Examining the Influence of Cosmetics on Jury Decisions

Many studies have examined how defendant characteristics influence jury decisions, but none have investigated the effect of cosmetics. We therefore examined how cosmetics influence jury decisions for young and middle-aged female defendants. In Study 1, participants were more likely to assign guilty verdicts to middle-aged defendants than young defendants and when presented with cosmetics, male participants gave young defendants longer sentences and middle-aged defendants shorter sentences. In Study 2, however, we did not replicate the age or the cosmetics effects on jury sentences, suggesting that comparisons between defendants may have influenced jury decisions in Study 1. Further work is thus still needed, but our two well-powered studies (N = 1127) provide a first exploration into the influence of cosmetics on jury decisions.


Introduction
Trials by jury have existed since the fourth century BC [1] and today are used in many countries. In this legal proceeding, ordinary citizens are asked to hear a case and make decisions regarding its outcome. Jurors are given the ability to convict or exonerate defendants. In the case of guilty verdicts, jurors may also have the authority to issue sentences. In many countries, juries can even administer death penalties.
As a result of the jury's vast power, much research has been devoted to understanding the factors that influence verdicts and sentences (e.g., [2,3]). Among this literature, many studies have investigated how defendant characteristics influence jury decisions. For example, it has been found that for the same crime, women tend to receive significantly less severe sentences than men [4,5]. Research has also demonstrated that jurors impose harsher sentences for Black defendants when compared to White defendants [6,7].
Age perceptions have also been found to influence jury decisions. For instance, Smith and Hed [8] found that older female defendants were judged more harshly than younger defendants. Another study found that in small claims court, as defendants increased in "baby-facedness", they were more likely to win cases [9]. Additionally, Warling and Peterson-Badali [10] found that although the defendant's age did not influence jurors' verdicts of guilt, it did influence the length of the imposed sentences. More specifically, younger defendants were granted shorter sentences than older defendants.
Across gender, race, and age, the defendant's attractiveness level has been found to play a crucial role in jury decisions. Studies using mock trials have consistently found that when compared to unattractive defendants, attractive defendants are convicted less, and if convicted, they are punished less severely for the same crime [5,8,11]. This finding extends to the real-world, where in actual trials,

Study 1 Materials and Methods
Research has found that jury simulations can be generalized to the behavior of real jurors [21], and, therefore, participants were recruited through MTurk, which provides a diverse pool of jury-eligible participants [22]. Ethical approval was received from the Gettysburg College Institutional Review Board. All participants were US citizens located in the United States who provided consent after being presented with the consent information. Previous research has found that gender significantly influences defendant sentencing [2] as well as cosmetics judgements [14,15] and thus, juror sex was recorded. Participants were then instructed that they would be presented with four separate trial vignettes, where they would have to give verdicts and sentences.
Cases were selected from the mock trial sample cases provided by the Michigan Center for Civic Education [23] and slightly altered to fit the purpose of this study (e.g., changing the sex of the defendant from male to female). A total of four cases were used as trial vignettes (see Appendix A). Each trial vignette was presented to participants individually with a picture of the defendant and participants were asked to choose a verdict/sentence out of nine options (ranging from "Not Guilty" to the most severe sentence). The defendant photograph presented was one of four identities (i.e., two young women (19 and 20 years old) and two middle-aged women (48 and 49 years old) in one of the two conditions (i.e., with or without cosmetics)). The cosmetics were professionally applied (e.g., skin, eye, and lip makeup) and the photographs were standardized with direct gaze and neutral expressions.
Each participant was individually presented with all four trial vignettes and whether each face appeared with or without cosmetics was randomized across participants so that participants would Cosmetics 2020, 7, 64 3 of 11 not see the same identity twice. Given no previous study has examined the influence of cosmetics on jury decisions, sample size to be collected (220) was predetermined with a sample size calculator [24] using the average effect size from 25 experiments examining the physical attractiveness of defendants on judgments of mock jurors [5]. One participant was excluded because they submitted their responses twice and those responses differed, and another participant was excluded because they entered in their age as 2 years old. Two hundred and eighteen participants (127 male, 89 female, and 2 other), aged 21-69 years old (Mage = 34.99 years, SD = 9.95) completed the study successfully. Each participant was paid $0.80 US dollars upon completion of the study through MTurk.

Study 1 Results
With each trial vignette, participants were asked to give a verdict/sentence out of nine options (ranging from "Not Guilty" to the most severe sentence). For example, in one of the vignettes the options were: "Not Guilty", "Guilty: 1 year", "Guilty: 2 years", "Guilty: 3 years", "Guilty: 4 years", "Guilty: 5 years", "Guilty: 6 years", "Guilty: 7 years", or "Guilty: 8 years". For the verdict analyses, participants' responses were coded as "Not Guilty" (for those that selected "Not Guilty") or "Guilty" (for those that selected any of the eight guilty options). Using those guidelines, 35.8% of the verdicts given by male participants were "Not Guilty" and 64.2% "Guilty" (see Table 1). For female participants, 34.0% of the verdicts given were "Not Guilty" and 66.0% were "Guilty" (see Table 1). For the sentencing analyses, "Not Guilty" responses were not included.

Verdicts
We conducted a logistic regression with face age (young/middle-aged), cosmetics (with/without), and sex of participant (male/female) entered as the independent variables and verdict as the dependent variable. The logistic regression model was statistically significant, χ 2 (3) = 26.55, p < 0.001. Cosmetics (p = 0.515) and sex of participant (p = 0.572) did not add significantly to the model but age did add significantly to the model (p < 0.001). More specifically, middle-aged defendants were given more guilty verdicts when compared to young defendants (see Figure 1).
Cosmetics 2020, 7, x 3 of 11 not see the same identity twice. Given no previous study has examined the influence of cosmetics on jury decisions, sample size to be collected (220) was predetermined with a sample size calculator [24] using the average effect size from 25 experiments examining the physical attractiveness of defendants on judgments of mock jurors [5]. One participant was excluded because they submitted their responses twice and those responses differed, and another participant was excluded because they entered in their age as 2 years old. Two hundred and eighteen participants (127 male, 89 female, and 2 other), aged 21-69 years old (Mage = 34.99 years, SD = 9.95) completed the study successfully. Each participant was paid $0.80 US dollars upon completion of the study through MTurk.

Study 1 Results
With each trial vignette, participants were asked to give a verdict/sentence out of nine options (ranging from "Not Guilty" to the most severe sentence). For example, in one of the vignettes the options were: "Not Guilty", "Guilty: 1 year", "Guilty: 2 years", "Guilty: 3 years", "Guilty: 4 years", "Guilty: 5 years", "Guilty: 6 years", "Guilty: 7 years", or "Guilty: 8 years". For the verdict analyses, participants' responses were coded as "Not Guilty" (for those that selected "Not Guilty") or "Guilty" (for those that selected any of the eight guilty options). Using those guidelines, 35.8% of the verdicts given by male participants were "Not Guilty" and 64.2% "Guilty" (see Table 1). For female participants, 34.0% of the verdicts given were "Not Guilty" and 66.0% were "Guilty" (see Table 1). For the sentencing analyses, "Not Guilty" responses were not included.

Verdicts
We conducted a logistic regression with face age (young/middle-aged), cosmetics (with/without), and sex of participant (male/female) entered as the independent variables and verdict as the dependent variable. The logistic regression model was statistically significant, χ 2 (3) = 26.55, p < 0.001. Cosmetics (p = 0.515) and sex of participant (p = 0.572) did not add significantly to the model but age did add significantly to the model (p < 0.001). More specifically, middle-aged defendants were given more guilty verdicts when compared to young defendants (see Figure 1).

Sentences
Among the participants who gave guilty verdicts, we conducted repeated-measures ANOVAs for their sentencing decisions. Face age (young/middle-aged) and cosmetics (with/without) were
In order to better understand the three-way interaction between face age, cosmetics, and sex of participant, we re-ran repeated-measures ANOVAs for the sentencing decisions, but this time splitting by participant sex. We found a significant interaction between face age and cosmetics for male participants (F(1, 16) = 4.59, p = 0.048, η2partial = 0.223) but not for female participants (F(1, 21) = 0.16, p = 0.690, η2partial = 0.008). Male participants gave young defendants with cosmetics longer sentences but gave middle-aged defendants with cosmetics shorter sentences, compared to the same faces without cosmetics (see Figure 2).
In order to better understand the three-way interaction between face age, cosmetics, and sex of participant, we re-ran repeated-measures ANOVAs for the sentencing decisions, but this time splitting by participant sex. We found a significant interaction between face age and cosmetics for male participants (F(1, 16) = 4.59, p = 0.048, η2partial = 0.223) but not for female participants (F(1, 21) = 0.16, p = 0.690, η2partial = 0.008). Male participants gave young defendants with cosmetics longer sentences but gave middle-aged defendants with cosmetics shorter sentences, compared to the same faces without cosmetics (see Figure 2).

Study 1 Discussion
We found no significant main effect of cosmetics on the verdicts or sentences given to defendants. We did, however, find that both male and female participants were more likely to assign guilty verdicts to middle-aged defendants than young defendants. We also found that male participants gave young defendants with cosmetics longer sentences than young defendants without cosmetics and middle-aged defendants with cosmetics shorter sentences than middle-aged defendants without cosmetics. This is consistent with the hypothesis that apparent age has a stronger effect on jury decisions than attractiveness.
Our finding that both male and female participants were more likely to assign guilty verdicts to middle-aged defendants than young defendants, regardless of cosmetics use, is in line with research showing that older defendants are judged more harshly than younger defendants [8]. Our finding is

Study 1 Discussion
We found no significant main effect of cosmetics on the verdicts or sentences given to defendants. We did, however, find that both male and female participants were more likely to assign guilty verdicts to middle-aged defendants than young defendants. We also found that male participants gave young defendants with cosmetics longer sentences than young defendants without cosmetics and middle-aged defendants with cosmetics shorter sentences than middle-aged defendants without cosmetics. This is consistent with the hypothesis that apparent age has a stronger effect on jury decisions than attractiveness. Our finding that both male and female participants were more likely to assign guilty verdicts to middle-aged defendants than young defendants, regardless of cosmetics use, is in line with research showing that older defendants are judged more harshly than younger defendants [8]. Our finding is also consistent with Zebrowitz and McDonald's [9] study which found that in small claims court, as defendants increased in "baby-facedness" they were more likely to win cases. One possibility for this effect as presented by Smith and Hed [8] is that jurors might not want to ruin a young defendant's future and may, therefore, be more inclined to give them the "benefit of the doubt". Additionally, it could be that older defendants may be attributed with more agency/responsibility over their actions and are, thus, more likely to be seen as guilty. In order to better understand the effect of defendant age on verdict outcomes, however, further research is needed to explore the motives behind juror decisions, perhaps through follow-up interviews after a verdict has been reached.
Surprisingly, we did not find a main effect of cosmetics on verdicts. One possibility for this null finding may be that including defendants of different ages suppressed any effect that might be present through cosmetics since the effect size of age was quite large (age β = 0.733; cosmetics β = 0.094). Alternatively, we may not have had enough trial vignettes or defendant identities to capture a cosmetics effect. Future work would thus benefit from narrowing the ages of the defendants presented to each participant, increasing the number of trial vignettes, and increasing the number of defendant identities.
Although no main effect of cosmetics on sentencing was found, we did find a significant interaction between face age and cosmetics for the sentencing decisions of male participants. More specifically, we found that male participants gave young defendants longer sentences when presented with cosmetics and middle-aged defendants shorter sentences when presented with cosmetics. One possibility for only finding a significant interaction in male participants is that female participants may have greater insight into the various motivations behind wearing cosmetics (e.g., looking more professional) and could, therefore, be less prejudiced towards their effect. Indeed, research has found that sex of participant significantly influences the effect of cosmetics on judgements [14,25]. For instance, McKeachie [26] found that male participants rated women as more frivolous and less conscientious when wearing cosmetics while Cash et al. [14] found that the ratings of female participants were not affected by the presence or absence of cosmetics.
Initially, we had hypothesized that if apparent age had a stronger effect than attractiveness, young women should receive harsher sentences when presented with cosmetics (since cosmetics make them appear older; Russell, et al., [20]) while middle-aged women should receive more lenient sentences when presented with cosmetics (since cosmetics make them appear younger; Russell, et al., [20]). However, we did not predict that young women with cosmetics would receive harsher sentences than middle-aged women with cosmetics since although young women (i.e., 18-22 years old) do look older with cosmetics, they still look younger than middle-aged women (i.e., 48-52 years old) (Russell; et al., [20]). This suggests that there must be an additional mechanism, other than age perceptions, driving the effect of cosmetics on the sentencing decisions of male participants.
One possibility is that cosmetics may elicit different social associations depending on the age of the wearer. For example, one study conducted in-depth interviews with women aged 71 to 93 and found that in their youth, women used cosmetics as an act of rebellion while later in life, cosmetics were worn in order to look respectable [27]. Some older women even commented that wearing cosmetics was a way for them to signal that they were still engaged in the community [27]. Thus, it appears that the motivations behind wearing cosmetics change with age, which could in turn result in different social associations. Future research thus needs to examine the influence of cosmetics on perceptions across different age groups. If the social associations of cosmetics use change with the age of the wearer, it could explain why in our study, young women were given harsher sentences when wearing cosmetics while middle-aged women were given more lenient sentences when wearing cosmetics.
In Study 1, only a subset of the participants gave sentencing decisions since one of the options was "Not Guilty". We therefore conducted a second study where the participants were told that a jury had found the defendant guilty and were asked to estimate the appropriate length of sentence, Cosmetics 2020, 7, 64 6 of 11 ensuring that all participants gave a sentencing decision. Additionally, Study 1 was conducted using a within-subjects design, which could possibly lead to comparisons between defendants, and we thus conducted Study 2 using a between-subjects design.

Study 2 Materials and Methods
Participants were again recruited through MTurk, ethical approval was received from the Gettysburg College Institutional Review Board, and all participants were US citizens located in the United States who provided consent after being presented with the consent information. Participants were instructed that they would be presented with one trial vignette where they would have to give a sentence, since the defendant had been found guilty.
The case used was selected from the mock trial sample cases provided by the Michigan Center for Civic Education [23] and slightly altered to fit the purpose of this study (e.g., changing the sex of the defendant from male to female, see Appendix B). The trial vignette was presented to participants individually with a picture of the defendant and participants were asked to choose a sentence out of eight options (ranging from "1 year" to "8 years"). As in Study 1, the defendant photograph presented was one of four identities (i.e., two young women (19 and 20 years old) and two middle-aged women (48 and 49 years old) in one of the two conditions (i.e., with or without cosmetics)).
Each participant was individually presented with the same trial vignette and which identity and whether the face appeared with or without cosmetics was randomized. Sample size to be collected (910) was predetermined with a sample size calculator for a between-subjects analysis [24] using the average effect size from 25 experiments examining the physical attractiveness of defendants on judgments of mock jurors [5]. One participant was excluded because they entered in their age as 4 years old. Nine hundred and nine participants (440 male, 466 female, and 3 other), aged 19-76 years old (Mage = 38.55 years, SD = 11.67) completed the study successfully. Each participant was paid $0.50 US dollars upon completion of the study through MTurk.

Study 2 Discussion
We found no significant main effects of cosmetics, face age, or sex of participant on the sentences given to defendants. We also did not find a significant interaction between cosmetics, face age, and sex of participant. It is surprising that neither the main effect of face age nor the three-way interaction (both found in Study 1) were replicated in Study 2. One possibility is that there is simply no effect of cosmetics or age on jury decisions. However, given the strong effect of cosmetics on attractiveness [13][14][15] and the link between attractiveness and jury decisions [5,8,11], this seems unlikely. Additionally, previous studies have already found an effect of age on jury decisions [8,10].
Another possibility for finding the effects in Study 1 but not in Study 2 is that having to sentence four defendants in a row led to comparison effects. For example, when compared to older defendants, younger defendants were given harsher sentences in Study 1 but when presented alone, in Study 2, this was not the case. That presenting all four defendants in a row led to different effects than presenting each defendant on their own also raises additional questions. For instance, could comparisons with other people also affect jury decisions? Say if the defendant's attorney is younger/older than the defendant, would jurors be influenced by such a comparison as they were with the presentation of other defendants in Study 1? Future work is needed to clarify such possible comparison effects on jury decisions.
Alternatively, the trial vignette we used in Study 2 could have elicited different jury decisions than the trial vignettes we used in Study 1. Indeed, age and cosmetics may cause women to look more guilty for certain crimes but not for others. An additional possibility is that the participants might not have had an understanding of what would be a lenient or a harsh sentence and for that reason, tended to choose the middle sentence for all images. Future studies would, therefore, benefit from examining different types of crimes as well as by providing the sentenced term and asking the participants if they would have decided on a harsher or more lenient sentence.

Conclusions
In Study 1, we found that regardless of cosmetics use, male and female participants were more likely to assign guilty verdicts to middle-aged defendants than young defendants. We also found that male participants gave young defendants longer sentences when presented with cosmetics and middle-aged defendants shorter sentences when presented with cosmetics. This finding suggests that the use of cosmetics affects jury decisions. More specifically, wearing cosmetics is beneficial for middle-aged defendants but detrimental for young defendants.
In Study 2, however, we found no main effects of age or cosmetics and no significant interaction between age, sex of participant, and cosmetics on sentencing decisions. One possibility for this lack of replication is that in Study 1, participants saw all four identities whereas in Study 2, participants only saw one identity. Thus, comparison effects might be influencing the results from Study 1.
Across both studies, we found no consistent effect of cosmetics on the decisions of jurors. Nevertheless, our two well-powered studies (N = 1127) provide an initial examination into the influence of cosmetics on jury verdicts and sentences. Future work is still needed to better understand how cosmetics influence defendant perceptions, but our studies provide a meaningful first step in understanding the effect of cosmetics on jury decisions.