



Discussion

Replicating Study 1, participants perceived both the Insanity and the Diminished Capacity defenses to be more applicable after reading a genetic explanation for Patrick’s behavior compared with an environmental explanation – this was not the case with the Intoxication defense. Again, this did not translate into different verdicts, suggesting that there are additional factors beyond simply explaining violent behavior as etiologically genetic or environmental.

Also like Study 1, judgments and perceptions in the Genetic, but not the Environmental, condition tended to deviate from the Control condition. In contrast, both experimental conditions tended to deviate from the Control condition with causal attributions, suggesting that explicitly referencing any cause impacts one’s causal attributions. Notably, causal attributions for all groups tended to cluster at the scales’ extremes, suggesting general agreement in the attributional profile of different causes. Despite this, genetic explanations still seem to have a statistically different profile than environmental ones – one that is more internal and specific, and less malleable, resembling the profile found by Dar-Nimrod, Cheung, Ruby, and Heine (2014) regarding obesity perceptions.

Differences in the perceived applicability of criminal defenses again did not lead to differences in sentencing and perceived criminal responsibility; however, the mediation analyses in the present study uncovered two mechanisms that explain the lack of significant differences in sentencing across conditions. Study 2 revealed a significant indirect effect of participants’ condition on the length of prescribed prison sentences through perceived conscious control and internal attributions. That is, genetic explanations led to less perceived conscious control, resulting in shorter sentences. Conversely, genetic explanations also led to stronger internal attributions, resulting in lengthier sentences, similar to previous research (e.g. Carroll & Payne, 1977; Sanderson, Zanna, & Darley, 2000; Tam et al., 2013). This tension between shorter sentences due to less perceived conscious control and lengthier sentences due to stronger internal attributions may explain why there is no direct effect of behavioral explanation on sentence length. Moreover, the mediated direct effect through internal attribution is larger than the non-mediated direct effect, potentially showing a positive cooperative suppression effect (Conger, 1974; Krus & Wilkinson, 1986), which we attempt to replicate in Study 3. Overall, seeing a genetic cause behind violent behaviours led participants to view the behaviors as less controllable and the cause as more internal. In conjunction with the mediation in Study 1 regarding perceived recidivism, genetic attributions appear to be seen as double-edged swords that can be simultaneously mitigating and aggravating.



Study 3

The findings from Studies 1 and 2 are consistent – genetic accounts, compared to environmental explanations, lead people to consider different defenses and other aspects of the legal decision-making process differently; however, it is not clear whether genetic ascriptions affect people’s judgments because they expect perpetrators to have stronger impulses towards violence or have less control of their impulses. Study 3 sought to clarify this issue, and to replicate findings from Studies 1 and 2.



Methods

Participants. Because this study has an extra condition (described below), and that some of our previous effects had rather wide confidence intervals, we boosted this study’s sample size to 303 American participants from Amazon’s MTurk, with an effective sample size of 298 participants who passed our comprehension check questions (42% males; Mage=34.62, SD=11.26). The sample consisted of 68% Euro-Americans, 7% African Americans, 3% Asian Americans, 22% “Others”. Participants’ political orientation was assessed using the same measure as before (M=2.63; SD=1.05). We again conducted one-way ANCOVAs with political orientation as the covariate.

Materials. The study used the same Genetic and Environmental scenarios as in Studies 1 and 2, with the addition of two other scenarios (creating four conditions). In the Gene-Impulse scenario, participants learned that Patrick’s gene had created a stronger impulse to act violently compared to people without that gene. In the Gene-Control scenario, participants learned that Patrick’s gene had decreased his ability to control his violent impulses relative to people without that gene. Participants answered the same questions regarding the perceived appropriateness of criminal defenses, appropriate sentence length, and perceptions of conscious control, criminal intent, and criminal responsibility on the part of Patrick. We changed our sentence length measure from a scale with five-year increments to one with one-year increments. This brought our measure more in line with other work that have demonstrated significant differences involving genetic manipulations (e.g. Aspinwall et al., 2012; Dar-Nimrod et al., 2011). Participants also received the same attribution style questions used in Study 2. We also reinserted the corrected question about perceived recidivism from Study 1 (see Footnote 2); but the remainder of the questions were left out of Study 3 as the two previous studies demonstrated strongly that we should not expect differences between the Genetic and Environmental conditions for them. Given Study 3’s research question, the Control condition from previous studies was not relevant, and was dropped.

Results

Defense claims. There was a main effect of Condition on the applicability of the insanity defense, F(3,290)=4.53, p=.004, partial η2=0.04, CI.95=[0.00,0.09]. It was endorsed less in the Environmental condition (M=2.06, SE=0.22) than in the Genetic (M=2.92, SE=0.21), p=.005, d=0.46, CI.95=[0.13,0.78], Gene-Impulse (M=2.89, SE=0.23), p=.009, d=0.44, CI.95=[0.11,0.77], and Gene-Control conditions (M=3.09, SE=0.22), p=.001, d=0.55, CI.95=[0.22,0.88]. The latter three were not significantly different, ps>.500.

There was a main effect of Condition for the diminished capacity defense, F(3,289)=12.82, p<.001, partial η2=0.12, CI.95=[0.05,0.18]. It was endorsed less in the Environmental condition (M=2.96, SE=0.22) than in the Genetic (M=4.56, SE=0.22), p<.001, d=0.83, CI.95=[0.50,1.16], Gene-Impulse (M=4.62, SE=0.23), p<.001, d=0.87, CI.95=[0.53,1.21], and Gene-Control conditions (M=4.45, SE=0.22), p<.001, d=0.78, CI.95=[0.44,1.11]. The latter three were not significantly different, ps>.500.

There was no main effect of Condition on the intoxication defense, F(3,289)=1.10, p=.349, partial η2=0.01, CI.95=[0.00,0.04].

Sentencing. Unlike Studies 1 and 2, but replicating other research (e.g., Aspinwall et al., 2012; Dar-Nimrod et al., 2011), there was a main effect of Condition, F(3,291)=7.48, p<.001, partial η2=0.07, CI.95=[0.02,0.13]. It was significantly longer in the Environmental condition (M=29.70, SE=1.70) than in the Genetic (M=23.96, SE=1.68), p=.017, d=-0.39, CI.95=[-0.71,-0.07], Gene-Impulse (M=18.76, SE=1.77), p<.001, d=-0.74, CI.95=[-1.08,-0.40], and Gene-Control conditions (M=21.09, SE=1.73), p<.001, d=-0.58, CI.95=[-0.91,-0.25]. The only other significantly different contrast was between the Genetic and Gene-Impulse conditions, p=.034, d=0.35, CI.95=[0.02,0.68]. The remaining contrasts were not significant, p’s>.200.

Perpetrator-relevant perceptions. Replicating previous studies, there was a significant effect of Condition on perceived control, F(3,290)=10.82, p<.001, partial η2=0.10, CI.95=[0.04,0.16]. It was significantly higher in the Environmental condition (M=6.16, SE=0.15) than in the Genetic (M=5.10, SE=0.15), p<.001, d=-0.78, CI.95=[-1.11,-0.45], Gene-Impulse (M=5.17, SE=0.16), p<.001, d=-0.74, CI.95=[-1.07,-0.40], and Gene-Control conditions (M=5.14, SE=0.16), p<.001, d=-0.76, CI.95=[-1.09,-0.42]. The latter three were not significantly different, ps>.700.

Condition did not affect perceived recidivism or intention to kill Fs<2.00, ps>.200, partial η2s=0.01.

Condition significantly affected criminal responsibility, F(3,291)=4.31, p=.005, partial η2=0.04, CI.95=[0.00,0.09]. It was significantly lower in the Environmental condition (M=6.37, SE=0.14) than in the Genetic (M=5.83, SE=0.14), p=.005, d=-0.45, CI.95=[-0.77,-0.13], Gene-Impulse (M=5.75, SE=0.14), p=.002, d=-0.51, CI.95=[-0.84,-0.18], and Gene-Control conditions (M=5.81, SE=0.14), p=.005, d=-0.47, CI.95=[-0.79,-0.14]. The latter three were not significantly different, ps>.700.

Causal attributions. Condition significantly affected causal locus, F(3,290)=8.65, p<.001, partial η2=0.08, CI.95=[0.03,0.14]. It is significantly more external (lower) in the Environmental condition (M=4.72, SE=0.17) than in the Genetic condition (M=5.80, SE=0.17), p<.001, d=0.74, CI.95=[0.41,1.06], Gene-Impulse (M=5.57, SE=0.17), p=.001, d=0.58, CI.95=[0.24,0.91], and Gene-Control conditions (M=5.73, SE=0.17), p<.001, d=0.69, CI.95=[0.35,1.02]. The latter three were not significantly different, ps>.300.

Condition did not affect causal stability, F(3,291)=0.67, p=.568, partial η2=.01, CI.95=[0.00,0.03].

Condition significantly affected causal specificity, F(3,289)=11.50, p<.001, partial η2=0.11, CI.95=[0.04,0.17]. Participants saw less causal specificity in the Environmental condition (M=2.25, SE=0.18) than in the Genetic (M=3.08, SE=0.18), p=.001, d=0.54, CI.95=[0.21,0.86], Gene-Impulse (M=3.74, SE=0.19), p<.001, d=0.96, CI.95=[0.62,1.30], and Gene-Control conditions (M=3.22, SE=0.18), p<.001, d=0.63, CI.95=[0.30,0.96]. The Gene-Impulse condition also elicited greater causal specificity than the Genetic, p=.011, d=-0.43, CI.95=[-0.75,-0.10], and Gene-Control conditions, p=.047, d=0.33, CI.95=[0.00,0.66]. The latter two did not differ significantly, p=.572.

Condition did not affect causal control, F(3,289)=0.60, p=.613, partial η2=0.01, CI.95=[0.00,0.02].

Condition marginally affected causal malleability, F(3,290)=2.15, p=.095, partial η2=0.02, CI.95=[0.00,0.06]. Participants saw greater causal malleability in the Environmental condition (M=4.68, SE=0.17) than in the Gene-Impulse condition, (M=4.06, SE=0.18), p=.014, d=-0.41, CI.95=[-0.74,-0.08], but not in the other conditions, nor do the other conditions differ from each other, ps>.100.

Mediation analyses. Mirroring Studies 1 and 2, we performed a mediation analysis on the Genetic-Environmental contrast predicting sentencing through perceived conscious control. Participants in the Environmental condition perceived greater conscious control than those in the Genetic condition, =0.75, t(286)=4.82, p<.001, CI.95=[0.46,1.04]. This, in turn, predicted lengthier sentences, =0.52, t(285)=10.08, p<.001, CI.95=[0.43,0.60]. The indirect effect was significant, =0.35, CI.95=[0.23,0.54]. Compared to the non-mediated direct effect, =0.35, t(287)=2.30, p=.022, CI.95=[0.04,0.67], the mediated direct effect becomes non-significant, =-0.04, t(285)=-0.28, p=.779, CI.95=[-0.33,0.25], indicating full mediation (see Figure 2A).

Another mediation analysis examining the indirect effect of the Genetic-Environmental contrast on sentencing through internal attributions replicated Study 2’s results. Participants in the Environmental condition made fewer internal attributions than those in the Genetic condition, =-0.71, t(286)=-4.50, p<.001, CI.95=[-1.01,-0.39]. Having stronger internal attributions also predicted lengthier sentences, =0.20, t(285)=3.54, p<.001, CI.95=[0.09,0.31]. The indirect effect was significant, =-0.14, CI.95=[-0.24,-0.06]. The mediated direct effect, =0.49, t(285)=3.10, p=.002, CI.95=[0.18,0.79], was larger than the non-mediated direct effect, =0.35, also replicating the suppression effect found in Study 2 (see Figure 2B).

Reflecting our findings from Study 1, the Genetic-Environmental contrast on sentencing through the perceived recidivism was not significant =-0.24, t(287)=-1.57, p=.117, CI.95=[-0.54,0.05], but perceived recidivism, again, predicted lengthier sentences, =0.37, t(286)=6.73, p<.001, CI.95=[0.24,0.48]. The indirect effect was in the predicted direction, =-0.09, CI.95=[-0.21,0.02]. The mediated direct effect, =0.44, t(286)=3.08, p=.002, CI.95=[0.16,0.73], became larger than the non-mediated direct effect, =0.35 as reported above, also replicating the suppression effect found in Study 1 (see Figure 2C).

Discussion

Results from Study 3 replicated all significant effects of Condition between the Genetic and Environmental conditions in Studies 1 and 2. Most importantly, we demonstrated a reliable mediation between Genetic versus Environmental explanations and prescribed sentence length through perceived conscious control, internal attributions, and marginally significantly through perceived recidivism. The latter two analyses also showed a replicable suppression effect across our studies (see Paulhus, Robins, Trzesniewski, & Tracy, 2004). One main difference between this study and the two previous studies is that a change in methodology regarding how participants recommended a sentence yielded a significant difference between the Genetic and Environmental conditions in prescribed prison sentence length, more in line with past research (e.g., Aspinwall et al., 2012; Dar-Nimrod et al., 2011). This suggests the importance of methodological concerns when examining sentencing practices.

Moreover, the fact that our Gene-Impulse and Gene-Control conditions did not generally yield significant differences from the Genetic condition suggests several possibilities. When people encounter genetic behavioral explanations, they do not consider what mechanisms link genotypes to phenotypes, instead defaulting to a simplistic “cause-and-effect” schematic of genetic effects on behavior. Another possibility is that people consider one or both of these mechanisms by default, rendering the three genetic conditions mostly equivalent. It is of note, though, when significant contrasts emerge between these conditions, it was always between the Genetic and the Gene-Impulse conditions, suggesting a greater likelihood that people pair genetic explanations with a lack of control rather than experiencing a greater impulse.

Additional analyses from aggregate dataset

To better estimate the effect sizes of the difference between the Genetic and Environmental conditions in terms of all of our criterion variables, we aggregated participants’ data in these two conditions into one aggregate dataset across the three studies. This resulted in a sample size of 350 participants for our one-way ANCOVAs, with political orientation as the covariate (see Table 1 for all effect sizes from the aggregate dataset).



Defense claims. There was a main effect of Condition for the insanity defense, F(1,340)=27.83, p<.001, partial η2=0.08, CI.95=[0.03,0.13] (see Figure 1D). It was significantly higher in the Genetic condition (M=3.01, SE=0.15) than in the Environmental condition (M=1.90, SE=0.15), d=0.57, CI.95=[0.35,0.79].

There was a main effect of Condition for the diminished capacity defense, F(1,339)=35.30, p<.001, partial η2=0.09, CI.95=[0.04,0.16]. It was significantly higher in the Genetic condition (M=3.65, SE=0.16) than in the Environmental condition (M=2.28, SE=0.16), d=0.64, CI.95=[0.42,0.86].

There was no main effect of Condition for the intoxication defense, F(1,338)=0.91, p=.341, partial η2=0.00, CI.95=[0.00,0.02].

Verdicts. There were no main effects of Condition, Fs<2.00, ps>.200, partial η2s<.01.

Sentencing. Condition significantly affected sentencing, F(1,340)=4.01, p=.046, partial η2=0.01, CI.95=[0.00,0.05]. Sentences were longer in the Environmental condition (M=25.97, SE=1.13) than in the Genetic condition (M=22.79, SE=1.12), d=-0.22, CI.95=[-0.43,-0.00].

Perpetrator-relevant perceptions. Condition significantly affected perceived control, F(1,338)=29.09, p<.001, partial η2=0.08, CI.95=[0.03,0.14]. It was significantly higher in the Environmental condition (M=5.26, SE=0.12) than in the Genetic condition (M=4.37, SE=0.12), d=-0.58, CI.95=[-0.80,-0.37].

Condition significantly affected intent to kill, F(1,339)=5.33, p=.022, partial η2=0.01, CI.95=[0.00,0.05]. It was significantly higher in the Environmental condition (M=3.87, SE=0.14) than in the Genetic condition (M=3.40, SE=0.14), d=-0.25, CI.95=[-0.46,-0.04].

Condition did not affect knowledge, F(1,191)=1.10, p=.295, partial η2=0.01, CI.95=[0.00,0.04].

Condition marginally affected criminal responsibility, F(1,340)=3.05, p=.082, partial η2=0.01, CI.95=[0.00,0.04]. It was higher in the Environmental condition (M=7.45, SE=0.14) than in the Genetic condition (M=7.10, SE=0.14), d=-0.19, CI.95=[-0.40,0.02].

Condition marginally affected perceived recidivism, F(1,230)=3.04, p=.083, partial η2=0.01, CI.95=[0.00,0.06]. It was higher in the Genetic condition (M=5.36, SE=0.13) than in the Environmental condition (M=5.02, SE=0.14), d=0.23, CI.95=[-0.03,0.49].

Causal attributions. Condition significantly affected causal locus, F(1,255)=28.65, p<.001, partial η2=0.10, CI.95=[0.04,0.17]. It was more internal (higher) in the Genetic condition (M=5.66, SE=0.12) than in the Environmental condition (M=4.75, SE=0.12), d=0.67, CI.95=[0.42,0.92].

Condition marginally affected causal stability, F(1,256)=3.45, p=.065, partial η2=0.01, CI.95=[0.00,0.05]. It was more stable in the Genetic condition (M=5.62, SE=0.10) than in the Environmental condition (M=5.36, SE=0.10), d=0.23, CI.95=[-0.01,0.47].

Condition significantly affected causal specificity, F(1,254)=22.19, p<.001, partial η2=0.08, CI.95=[0.03,0.15]. It was more specific in the Genetic condition (M=2.80, SE=0.11) than in the Environmental condition (M=2.07, SE=0.11), d=0.59, CI.95=[0.34,0.84].

Condition did not affect causal control, F(1,255)=0.24, p=.624, partial η2=0.00, CI.95=[0.00,0.02].

Condition significantly affected causal malleability F(1,256)=11.66, p=.001, partial η2=0.04, CI.95=[0.01,0.10]. It was more malleable in the Environmental condition (M=4.92, SE=0.12) than in the Genetic condition (M=4.33, SE=0.12), d=-0.42, CI.95=[-0.67,-0.18].

Mediation analyses. Across all three studies, the Genetic condition was associated with higher ascriptions of conscious control compared to the Environmental condition, =0.56, t(338)=5.39, p<.001, CI.95=[0.36,0.76], which predicts lengthier sentences, =0.30, t(337)=5.69, p<.001, CI.95=[0.18,0.41]. The indirect effect is significant, =0.17, CI.95=[0.08,0.26]. Compared to the non-mediated direct effect,=0.21, t(340)=2.00, p=.041, CI.95=[0.00,0.42], the mediated direct effect is no longer significant, =0.03, t(334)=0.25, p=.799, CI.95=[-0.18,0.23], indicating full mediation (see Figure 2A).

Another mediation analysis found that the Genetic condition leads to stronger internal causal attributions than the Environmental condition, =-0.63, t(255)=-5.35, p<.001, CI.95=[-0.87,-0.40], which predicts lengthier sentences, =0.21, t(254)=3.33, p=.001, CI.95=[0.08,0.34]. The indirect effect is significant, =-0.13, CI.95=[-0.23,-0.05], while the mediated direct effect is even more significant, =0.38, t(254)=3.10, p=.002, CI.95=[0.14,0.61] than the non-mediated direct effect,=0.21, indicating a reliable suppression effect (see Figure 2B).

Finally, a mediation analysis revealed that the perceived recidivism is also important in explaining how the Genetic condition may affect prison sentences differently than the Environmental condition. Specifically, there was a marginal trend for the Genetic condition to elicit greater perceived recidivism than the Environmental condition, =-0.22, t(230)=-1.74, p=.083, CI.95=[-0.48,0.02]. This, in turn, predicted lengthier sentences overall, =0.40, t(229)=6.37, p<.001, CI.95=[0.28,0.51]. The indirect effect of recidivism was in the predicted direction, =-0.09, CI,95=[-0.20,0.01]. Similar to the mediation analyses involving internal attributions, this mediation analysis demonstrates another reliable suppression effect, with the mediated direct effect, =0.34, t(229)=2.74, p=.006, CI.95=[0.11,0.57], being stronger than the non-mediated direct effect, =0.21 (see Figure 2C).

Path analysis. We conducted a path analysis using the “lavaan” package in R (Rosseel, 2012) to determine the effect of all three mediators in simultaneously explaining variation in prison sentence length between participants in the Genetic and Environmental conditions. Furthermore, because Aspinwall et al.’s (2012) study revealed that both the mitigating factor of lower perceived conscious control and the aggravating factor of higher perceived recidivism appear to be phenomenologically related, we also predicted a pathway leading from the mitigating factor to the aggravating factor. This model is shown in Figure 3. The correlation matrix for all predictors (including covariates) can be found in Table 2. Our model demonstrates that internal attributions, perceived conscious control, and perceived recidivism all mediate the path from the contrast between the Genetic and Environmental conditions and prison sentence length, χ2(2)=7.62, p=.022, CFI=0.97, RMSEA=0.09, SRMR=0.03.

General Discussion

We examined whether genetic explanations of criminal behavior affect legal decision-making differently than environmental explanations of criminal behavior, resulting in findings that carry important implications.

One particularly important implication of these studies pertains to mens rea, a legal concept pertaining to one’s malicious intent and volition to commit a crime, and is necessary for a conviction (American Law Institute, 1962). Perceiving someone’s actions as being beyond their control likely leads to the perception that the perpetrator lacked mens rea. Indeed, many defense claims, some of which we adopted for our study, are meant to mitigate mens rea by arguing for lack of intention and/or control. In line with previous work (Dar-Nimrod & Heine, 2011), our studies support the idea that genetic explanations more robustly impact mens rea-related attributions differently than other kinds of explanations. It is important to keep in mind that the means for participants’ acceptance of defenses were generally below the midpoints of the scales, suggesting that participants did not perceive them to be highly applicable; however, the pattern of differences across conditions informs our hypotheses. As such, across all three studies, genetic explanations reliably increased the perceived applicability of both the Insanity and Diminished Capacity defenses compared to environmental explanations involving one’s upbringing. Related to these findings, compared to environmental explanations, genetic explanations overall lowered one’s perceptions of the perpetrator’s control over his actions as well as his perceived intention to harm the victim; but they do not impact perceptions of whether the perpetrator knew the potential outcome of his actions. Altogether, these results suggest that genetic explanations diminish one’s agency – despite knowing that his actions could have killed the victim, he neither was able to control his behavior, nor did he really intend to kill the victim. Both of these beliefs may partially mitigate a person’s perceived guilt. Unfortunately, our research cannot determine whether people understand genetic causes as impacting one’s impulses or one’s ability to inhibit their impulses. Additional research is needed.

Despite the impact of genetic explanations on general perceptions of mens rea, they do not appear to explicitly affect perceptions of criminal responsibility or ultimate verdicts compared to environmental explanations. Across the 3 studies, only Study 3 revealed lower levels of criminal responsibility in response to a genetic explanation compared to either an environmental explanation or people’s default perception. Furthermore, participants’ perceptions of final verdicts were not affected across the three studies, suggesting that genetic explanations may not affect ultimate legal judgments. While this may, prima facie, appear to be the case, the prescribed sentence length provides important insight into potentially latent mechanisms. Specifically, our studies revealed three mechanisms. First, genetic, versus environmental, explanations led to lower perceptions of conscious behavioral control, which predicted lighter prison sentences. On the other hand, genetic explanations triggered more internal causal attributions and expectations of reoffending than environmental explanations, subsequently predicting lengthier prison sentences, potentially mapping onto distinctions between retributive versus rehabilitative punishment due to one’s implicit theories of behavior (Plaks, Levy, & Dweck, 2009). Genetic evidence may, thus, force people to concurrently consider and reconcile this opposing combination of mitigating and aggravating points, which may not necessarily neutralize each other (having done so in Studies 1 and 2, but not in Study 3). These findings are reminiscent of studies examining the effect of neuroimaging evidence (Schweitzer, et al., 2011), potentially stemming from genetic evidence being seen as being a double-edged sword (Dar-Nimrod & Heine, 2011). Overall, our path analysis supports these ideas, revealing that all three mediators simultaneously impact people’s deliberation over sentencing as a result of being exposed to genetic versus environmental explanations.

Also important is the finding that genetic explanations led to different types of causal attributions than environmental explanations. Overall, genetic causes are seen as being more internal, more stable, more domain-specific, and less malleable than environmental causes, but not more controllable. This different pattern of attributions may help future researchers develop new hypotheses for how genetic and environmental explanations may affect legal decision-making.

The replication of most of our findings across three studies suggest that genetic explanations cause people to think differently than environmental explanations do, particularly in legal contexts, by reliably affecting certain aspects of the legal decision-making process, such as the applicability of defense claims, certain mens rea-relevant perceptions, and sentencing. Furthermore, the penchant for genetic explanations, more than environmental explanations, to differ from responses in the control condition suggests that in many cases, genetic explanations are the engines that drive the shifts in people’s thinking away from their default ways of making sense of criminal responsibility.

Future work should compare genetic explanations to a wider range of alternative behavioral explanations, such as neurological explanations of criminal behaviour (Gazzaniga, 2011), to determine the extent to which genes hold special explanatory power. Given that work by Appelbaum, Scurich, and Raad (2015) suggests that biological explanations may be equally essentialized, our results potentially highlight key differences that are more applicable to delineations of “nature”- versus “nurture”-type explanations.

Limitations

Our three studies required participants to make legal decisions based on limited facts presented in a brief vignette. It remains uncertain whether such effects are generalizable to trial settings during which jurors encounter days’ worth of testimony and information. Furthermore, as jury decisions are usually made in groups, they may also be subject to the influence of group processes such as polarization (Isenberg, 1986) and groupthink (Janis, 1972), neither of which we examined here.

There may also be external validity concerns as our online and college samples are not representative of their respective populations; however, given that our results mimic those from a large representative sample (Appelbaum et al., 2015), our results should be generalizable. In addition, as this research was all conducted with North Americans, it is unclear whether the results would generalize to other cultures. Some research indicates that East Asians are less likely to attend to dispositions (Choi, Nisbett, & Norenzayan, 1999) and to have weaker entity theories of self (Heine, et al., 2001) than North Americans, suggesting that East Asians might be affected by genetic attributions less, although we could not test for the effect of ethnicity in our studies.

Furthermore, the use of single items as dependent variables may lead to unstable results, suggesting issues with reliability (see Wanous & Reichers, 1996); however, many single items are used to replace lengthier scales while retaining good psychometric properties (e.g. Rammstedt & Rammsayer, 2002; Robins, Hendin, & Trzesniewski, 2001). This, in conjunction with the fact that our results replicate across studies (both significant and non-significant), lead us to believe that our measures captured meaningful variance attributable to our conditions.


  1   2


The database is protected by copyright ©essaydocs.org 2016
send message

    Main page