Juror perceptions of the stereotypical violent crime defendant

The criminal stereotype

Stereotyping is an adaptive cognitive mechanism that efficiently categorises and processes social stimuli to decrease the cognitive demand of individuating people (Hilton & Von Hippel, 1996; Mackie et al., 1996; Macrae et al., 1994; Zebrowitz, 1996). Stereotypes also aid decision-making, especially when the available information is limited or ambiguous (Bornstein & Greene, 2011; Brewer, 1996; D. J. Devine et al., 2009; Hugenberg & Sacco, 2008; Kahneman, 2011). Furthermore, stereotype activation can lead to cognitive biases, such as confirmatory information processing (C. S. Jones & Kaplan, 2003), as well as behavioral and personality trait inferences (Macrae & Bodenhausen, 2000; Todorov et al., 2015; Zebrowitz, 1996; Zebrowitz et al., 2007; Zebrowitz & Montepare, 2008).

The stereotypes pertaining to the appearance of the typical violent crime offender include male, Black, unattractive, tall, thin, beady eyes, dark clothes and dark hair (D. J. Devine & Caughlin, 2014; Goldstein et al., 1984; MacLin & Herrera, 2006; Skorinko & Spellman, 2013). Although Black people tend to be associated with crime in general (Eberhardt et al., 2004), certain crimes are perceived as more stereotypical of specific races or ethnicities (e.g. blue collar vs. white collar crime; MacLin & Herrera, 2006; Skorinko & Spellman, 2013; Smalarz et al., 2016; Sunnafrank & Fontes, 1983; Willis-Esqueda, 1997). This phenomenon has been coined race–crime congruency (C. S. Jones & Kaplan, 2003) and consistently influences juror decision-making. The traits most strongly associated with violent, blue collar crimes include: dark skin tone, facial untrustworthiness and tattoos. Therefore, these traits were chosen for the current study.

Skin tone

Prior findings suggest a skin tone bias in which individuals with lighter skin tone are perceived and treated better than those with darker skin tone (E. A. Adams et al., 2016; K. T. Brown et al., 1999; Hunter, 2007; T. Jones, 2000). Studies assessing judgments of Blacks consistently demonstrate increased negative implicit (Livingston & Brewer, 2002) and explicit attitudes (Hagiwara et al., 2012) towards Blacks with darker skin compared to Blacks with lighter skin. Further, both Black and White participants associate dark-skinned Black people with negative Black stereotypes (e.g. criminal, poor, lazy), while light-skinned Black people are associated with positive Black stereotypes (e.g. intelligent, wealthy, rhythmic; Maddox, 2004; Maddox & Gray, 2002).

Light-skinned Black people are also perceived more positively in the criminal justice system, resulting in more lenient sentencing outcomes than those for dark-skinned Black people. In one study, participants exposed to a photo of a dark-skinned perpetrator were more likely to judge a piece of evidence as indicating criminal responsibility and vote guilty on a violent crime charge than participants exposed to a light-skinned perpetrator (Levinson & Young, 2010). Dark-skinned Black convicts have been shown to receive sentences of 200–400 days longer than those of White and light-skinned Black convicts (Burch, 2015). However, King and Johnson (2016) found that dark-skinned Black and darker skinned White people received harsher sentences than lighter skinned individuals of their respective race. Finally, dark-skinned Black murder defendants were more likely to receive a death sentence than light-skinned Black murder defendants (Eberhardt et al., 2006). Much of the literature has focused on interracial differences, but recently it has become apparent that there are disparities within races. Dark-skinned Black people, compared to light-skinned Black people, are at a significant disadvantage. Thus, the severity of the consequences that dark-skinned Black people endure warrants further investigation into the salience of skin tone in criminal stereotype activation.

Facial trustworthiness

Trustworthiness judgments are made spontaneously based on a person’s facial appearance (Willis & Todorov, 2006), are relatively stable over time (Klapper et al., 2016), have high consensus among those making judgments, but may not reflect objective reality (Olivola et al., 2014; Rule et al., 2013). Factors including typicality, facial structure and emotional resemblance influence trustworthiness perceptions. Participants perceive typical faces (i.e. not very attractive, nor very unattractive; Sofer et al., 2015) and those with lower facial width-to-height ratios (Kleisner et al., 2013; Ormiston et al., 2017; Stirrat & Perrett, 2010) as appearing more trustworthy. In addition, the emotion overgeneralization effect suggests that people whose facial appearance resembles an emotion may be perceived to have personality traits associated with that emotion (Montepare & Dobish, 2003; Zebrowitz, 1996). Facial trustworthiness has been empirically associated with an angry–happy continuum, such that the angrier a face appears, the more untrustworthy it is perceived to be, and the happier a face appears, the more trustworthy it is perceived to be (Dong et al., 2014; Flowe, 2012; Oosterhof & Todorov, 2008, 2009).

Importantly, facial trustworthiness has been found to predict criminal appearance and perceived likelihood of committing a crime (Flowe, 2012; Klatt et al., 2016; Korva et al., 2013). Funk and Todorov (2013) had participants rate how criminal a face appears and found that participants rated untrustworthy faces as significantly more criminal looking than trustworthy faces. Wilson and Rule (2015) conducted two studies investigating the role of facial trustworthiness on sentencing outcomes for convicted murderers. In the first study, participants were asked to rate the trustworthiness of the faces of inmates who had been sentenced to either life in prison or sentenced to death. The faces of inmates who had been sentenced to death were more likely to be rated as untrustworthy than those sentenced to life in prison. In the second study, participants were asked to rate the faces of people who had been sentenced to life in prison or to death, but were later exonerated. Wilson and Rule (2015) found that those who had formerly been sentenced to death were less trustworthy looking than those formerly sentenced to life in prison. Further, the authors found that when participants were asked to sentence convicted murderers, their hypothetical assigned sentence matched the inmate’s actual sentence, and facial trustworthiness was a unique predictor of both hypothetical and actual sentences (Wilson & Rule, 2016). Lastly, untrustworthy faces can lead to confirmatory information processing, such that participants do not require as much evidence to be highly confident in a guilty verdict for someone with an untrustworthy face accused of a serious, violent crime (Porter et al., 2010).

Findings from the available literature suggest that facial untrustworthiness contributes to activating a violent criminal stereotype. Specifically, someone with an untrustworthy face may be criminal-stereotype congruent, and once this stereotype is activated, less evidence is required for a guilty verdict. Thus, facial trustworthiness appears to be particularly important in understanding criminal stereotype activation.

Tattoos

Unlike facial appearance or race, tattoos are controllable features, which individuals use as a justification for their stigma towards tattooed individuals (Larsen et al., 2014). Furthermore, the tattoo is the main physical characteristic associated with criminals (MacLin & Herrera, 2006), with 15–32% of inmates being tattooed (Manuel & Retzlaff, 2002; Palermo, 2004). However, tattoos are growing in prevalence in mainstream society, with one estimate stating 24% of adults between the ages of 18–50 have at least one tattoo (Laumann & Derick, 2006), creating a need to understand perceptions of tattooed individuals and specifically alleged criminals.

Studies have found that tattooed individuals are rated more negatively on several traits (i.e. competence, credibility, creativeness, determination, motivation and trustworthiness; Broussard & Harton, 2018; Degelman & Price, 2002; Seiter & Hatch, 2005) and are also perceived to be more thrill-seeking, more prone to boredom, less inhibited (Wohlrab et al., 2009) and more threatening (Johnson & King, 2017). However, in recent years, studies have shown that not all tattoos are perceived equally. Tribal tattoos are categorized as more aggressive and less friendly looking than small, brightly colored tattoos (Burgess & Clark, 2010). Barbed wire and spider web tattoos are perceived as more dangerous and more typical of an offender than a butterfly or star tattoo (K. A. Brown et al., 2018). Lastly, a person with a violence or nudity-themed tattoo is perceived as less trustworthy than a person with a nature-themed tattoo (Timming & Perrett, 2017).

Studies have found that specifically aggressive-themed tattoos are associated with crime, and subsequently trigger criminal stereotypes. It is not the presence of tattoos in and of themselves that creates negative perceptions, but the personal attributions made based on the tattoo. In the study conducted by K. A. Brown et al. (2018), participants exposed to a defendant with a prison-styled tattoo (i.e. a tattoo characterized by imagery relating to the convict’s background or situation in prison that appears primitive or poorly made; see DeMello, 1993 for a review) perceived the defendant to be more dangerous, and more likely to be guilty of assault. Similarly, Funk and Todorov (2013) found perceived criminal appearance of the defendant (i.e. criminal, aggressive, dangerous) to be a mediator between the presence of a facial tattoo and verdict, such that facial tattoos predicted higher criminal appearance of the defendant and subsequently increased likelihood of guilt. In addition, Johnson and King (2017) found that defendants with a facial tattoo were over twice as likely to receive a prison sentence than a defendant without a facial tattoo.

Prior works examining the effects of tattoos on juror decision-making has focused on prison-styled tattoos, creating a gap in the literature on aggressive, but not prison-styled, tattoos. This is an important distinction, since prison-styled tattoos can bias the perceiver towards guilt due to the implication of prior convictions (DeMello, 1993; D. J. Devine & Caughlin, 2014; D. J. Devine et al., 2001; Greene & Dodge, 1995). Aggressive tattoos (i.e. tribal, tigers, snakes, dragons) harbor their own negative prejudices (Zeiler & Kasten, 2016), suggesting a similar effect to prison-styled tattoos, without, however, suggesting prior convictions. Considering the prevalence of tattoos in the criminal population (Manuel & Retzlaff, 2002; Palermo, 2004) and the negative perceptions and outcomes associated with prison-styled tattoos, it is an important extension of the literature to understand the effect of aggressive tattoos on juror perceptions and decision-making.

Prejudice and motivation to respond without prejudice

Negative stereotypes are theorized to have a bi-directional relationship with prejudice (Allport, 1954; P. G. Devine, 1989; E. R. Smith, 1993), and those who have higher levels of prejudice are more likely to apply negative stereotypes in judgments of out-group members (Dotsch et al., 2011; Lepore & Brown, 1997; Locke et al., 1994). Furthermore, scholars have investigated the influence of prejudice on legal decision-making, with most of the literature focusing on racial prejudice (Dovidio & Gaertner, 2004).

As society has shifted away from overt, blatant forms of prejudice, several investigators argue that there is evidence of a new type of subtle, contemporary prejudice (Gaertner & Dovidio, 1986; McConahay, 1986; Sears, 1988) referred to as aversive racism. Aversive racism suggests that White Americans feel as though they are not prejudiced, yet still have internal, unconscious negative feelings and biases towards minorities (Dovidio, 2001; Dovidio & Gaertner, 1998, 2004; Gaertner & Dovidio, 1986), leading to internal conflict (P. G. Devine, 1989; Pearson et al., 2009). Additionally, studies investigating implicit and explicit attitudes towards Black individuals are congruent with aversive racism (Dovidio et al., 2001; Fazio & Dunton, 1997; Hofmann et al., 2005), such that implicit attitudes towards Black individuals were more negative than self-reported, explicit attitudes (Dovidio et al., 1997). Aversive racists are more likely to let prejudice influence their behavior or decisions when the situation is race neutral or the person has non-race-related explanations (Dovidio & Gaertner, 1998, 2000; Gaertner & Dovidio, 1986; Sommers & Ellsworth, 2000, 2001; Son Hing et al., 2008).

In addition, several studies have found a pattern of negative attitudes towards tattooed individuals. For example, through the Implicit Association Test (IAT), Zestcott and colleagues have found that participants have an implicit preference for non-tattooed individuals compared to tattooed individuals (Zestcott et al., 2017, 2018). As previously mentioned, not all tattoos are perceived negatively, and while prejudicial attitudes are not completely eliminated when someone has a tattoo with innocuous content, people have attenuated negative attitudes towards tattoos that are positively valenced (i.e. a heart) compared to negatively valenced tattoos (i.e. tribal-styled; Zestcott et al., 2017).

As previously mentioned, there has been a shift in intergroup attitudes in the United States that has resulted in the rise of egalitarian views of minority groups, as well as the motivation to control biases and appear non-prejudiced (P. G. Devine et al., 2002; Johns et al., 2008; Sinclair et al., 2005). As a result, several scholars have examined the role of motivation in reducing prejudiced behavior and responses (Butz & Plant, 2009; Dunton & Fazio, 1997; Glaser & Knowles, 2008; Plant & Devine, 1998). Plant and Devine (1998) suggested that there are two motivational dimensions: internal and external. When anticipating an interracial interaction, White individuals that are higher in external motivation are more focused on avoiding being perceived as non-prejudiced, while those higher in internal motivation are focused on being approachable and egalitarian (Plant & Devine, 2009; Plant et al., 2010). Along with racial attitudes, these constructs have also been linked to attitudes towards tattooed individuals. Generally, individuals higher in internal motivation score lower on tattoo prejudice and respond more positively to tattooed individuals and those higher in external motivation score higher on tattoo prejudice and respond more negatively to tattooed individuals (Zestcott et al., 2018).

Stimuli and materials

A pilot study was conducted to determine which face and tattoo images to utilise in the current study. Male face images from the Chicago Face Database were chosen to be included in the pilot study based on normed data provided by the authors (Ma et al., 2015). The face and tattoo images were rated on several appearance traits, including perceived facial trustworthiness (see online Supplemental Materials for details). A total of 27 images were developed to be used as stimuli. Three different versions of each face image were used: one with an aggressive tattoo, one with a non-aggressive tattoo and one with no tattoo as a control. The tattoos were digitally edited onto the neck of the male images. Tattoo placement on the neck was chosen for two reasons. First, visible tattoos, such as those on the neck and face, are perceived more negatively and associated more with deviant behavior than non-visible tattoos (J. Adams, 2009; K. A. Brown et al., 2018 ). Second, neck tattoos are relatively more common than facial tattoos in the population (Laumann & Derick, 2006).

The liberation hypothesis suggests when case evidence is ambiguous, and does not clearly favor one side, jurors are more likely to utilise extralegal factors (i.e. factors that are not related to the case, such as stereotypes or personal knowledge) to guide decision-making (D. J. Devine et al., 2009; Hugenberg & Sacco, 2008; Kalven & Zeisel, 1966; Reskin & Visher, 1986; Shoemaker et al., 1973). Thus, the current study included a modified version of a criminal case vignette utilised by K. A. Brown et al. (2018), in which the evidence was meant to be ambiguous (i.e. 75% eyewitness confidence) to elicit participant stereotype use in decision-making. The vignette described an assault committed outside of a bar. Assault was chosen as the crime for the following reasons. A stereotype of those with tattoos is that they are more violent and aggressive (Swami et al., 2015), and assault has also been shown to be a stereotypical crime of people with tattoos (K. A. Brown et al., 2018). As mentioned above, violence and aggression is also a stereotype of Black men, and assault is a blue collar crime, which is stereotypically associated with Black men. Furthermore, some literature suggests that the more violent or serious the crime is, the more likely it is that people will rely on crime stereotypes (Skorinko & Spellman, 2013).

Case judgments

In total, there were 290 (68.1%) ‘not guilty’ and 136 (31.9%) ‘guilty’ verdicts. A generalized linear model with a specified binary logistic link function was conducted to determine whether the physical traits (i.e. tattoos, facial trustworthiness, skin tone), prejudice (i.e. MRS, MSATS) or perceived criminal appearance were significant predictors of dichotomous verdict, while controlling for motivation to respond without prejudice and Afrocentric appearance. MRS was a significant predictor, Wald χ²(1) = 7.42, p = .006. The direction of the beta (b = 0.38) suggests that as racial prejudice increased, the likelihood of a guilty verdict increased. Perceived criminal appearance was a significant predictor, Wald χ²(1) = 30.79, p < .001. The direction of the beta (b = 0.67) suggests that as perceived criminal appearance increased, the likelihood of a guilty verdict increased. Once perceived criminal appearance was entered, MRS became only marginally significant, Wald χ²(1) = 3.71, p = .054, b = 0.28. Tattoos, trustworthiness, skin tone and MSATS were not significant predictors.

Criminal appearance

A generalized linear model analysis was conducted to test the physical traits and prejudice as predictors of perceived criminal appearance. Tattoo possession was a significant predictor, Wald χ²(2) = 15.58, p < .001, such that defendants with aggressive tattoos (M = 4.04, SE = 0.08) and non-aggressive tattoos (M = 4.13, SE = 0.08) appeared more criminal looking than defendants with no tattoo (M = 3.70, SE = 0.08). Trustworthiness was a significant predictor, Wald χ²(2) = 64.87, p < .001, such that untrustworthy defendants appeared more criminal looking (M = 4.47, SE = 0.08) than the neutral (M = 3.84, SE = 0.08) or trustworthy (M = 3.56, SE = 0.08) defendants. Skin tone was a significant predictor, Wald χ²(2) = 37.68, p < .001, such that white defendants appeared more criminal looking (M = 4.25, SE = 0.12) than light-skinned (M = 3.83, SE = 0.08) and dark-skinned defendants (M = 3.79, SE = 0.11). MRS was a significant predictor, Wald χ²(1) = 7.1, p = .008, with the direction of the beta (b = 0.37) suggesting as racial prejudice increased, perceived criminal appearance increased. MSATS was a significant predictor, Wald χ²(1) = 8.04, p = .005, with the direction of the beta (b = 0.01) suggesting that as tattoo prejudice increased, perceived criminal appearance increased.

The interaction between skin tone and MRS was significant, Wald χ²(2) = 34.97, p < .001. More specifically, the interactions between the White and dark-skinned defendant contrast and MRS (b = −0.59), t(412) = −5.20, p < .001, and the White and light-skinned defendant contrast and MRS (b = 0.59), t(412) = 4.75, p < .001, were significant, while the interaction between the light-skinned and dark-skinned defendant contrast and MRS was not (b = 0.01), t(412) = 0.04, p = .969. Examination of the interaction plot (see Figure 1) suggests that at low levels of prejudice, white defendants appeared more criminal than those with light skin or dark skin. As racial prejudice increased, perceived criminal appearance of light-skinned and dark-skinned defendants increased, and criminal appearance of white defendants decreased.

Open in a separate window

Figure 1.

Defendant Skin Tone × MRS interaction on criminal appearance ratings. MRS = Modern Racism Scale.

An additional stepwise linear regression analysis was conducted to estimate the unique variance explained by each physical trait. The covariates were entered into the first model, accounting for approximately 9% of unique variance, F(6, 419) = 6.95, p < .001, R² = .09. Skin tone was entered into the second model, accounting for approximately 0% of unique variance, F_change(1, 418) = 0.11, p = .737, R²_change = .00. However, as was mentioned in the Skin Tone × MRS analysis, the interaction accounted for 6% of unique variance. Facial trustworthiness was entered into the third model, accounting for approximately 10% of unique variance, F_change(1, 417) = 49.4, p < .001, R²_change = .10. Tattoo possession was entered into the fourth model, accounting for approximately 2% of unique variance, F_change(1, 416) = 8.06, p = .005, R²_change = .02.

Limitations

One of the largest limitations of the current research is that the face and tattoo images may not have been effective or realistic enough to elicit the expected effects. There also may have been some unknown variance in the images that resulted in the produced effects. Prior research suggests that there are some physical features that influence juror decision-making that were not considered in the current research design, such as attractiveness (D. J. Devine et al., 2001; Johnson & King, 2017; MacCoun, 1990; Sigall & Ostrove, 1975; Stewart, 1980) and baby-faced appearance (Berry & McArthur, 1986; Montepare & Zebrowitz, 1998; Zebrowitz & McDonald, 1991). Although the manipulations were pilot tested individually, the images were not pilot tested after editing the tattoo images onto the necks. Therefore, it is possible that the images were not realistic to participants, thus influencing their responses in an unanticipated direction.

Additionally, although the case vignette was rated by participants as ambiguous, it is possible that information within the vignette influenced juror decisions in unexpected ways. Past research suggests that when jurors are provided with the instructions to only find the defendant guilty if the evidence proves guilt beyond a reasonable doubt, the influence of prejudice over verdict decisions diminishes (Pfeifer & Ogloff, 1991; Rector et al., 1993). Thus, the ‘guilty beyond a reasonable doubt’ instruction provided in the vignette may have diminished the effect of stereotype activation on decision-making. Additionally, the crime of assault may not be as strongly associated with Black or tattooed individuals as other crimes (i.e. drug crimes, burglary, gang activity; Gordon et al., 1988; Skorinko & Spellman, 2013) to have elicited the race–crime congruency effect in participants. The homogeneity of the sample’s demographic characteristics made it difficult to analyze effects based on participant race and gender. The ecological validity of the findings was also limited, as the design did not consider the effects of juror deliberation. However, findings of Bornstein (1999) suggest that individual mock juror judgments are similar to those of jury deliberation judgments.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee at the University of North Dakota and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.