Physical attractiveness and criminal justice processing: results from a longitudinal sample of youth and young adults

Abstract

A substantial body of research examining the role that attractiveness plays in a wide range of outcomes has revealed that attractiveness is a beneficial characteristic across multiple domains of life, including some related to crime and the criminal justice system. The current study uses these findings as a springboard to examine the potential association between attractiveness and multiple measures of criminal justice processing, including being arrested, being convicted, being sentenced to probation and being incarcerated. Analysis of data drawn from the National Longitudinal Study of Adolescent to Adult Health revealed that more attractive persons were less likely to be arrested and convicted than less attractive persons, but there was no association with odds of being sentenced to probation or incarcerated. Follow-up analyses revealed that the beneficial effect of being attractive was confined solely to females. We discuss possible reasons for these results and provide suggestions for future research.

Introduction

When Casey Anthony, who was accused of murdering her daughter, was found not guilty in 2011, the typical reaction among many viewers of news and news shows was one of speechlessness. After all, many wondered, how could she escape a guilty verdict when the evidence against her seemed to be so overwhelming? The final verdict seemed to be such a disconnect with the evidence as the public understood it that immediately after the verdict was announced there was public outrage, finger pointing and outright attacks leveled against those who may have been responsible for such a verdict. The jury, which had been brought in from a nearby county to Orange County, Florida, for the trial, immediately left after the decision, apparently in fear of a possible mob gathered outside the courthouse (the judge in the case also delayed releasing jurors’ names for several months). Attempts to explain how the jury could have rendered a ‘not guilty’ decision were immediate. In the days, months and even years following the trial, experts and social commentators shared various explanations, thoughts and ideas on how the jury could have possibly returned a not guilty verdict. One inescapable possibility was that her appearance – namely, the fact that she was viewed by many as an attractive defendant – played a pivotal role in the way in which she was viewed by the jury and ultimately led to a verdict in her favor. The belief was that Anthony’s attractiveness led the jury either consciously or unconsciously to treat her with more leniency or to believe her side of the story. Whether attractiveness was a contributing factor in the Anthony trial is unknown, but it remains a viable and interesting extralegal factor that is worthy of empirical research.

A substantial body of research has examined a list of extralegal factors – factors that should be irrelevant to the criminal justice system, but may not be – to determine their potential role in the criminal justice system (Engen & Gainey, 2000; Hagan, 1974; Meyer & Gray, 1997). Entire knowledge bases exist examining the way extralegal factors, such as age, gender and race, might structure how persons are differentially processed through the criminal justice system from arrest through sentencing (Beaver et al., 2013; Steffensmeier, Painter-Davis, & Ulmer, 2017; Steffensmeier, Ulmer, & Kramer, 1998). Comparatively less research, however, has focused on other extralegal factors, such as attractiveness. The current study is designed to build off the existing literature and examine whether attractiveness matters when it comes to being arrested, being convicted, being incarcerated and being sentenced to probation. To do so, data drawn from a nationally representative and longitudinal sample will be analyzed.

Attractiveness and criminal justice processing

There has been a tremendous amount of interest regarding attractiveness and the effects that it has on people’s various life outcomes. Studies have found, for instance, that attractive persons, when compared to less attractive persons, are more successful reproductively (Jokela, 2009), report being healthier across a range of outcomes (Nedelec & Beaver, 2014), fare better when it comes to dating and marriage (Buss, 2003), are more successful in the labor market (Hamermesh & Biddle, 1994) and are viewed as having more successful political careers (Lewis & Bierly, 1990; Rosar, Klein, & Beckers, 2008). Taken together, the available evidence tends to suggest that attractive persons enjoy a greater range of benefits across multiple domains of life.

Of particular importance to the current study is whether attractiveness confers advantage within the criminal justice system. A body of research studies has examined this association. These studies have been heterogeneous, analyzing very different samples, using different methodologies, and examining different criminal justice outcomes. Overall, the findings from these studies tend to suggest that attractiveness is a beneficial characteristic within the criminal justice system, although effect sizes may be relatively weak. Much of this research has been conducted using mock jury trials to isolate the effect of defendants’ attractiveness on verdict decisions. A meta-analysis conducted on 25 such studies, netting a total of 56 effect sizes, was quite revealing (Mazzella & Feingold, 1994). Across these studies, the results of the meta-analysis indicated that mock jurors were less likely to return guilty verdicts for attractive defendants (d = 0.19) and also recommended less punitive punishments for attractive defendants (d = 0.12). However, note that the effect sizes were relatively small for these associations, though some of these effects were moderated by crime type.

In a more recent study, Ahola, Christianson, and Hellström (2009) presented subjects with pictures of persons (males and females) of varying levels of attractiveness along with a description of a crime that they may have committed. Subjects were then asked to determine (i.e., they were the sole assessor) the guilt or innocence of the person along with a number of other characteristics (e.g., the person’s reliability, aggressiveness, ruthlessness, etc.). The results of the analysis indicated that there was a small tendency for more lenient assessments of more attractive females. There was no such effect detected for males.

Other studies that have not used a mock jury or an experimental design have also detected an association between the attractiveness of the defendant and verdicts. In one study observers attended criminal trials and rated the attractiveness of the defendant (Stewart, 1980). These measures of attractiveness were then examined in relation to the sentences and the results across 67 defendants; this revealed a statistically significant association between attractiveness and sentence length. The more attractive defendants received less severe sentences than less attractive defendants (see also Stewart, 1985).

Precisely why attractiveness might be related to criminal justice processing variables is not known at this point, though research findings derived from studies regarding perceptions of attractive people are insightful. A body of research findings has examined whether attractive people are viewed differently from less attractive people. The results generated from these studies have consistently shown that attractive people are perceived as being more intelligent, more socially competent, more mentally stable and more trustworthy (Eagly, Ashmore, Makhijani, & Longo, 1991; Langlois et al., 2000). To use the words of Dion, Berscheid, and Walster (1972, p. 285), people tend to believe ‘what is beautiful is good’.

These ascribed traits are particularly important for those making decisions in the criminal justice system. For instance, law enforcement agents have a great deal of discretion regarding whom to arrest, and often their decisions are structured, in part, by whether they believe the person they are questioning. Absent any clear-cut physical evidence and any background information about the person, they may rely on their perceptions regarding the person’s honesty, believability and mental well-being, perceptions that might be structured partially by the attractiveness of the suspect. Similarly, it is quite possible that jury members find the testimony and/or the plea of attractive defendants much more believable than that of comparatively unattractive defendants, leading to more acquittals of attractive defendants than of less attractive defendants. Other actors in the criminal justice system, such as the district attorney and judge, also may be affected by the perceptions of the defendant, such that attractive defendants would not be prosecuted as harshly and would be more likely to receive more lenient sentences than less attractive defendants. Of course, physical attractiveness is likely to be only one small part of overall social attractiveness, along with speaking ability, social skills, cordiality, and so on. However, taken together, decision-making processes among those in the criminal justice system might be affected – in part – by the physical attractiveness of the suspect.

Current study

The available evidence to date suggests that more attractive persons are less likely to be processed through the criminal justice and, if they are found guilty, tend to receive lighter sentences. The current study is designed to expand on these studies in three unique ways. First, rather than using a mock jury design as most studies have, we analyze data drawn from a nationally representative and longitudinal study. By using these data, our findings should be more generalizable than those that focus on an artificial scenario. Second, instead of focusing only on one or two criminal justice outcomes (e.g., sentence length, guilt/innocence), we analyze the association between attractiveness and multiple criminal justice outcomes, including arrest, conviction, incarceration and probation. By including a broader range of criminal justice outcomes, we are able to examine whether attractiveness has differential effects at different stages in the criminal justice system. Third, to measure attractiveness, we use multiple independent raters of attractiveness at different periods in the life course. These observers are independent of (and should have no knowledge of) the criminal justice involvement of the subjects. In this way, the measures of attractiveness are independent of the measures of criminal justice involvement, making for a rigorous assessment of the potential link between attractiveness and criminal justice outcomes.

Method

Data

This study analyzes data drawn from the National Longitudinal Study of Adolescent to Adult Health (Add Health). The Add Health is a longitudinal nationally representative sample of Americans that began data collection in the mid-1990s (1994–1995). The first wave of the survey was collected from over 90,000 adolescents enrolled in 132 middle schools and high schools in the 1994–1995 school year (Udry, 2003). A subsample of approximately 20,000 respondents were then followed up to take part in an in-home survey that has since collected an additional three waves of data. The second wave of the survey was administered in 1996 to approximately 15,000 respondents from the original in-home survey. Questions in the first two waves of the survey had a broad focus and asked about daily activities, relationships with parents and peers, contact with the criminal justice system and delinquent behavior. The third wave of the survey was collected five years later (2001–2002) from approximately 15,000 respondents when the majority of the respondents were between the ages of 18 to 26 years. The fourth wave of the survey was administered six years later in 2008 when the majority of the respondents were between the ages of 24 and 32 years (Harris et al., 2003). Questions in the later waves of the survey were adjusted to be age appropriate for the now adult sample and contained questions concerning labor market participation, educational achievement, criminal involvement and contact with the criminal justice system. Along with the questionnaire, information about the respondents was also collected according to the interviewer’s observations at each of the four waves. In particular, at each of the four waves the interviewers were asked to rate the respondents’ levels of physical attractiveness.

Measures

Outcome variables

Criminal involvement

Criminal involvement was assessed using items designed to tap delinquent and criminal involvement over all four waves of data. Each of the four waves of data contains a battery of questions concerning involvement in various types of criminal behavior. Respondents, for instance, were asked how often they stole something worth more than 50 dollars, broke into a building to steal something, were involved in a group fight, deliberately damaged property or used a weapon to get something in the last 12 months. Responses to each of these items were summed together at each of the four waves to assess criminal involvement at each wave. Then, dichotomous indicators were created to indicate whether respondents reported criminal activity at each of the waves. The dichotomous indicators for criminal involvement at each of the waves were then summed together to create a measure of criminal involvement over the four waves of available data. Values for the criminal involvement measure span from 0 to 4 with higher numbers reflecting criminal involvement at more waves. Descriptive statistics for this variable and all the other variables included in the analyses are presented in Table 1.

Table 1.

Descriptive statistics for scales and variables included in the analyses.

	M	%	SD	Range	N
Attractiveness average	3.525		0.522	1.750–5.000	9919
Criminal involvement	2.022		1.221	0–4	11,037
Arrest
Yes		27.83			4306
No		72.17			11,168
Arrest number	0.777		2.445	0–62	15,474
Conviction (arrested only)
Yes		44.31			1900
No		55.69			2388
Incarceration
Yes		54.57			2341
No		45.43			1949
Probation
Yes		47.33			2030
No		52.67			2259
Age	16.153		1.739	12–21	20,492
Sex
Male		49.19			10,087
Female		50.81			10,420
Race
Non-white		37.63			7711
White		62.37			12,783
Low SES
Yes		9.61			1669
No		90.39			15,702

Note. SES = socioeconomic status.

Arrest

Arrest was measured in two ways for the analyses of these study. First, a single item administered during Wave 4 was used as a dichotomous indictor or arrest. Specifically, respondents were asked whether they had ever been arrested. Responses to this item are coded so that 0 = no and 1 = yes.

Second, a measure of number of arrests was constructed using three additional items assessed at Wave 4. Respondents who indicated that they had been arrested in the original item were then asked to indicate how many times they had been arrested. Response categories for this item were coded so that 1 = once and 2 = more than once. Then, respondents who indicated that they had been arrested more than once were asked how many times they were arrested before and after their 18^th birthdays. Number of arrests was then calculated by summing the number of arrests before and after the respondents’ 18th birthdays, and respondents who indicated that they had only been arrested once were coded as 1s. Respondents who indicated that they had not been arrested in the original question were coded as 0s.

Conviction

Conviction was measured using a single item asked at Wave 4. Respondents, for example, were asked whether they had ever been convicted. Responses for this item are coded so that 0 = no and 1 = yes.

Incarceration

Incarceration was measured using a single item contained in the fourth wave of the survey. Specifically, respondents were asked to indicate whether they had ever been incarcerated. Responses to this item are coded so that 0 = no and 1 = yes.

Probation

Probation was measured as a dichotomous indicator at Wave 4. Respondents, for instance, were asked to indicate whether they had ever been sentenced to probation (0 = no, 1 = yes).

Predictor variable

Attractiveness

Attractiveness was measured using interviewers’ assessments of respondent attractiveness at all four waves of data. Interviewers, for instance, were asked to indicate ‘How physically attractive is the respondent’ at each of the four waves. Responses to these items were coded so that 1 = very unattractive, 2 = unattractive, 3 = about average, 4 = attractive and 5 = very attractive. Interviewer responses concerning respondent attractiveness were then averaged for each of the respondents across all four waves to create a measure of average attractiveness. This average attractiveness measure is coded so that higher values represent greater levels of average attractiveness. These reviewer-rated attractiveness scales are similar to previous measures of attractiveness used with Add Health data (Mocan & Tekin, 2010; Nedelec & Beaver, 2014).

Control variables

The analyses of this study were conducted using four separate control variables. First, age was measured continuously in years at Wave 1. Second, sex was measured dichotomously where 0 = female and 1 = male. Third, race was measured dichotomously where 0 = white and 1 = nonwhite. Finally, socioeconomic status (SES) was measured dichotomously using a single item administered during Wave 1. For this item respondent’s parents were asked to indicate whether they received public assistance such as welfare. Responses to this item are coded so that 0 = no and 1 = yes.

Analytic strategy

The analytic strategy for these analyses unfolded over several steps. First, ordinary least squares (OLS) regression was used to assess the relationship between attractiveness and criminal involvement for the full sample and for males and females separately. Second, logistic regression was used to assess the relationship between attractiveness and odds of arrest with minimal controls. Then, the relationship between attractiveness and odds of arrest was tested again after including criminal involvement as a control variable. Third, logistic regression was used to assess the relationship between attractiveness and odds of conviction using minimal controls among respondents who reported having at least one arrest. Then, number of arrests was added into the analyses to test whether the relationship between attractiveness and odds of arrest is affected by number of arrests. Fourth, logistic regression was used to assess the relationship between attractiveness and the odds of incarceration in respondents who reported having received at least one conviction. Fifth, logistic regression was used to assess the relationship between attractiveness and odds of being sentenced to probation among respondents who had received at least one conviction. Finally, the relationships between attractiveness, odds of arrest, odds of conviction, odds of incarceration and odds of being sentenced to probation were estimated separately for males and females.

Results

The analyses began by examining the relationship between attractiveness and criminal involvement in the full sample and in males and females separately. As can be seen in Table 2, attractiveness is negatively associated with criminal involvement in the full sample. In addition, the relationship between attractiveness and criminal involvement is negative and statistically significant in females. Importantly, the effect sizes for these associations are relatively small (βs range between –.031 and –.027). Moreover, the relationship between attractiveness and criminal involvement is not statistically significant for males.

Table 2.

OLS regression models for the association between attractiveness and criminal involvement.

	Full sample (N = 8484)		Males (N = 3770)		Females (N = 4714)
	Coeff	β	Coeff	β	Coeff	β
Predictor variables
Attractiveness	−.064	−.027*	−.052	−.019	−.065	−.031*
	(.025)		(.044)		(.031)
Controls
Age	−.061	−.080*	−.072	−.089*	−.055	−.078*
	(.008)		(.013)		(.010)
Sex	.477	.193*		—		—
	(.027)
Race	.067	.025*	−.085	−.030	.178	.075*
	(.028)		(.046)		(.035)
SES	−.038	−.009	−.079	−.016	−.022	−.006
	(.047)		(.080)		(.056)

Note. OLS = ordinary least squares; coeff = coefficient; SES = socioeconomic status.

Standard errors in parentheses.

*p < .05.

Next, we examined the relationship between attractiveness and odds of arrest. Table 3 reveals that attractiveness is negatively associated with the odds of arrest. In addition, the second model indicates that attractiveness maintains a significant and negative association with the odds of being arrested even after controlling for criminal involvement. These findings suggest that more attractive people are less likely to be arrested.

Table 3.

Logistic regression models for the associations between attractiveness and the odds of arrest.

	Odds of arrest (N = 8792)		Odds of arrest (N = 8484)
	Coeff	OR	Coeff	OR
Predictor variables
Attractiveness	−.153	0.858*	−.112	0.894*
	(.043)		(.047)
Criminal involvement	—		.407	1.502*
			(.033)
Controls
Age	.002	1.002	.034	1.034*
	(.016)		(.017)
Sex	1.201	3.324*	1.045	2.843*
	(.172)		(.155)
Race	.153	1.166*	.142	1.152*
	(.062)		(.065)
SES	.468	1.597*	.497	1.644
	(.135)		(.146)

Note. OR = odds ratio; coeff = coefficient; SES = socioeconomic status.

Standard errors in parentheses.

*p < .05.

Given the significant relationship between attractiveness and odds of arrest, we then tested whether attractiveness is associated with odds of conviction among respondents reporting at least one arrest. The first model of Table 4 reveals that attractiveness is negatively associated with the odds of conviction. Model 2 indicates that attractiveness is negatively associated with odds of conviction even after controlling for number of arrests. Together, these findings indicate that more attractive people are significantly less likely to be convicted upon being arrested.

Table 4.

Logistic regression models for the associations between attractiveness and the odds of conviction.

	Odds of conviction (N = 2335)		Odds of conviction (N = 2335)
	Coeff	OR	Coeff	OR
Predictor variables
Attractiveness	−.255	0.775*	−.200	0.819*
	(.067)		(.072)
Arrest number	—		.173	1.188*
			(.023)
Controls
Age	.004	1.004	.005	1.005
	(.026)		(.026)
Sex	.504	1.655*	.351	1.420*
	(.152)		(.134)
Race	−.263	0.769*	−.300	0.740*
	(.069)		(.068)
SES	−.017	0.983	−.058	0.944
	(.134)		(.132)

Note. OR = odds ratio; coeff = coefficient; SES = socioeconomic status.

Standard errors in parentheses.

*p < .05.

After establishing that attractiveness is significantly associated with the odds of conviction, we then test whether attractiveness affects the odds of being incarcerated or sentenced to probation among convicted individuals. As can be seen in Table 5, attractiveness does not appear to be significantly associated with the odds of being incarcerated or being sentenced to probation.

Table 5.

Logistic regression models for the associations between attractiveness and the odds of probation and incarceration.

	Odds of probation (N = 1036)		Odds of incarceration (N = 1037)
	Coeff	OR	Coeff	OR
Predictor variables
Attractiveness	−.087	0.917	−.164	0.849
	(.128)		(.117)
Controls
Age	−.028	0.972	.043	1.043
	(.039)		(.042)
Sex	.493	1.638*	.382	1.465*
	(.242)		(.216)
Race	.054	1.055	.301	1.352*
	(.155)		(.199)
SES	−.163	0.849	.569	1.766*
	(.181)		(.417)

Note. OR = odds ratio; coeff = coefficient; SES = socioeconomic status.

Standard errors in parentheses.

*p < .05.

We next examined whether these relationships differed based on the sex of the subject. Table 6 presents the results of the analyses examining the relationships between attractiveness and odds of arrest for males and females separately. As can be seen, attractiveness appears to be negatively and significantly associated with odds of arrest for females even when controlling for criminal involvement. However, the relationship between attractiveness and odds of arrest does not appear to be significant in males. These findings indicate that the significant relationship between attractiveness and odds of arrest may only be present for females.

Table 6.

Logistic regression models for the association between attractiveness and the odds of arrest for males and females.

	Males (N = 3960)		Males (N = 3770)		Females (N = 4832)		Females (N = 4714)
	Coeff	OR	Coeff	OR	Coeff	OR	Coeff	OR
Predictor variables
Attractiveness	−.103	0.902	−.072	0.930	−.214	0.807*	−.162	0.850*
	(.061)		(.066)		(.060)		(.066)
Criminal involvement	—		.383	1.465*	—		.450	1.569*
			(.040)				(.056)
Controls
Age	.039	1.041*	.074	1.077*	−.055	0.946*	−.025	0.975
	(.021)		(.023)		(.023)		(.025)
Race	.125	1.133	.166	1.181*	.200	1.221*	.114	1.120
	(.080)		(.089)		(.100)		(.096)
SES	.467	1.596*	.536	1.709*	.456	1.578*	.457	1.579*
	(.193)		(.219)		(.187)		(.198)

Note. OR = odds ratio; coeff = coefficient; SES = socioeconomic status.

Standard errors in parentheses.

*p < .05.

Table 7 presents the analyses of the relationship between attractiveness and criminal justice outcomes for males. Table 7 reveals that attractiveness is not significantly associated with odds of conviction, odds of incarceration or odds of being sentenced to probation in males.

Table 7.

Logistic regression models for the association between attractiveness and the odds of conviction, incarceration and probation for males.

	Odds of conviction (N = 1550)		Odds of conviction (N = 1550)		Odds of probation (N = 758)		Odds of incarceration (N = 758)
	Coeff	OR	Coeff	OR	Coeff	OR	Coeff	OR
Predictor variables
Attractiveness	−.150	0.860	−.082	0.922	−.294	0.745	−.244	0.783
	(.093)		(.103)		(.130)		(.134)
Number of arrests	—		.168	1.183*	—		—
			(.025)
Controls
Age	−.006	0.994	−.008	0.992	−.044	0.957	.038	1.039
	(.030)		(.031)		(.047)		(.050)
Race	−.234	0.791*	−.256	0.774*	.139	1.150	.369	1.446*
	(.087)		(.087)		(.205)		(.258)
SES	.192	1.212	.131	1.140	−.155	0.856	.745	2.106*
	(.210)		(.204)		(.220)		(.630)

Note. OR = odds ratio; coeff = coefficient; SES = socioeconomic status.

Standard errors in parentheses.

*p < .05.

Finally, Table 8 presents the analyses of the relationship between attractiveness and criminal justice outcomes for females. As can be seen, attractiveness is negatively and significantly associated with the odds of conviction for females even when controlling for criminal involvement. In contrast, further examination of Table 8 reveals that attractiveness is not significantly associated with odds of incarceration or odds of being sentenced to probation for females. These results reveal that attractiveness seems to be associated with several criminal justice outcomes in females.

Table 8.

Logistic regression models for the association between attractiveness and the odds of conviction, incarceration and probation for females.

	Odds of conviction (N = 785)		Odds of conviction (N = 785)		Odds of probation (N = 278)		Odds of incarceration (N = 279)
	Coeff	OR	Coeff	OR	Coeff	OR	Coeff	OR
Predictor variables
Attractiveness	−.455	0.634*	−.416	0.659*	.279	1.322	−.047	0.954
	(.092)		(.097)		(.319)		(.227)
Arrest number	—		.195	1.216*	—		—
			(.055)
Controls
Age	.025	1.025	.036	1.037	.007	1.007	.055	1.057
	(.049)		(.051)		(.074)		(.078)
Race	−.307	0.735*	−.389	0.678*	−.077	0.926	.201	1.223
	(.117)		(.111)		(.246)		(.327)
SES	−.383	0.682	−.367	0.693	−.094	0.910	.271	1.311
	(.159)		(.163)		(.358)		(.529)

Note. OR = odds ratio; coeff = coefficient; SES = socioeconomic status.

Standard errors in parentheses.

*p < .05.

Discussion

There has been a great deal of research focused on examining whether physical attractiveness confers any advantages to life outcomes. Overall, the findings generated from this body of research have revealed that persons who are viewed as being more physically attractive receive certain benefits in life that less attractive people do not (Dion et al., 1972; Eagly et al., 1991; Langlois et al., 2000). The current study sought to extend these findings by examining whether physical attractiveness was associated with differential processing through the criminal justice system.

The results generated from these analyses revealed one key finding – namely, and consistent with some existing literature (e.g., Ahola et al., 2009; Mazzella & Feingold, 1994), that average attractiveness confers a leniency effect when it comes to the odds of being arrested and convicted. What is important to note, however, is that additional analyses revealed that the attractiveness effect was confined to females, such that females who were rated as being more attractive were less likely to be arrested and convicted than females who were rated as less attractive. Attractiveness had no effect on males. This is an interesting finding as it shows that attractiveness is an extralegal factor that has application to females, but not males. Precisely why this is the case is not known but should be explored in future research. It is possible that this effect is confined solely to females as the majority of actors with discretion in the criminal justice system (e.g., law enforcement agents) are males and thus they may be swayed, either consciously or unconsciously, by the attractiveness of females. The same pattern does not parallel with males. As for now, however, this remains an open-empirical issue awaiting future research.

The good news is that, although these effects appear to be statistically significant, they are very small in effect size, even for females. Referring back to the Casey Anthony trial, were this a typical case (which admittedly it was not), based on the current evidence her attractiveness may have played a very small but probably not determinative influence on jury decisions. The influence of attractiveness may be expected to be most likely for cases that are quite marginal, wherein jurors are otherwise demonstrating difficulty coming to a decision. It is also possible that news media narratives that overly focus on attractiveness may distract from more salient aspects of jury decision-making.

These findings provide some evidence that physical attractiveness is associated with more lenient outcomes in the criminal justice system. Even so, there are a number of limitations with the study that should be addressed in future research. First, the measure of attractiveness is based on interviewer observations made by a single interviewer at each of the four waves. As a result, the ratings of attractiveness of each respondent may be influenced in part by their interviewer, with some interviewers being more likely to rate certain individuals as more attractive. However, since the attractiveness measure employed in these analyses used an average of interviewer ratings across all four waves, the overall attractiveness measure should be relatively valid and reliable as it is not based on a sole rater at a single point in time. Second, the Add Health is a nationally representative sample of American youth, and, as a result, the base rate for the criminal justice processing variables is relatively low. Whether the findings reported here would be reproduced with more criminal samples remains to be determined. This may also have truncated the effect sizes seen in the current analysis. Third, the majority of the respondents in this who were involved in crime were involved in relatively minor types of criminal behaviors. Future research should examine whether the association between attractiveness and criminal justice processing varies based on type of crime.

One of the guiding principles of the criminal justice system is that justice should be blind. Precautions have been employed in order to help reduce any biases that might result across demographic characteristics, such as race, class and gender. When it comes to physical attractiveness, however, it does not appear to be the case that justice is blind. Across a range of criminal justice outcomes, including being arrested and being convicted, physically attractive persons are treated much more leniently than less attractive persons. Ways to reduce the biases that might result from physical attractiveness is an important area of inquiry that future research needs to explore in order to more fully ensure that the system operates in a fair and impartial fashion.

Funding Statement

This research was funded by the National Institute of Child Health and Human Development [grant number PO1-HD31921]; with cooperative funding from 17 other agencies.

Ethical standards

Declaration of conflicts of interest

Kevin Beaver has declared no conflicts of interest

Cashen Boccio has declared no conflicts of interest

Sven Smith has declared no conflicts of interest

Chris Ferguson has declared no conflicts of interest

Ethical approval

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

Informed consent

Informed consent was obtained from all individual participants included in the study.

Acknowledgements

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. 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 (http://www.cpc.unc.edu/addhealth). No direct support was received from grant P01-HD31921 for this analysis.