When jurors’ moral judgments result in jury nullification: moral outrage at the law as a mediator of euthanasia attitudes on verdicts

Abstract

In a mock-trial study, jurors read evidence about a doctor who had killed a terminally ill patient at the patient’s request. We tested whether instructing jurors about jury nullification (ie jurors’ power to return a not-guilty verdict even when legal guilt is beyond doubt, often because the law would result in unjust convictions) would exacerbate the effect of pre-trial euthanasia attitudes on their verdicts – compared to standard, pattern jury instructions. We also hypothesized that anti-euthanasia pre-trial attitudes would result in moral outrage at the defendant and higher conviction rates, but pro-euthanasia attitudes would prompt feelings of moral outrage at the law and lower conviction rates. Moderated mediation analyses revealed that nullification instructions bolstered the effect of attitudes on verdicts by encouraging jurors to rely on their feelings of moral outrage toward the defendant. Jurors’ moral outrage toward the law mediated the effect of attitudes on verdicts regardless of nullification instructions.

In 1973, 23-year-old Lester Zygmanik stood trial for first-degree murder after shooting his paraplegic brother. George, Lester’s 26-year-old brother, had suffered a motorcycle accident that led to a fractured neck and damaged spinal cord, leaving him paralyzed from the neck down. George begged his brother Lester to kill him, arguing that he would have killed himself if he could. Lester complied and shot his brother in the head with a sawn-off shotgun, then turned himself in and confessed to killing his brother with premeditation. He had even packed the bullets with wax to ensure their effect would be deadly (Cook, 1976; Kassin & Wrightsman, 1988). Despite the indubitable legal guilt, the jury acquitted Lester Zygmanik after only three hours of deliberation. The 12 jurors decided that Lester had been legally insane at the time of the shooting, but they also found that he had regained his sanity and should be set free (Maguire, 1974). In other words, the jurors cleverly circumscribed the letter of the law to reach a not-guilty verdict reflecting their personal sense of justice – a phenomenon known as jury nullification. What determines jurors to nullify the law and to acquit defendants when guilt is established beyond a reasonable doubt?

Jurors’ verdicts in criminal trials often involve issues of morality and justice (eg sexual assault, assisted suicide, hate crimes). When that is the case, jurors are asked to judge a legal transgression – but they also indirectly judge the underlying moral transgression. Therefore, factual judgments of guilt can be colored by judgments about the morality or immorality of the act. For example, people required to make legal judgments of guilt take into consideration (whether they are aware of it or not) their own feelings of anger, disgust and moral outrage at the defendant’s actions (Bastian, Denson, & Haslam, 2013; Salerno & Peter-Hagene, 2013). Such feelings of moral outrage in reaction to the crime lead to increased certainty in the defendant’s guilt and a higher likelihood of conviction (eg Wiley & Bottoms, 2009), even though the outrageousness of the act has no bearing on the likelihood that the defendant committed the act (ie performed the illegal actions for which he or she stands trial). Thus, the perception of heightened moral blame, rather than the evidence of legal guilt, can drive verdicts in cases that rouse outrage.

But the opposite can also happen: when jurors do not find the defendant morally blameworthy, they might not convict even in the face of certain legal guilt. Trials can end in jury nullification, meaning that jurors choose to acquit because applying the law to the letter would violate their sense of justice, even though legal guilt is proven beyond a reasonable doubt (Finkel, 2000; Horowitz, Kerr, Park, & Gockel, 2006; Kerr, 2010; Kerr et al., 2008; Scheflin & Van Dyke, 1980). Nullification has been particularly well illustrated in high-profile euthanasia (ie mercy killing) and physician-assisted suicide cases such as the trial of Lester Zygmanik and the repeated trials of Jack Kevorkian (Johnson, 1999; Michigan v Kevorkian). Kerr (2010) delineates several reasons why juries nullify, such as the belief that a defendant’s actions were not immoral and the belief that the law itself is unjust and should be rejected (eg Northern United States juries refused to convict those legally guilty of breaking the Fugitive Slave Act of 1850 by helping slaves escape).

We reasoned that jurors’ personal attitudes toward a crime that carries with it some moral ambiguity (ie euthanasia) would influence whether they experience moral outrage at the defendant or moral outrage at the unjust law, and in turn, whether these two forms of moral outrage would motivate jurors to convict or to acquit. We also reasoned that the indirect effect of attitudes on verdicts through moral outrage would be stronger when jurors are given explicit instructions about nullification, allowing them to follow their own moral compass instead of the letter of the law.

Euthanasia attitudes

The legal system is ambivalent about the desirability of attitudes influencing verdicts. In Witherspoon v Illinois, the court ruled that attitudes toward the death penalty express the ‘conscience of the community’, and thus jurors who opposed the death penalty should not be universally excluded from the jury. Yet, other court decisions limit the allowed effect of attitudes such that they must constitute a ‘reasoned moral response’ to the evidence (Penry v Lynaugh), blocking jurors’ license to apply their moral standards regardless of the law and the evidence. The extent to which jurors understand these nuanced legal limits of relying on attitudes is unclear, but psychological research indicates jurors do rely on their pre-trial attitudes.

Because euthanasia involves the intentional killing of another person (albeit with the intent to end that person’s suffering, and often at the victim’s own request), it has historically been considered murder from a legal perspective (eg Barber v Superior Court; Gilbert v State; People v Kevorkian). Public opinion, however, has been split on its moral merits (Rosenfeld, 2000), and public support for euthanasia and physician-assisted suicide as revealed by Gallup polls has increased over the years, with 64% of those polled supporting euthanasia in 2012 (Emanuel, Onwuteaka-Philipsen, Urwin, & Cohen, 2016) and 72% supporting the practice in 2018 (Brenan, 2018). Laws and policies reflect this shift: physician-assisted suicide (ie physicians purposefully providing terminally ill or severely impaired patients with the means to take their own lives) has been legalized in Washington DC and six states in the United States (Oregon, Vermont, Washington, Montana, Colorado, Hawaii and California), as well as in countries such as Belgium, Canada, Colombia, Germany, Luxembourg, The Netherlands and Switzerland. Euthanasia, which is a more direct form of mercy killing in which the perpetrator (whether a medical professional or family member) not only provides the means, but also completes the action of killing the patient, is illegal in all states. Further, in most states, and in most countries worldwide, both forms of mercy killing are illegal and foster significant moral opposition (Bulmer, Böhnke, & Lewis, 2017).

How might this public split on the morality of euthanasia affect the legal system’s ability to prosecute and convict such cases? On most issues, the impact of pre-trial attitudes in the courtroom is robust: attitudes influence verdicts in cases involving the insanity defense (Rendell, Huss, & Jensen, 2010), death penalty (O’Neil, Patry, & Penrod, 2004), rape (Field, 1978; Kovera, 2002) and child sexual abuse (Bottoms, 1993; Bottoms et al., 2014). Giner-Sorolla, Chaiken, and Lutz (2002) found that attitudes that reflect moral/ideological beliefs – such as attitudes about the (im)morality of euthanasia – hold even stronger and more consistent sway on jurors’ verdicts compared to other types of attitudes (eg validity beliefs). The effect of pre-trial attitudes is explained by several juror decision-making models, including the story model (Pennington & Hastie, 1986) and the ‘director’s cut’ model (Devine, 2012). The latter, in particular, details how jurors arrive at a verdict via a process similar to constructing, then editing, a story. Jurors form initial representations of the trial based on their existing schemas (scripts, stereotypes) about the crime or about the defendant. These schemas interact with actual trial characteristics (ie evidence) to form stories about the trial, which are then evaluated for plausibility (similar to the story model). Thus, jurors’ pre-existing attitudes toward euthanasia, for example, would interact with features of the trial (eg testimony about a defendant’s motivation, jury instructions) to shape the most convincing story, which then guides jurors’ verdict decisions. The value of this approach is that it allows for specific, quantifiable predictions about how the exact same trial information – be it evidence or instructions – can shape jurors’ verdicts differently as a function of jurors’ pre-existing beliefs.

Two studies tested the effect of euthanasia attitudes, specifically, on jurors’ verdicts in cases involving (a) a man accused of killing his suffering spouse (Meissner, Brigham, & Pfeifer, 2003) and (b) a doctor accused of killing a terminally ill patient (Peter-Hagene & Bottoms, 2017). In both studies, pro-euthanasia jurors were more likely to acquit than anti-euthanasia jurors, despite the fact that the defendants’ legal guilt was established beyond reasonable doubt (although this effect was reduced by deliberation, Meissner et al., 2003). Finkel, Hurabiell, and Hughes (1993) also provided indirect evidence that attitudes are very important to jurors’ verdicts in euthanasia cases: jurors who acquitted cited pro-euthanasia beliefs (ie the killing was not an immoral act) as the main reason for their law-nullifying acquittals in several euthanasia cases.

Moral outrage as mediator of euthanasia attitudes on verdicts

What psychological processes underlie the strong effect of attitudes on legal judgments, especially when attitudes are at odds with black-letter law? According to Finkel et al. (1993), jurors who acquitted in a euthanasia case reasoned that the defendant had not acted with malice and that the law should be changed or disregarded. These judgments are clearly not evaluations of case evidence strength; instead, they indicate that jurors grounded their verdicts in moral evaluations of the defendant, but also of the law. Such judgments of moral blame and/or unfairness are important aspects of moral outrage, a constellation of affective, cognitive and behavioral responses to perceived moral transgressions (Skitka, Bauman, & Mullen, 2004). The emotions associated with moral outrage are anger and disgust – moral emotions commonly experienced in response to immoral acts (Batson et al., 2007; O’Mara, Jackson, Batson, & Gaertner, 2011; Okimoto & Brescoll, 2010; Salerno & Peter-Hagene, 2013). Cognitive appraisals associated with moral outrage include perceived unfairness (Antonetti & Maklan, 2016), perceived questionable motives (Crossley, 2009; Grégoire, Laufer, & Tripp, 2010), attributions of blame (Darley & Pittman, 2003) and perceived harm severity (Bastian et al., 2013; Salerno et al., 2010). Moral outrage is also a psychological response that elicits action: if an injustice has been committed, justice must be restored (Darley & Pittman, 2003). Behavioral response tendencies therefore include harsh punishment recommendations or support for punitive policies (Bastian et al., 2013; Salerno et al., 2010), but also an increased likelihood of attributing factual guilt to defendants (Salerno & Peter-Hagene, 2013; Wiley & Bottoms, 2009).

When moral outrage is absent, however, people are less prone to punish law-breakers such as vigilantes (Haas, Keijser, & Bruinsma, 2012). In one study specifically about jurors’ moral judgments in a euthanasia case, Peter-Hagene and Bottoms (2017) found that moral outrage toward the defendant mediated the effect of pre-existing attitudes on verdicts: pro-euthanasia jurors experienced little or no moral outrage toward the defendant, and in turn were more likely to acquit than anti-euthanasia jurors, who experienced higher levels of moral outrage. Moral outrage toward the defendant was an immediate predictor of jurors’ verdicts, but was grounded in jurors’ pre-existing beliefs about euthanasia as morally acceptable or not.

But to our knowledge, although both moral outrage and perceptions of the legal system have been studied extensively, no research has tested whether jurors experience feelings akin to moral outrage toward the law when their attitudes contradict the letter of the law. There are several reasons why we hypothesized such a psychological process takes place. People can feel moral outrage toward institutions or even political regimes when these behave unfairly (Lindenmeier, Schleer, & Pricl, 2012; Pagano & Huo, 2007), which is consequential, especially given that unfairness is a primary predictor of moral outrage (Antonetti & Maklan, 2016). As mentioned above, Finkel et al. (1993) noted that jurors who acquitted in the euthanasia trial also commented that the law should be changed or disregarded. This observation is consistent with research demonstrating that moral attitudes can affect not only verdicts and reactions to defendants or victims, but also people’s reactions to the legal system itself, including evaluations of legal institutions. For example, people who hold strong moral convictions about a legal outcome are more likely to criticize legal institutions when that outcome is not achieved (Skitka & Mullen, 2002) and to react with anger – a principal emotional component of moral outrage – when outcomes are not aligned with their moral beliefs (Mullen & Skitka, 2006).

In a study directly relevant to our legal issue – euthanasia committed by a doctor at a patient’s request – Skitka, Bauman, and Lytle (2009) measured people’s moral convictions about physician-assisted suicide before and after a Supreme Court ruling on the right of states to legalize it. People’s attitudes before the ruling significantly predicted not only their perception of the fairness of that ruling, but also their perceptions of the Supreme Court’s legitimacy. When the decision was in line with people’s pre-existing attitudes, they perceived it as fair and the Court as a legitimate source of such decisions; but when the decision contradicted their pre-ruling attitudes, people perceived it as unfair and rated the Court as less legitimate. These findings suggest that, not only are jurors capable of bypassing the law when it contradicts their moral beliefs (ie the authority-independence hypothesis), they are also likely to judge the law itself as immoral and to question its legitimacy. We are extending this research by testing the hypothesis that pre-existing attitudes about a crime can influence verdicts because they elicit a visceral, emotional reaction of moral outrage at an unfair law – and not just cognitive evaluations about legitimacy.

We reasoned that, when pro-euthanasia jurors are faced with a trial pitting their moral sense of right and wrong against the law, they might experience moral outrage toward the law and a desire to correct the injustice done to the defendant by the prosecution. These emotional and cognitive reactions might contribute to jurors’ propensity to find the defendant not guilty. Our measures were derived from literature on the cognitive appraisals and complex emotions associated with anger and moral outrage (eg Salerno & Peter-Hagene, 2013): desire to correct an injustice and belief that one has the power to right a wrong (Lerner & Tiedens, 2006). For pro-euthanasia jurors in particular, we expected (a) appraisals of the law as unjust, (b) desire to correct the injustice done to the defendant and (c) belief that they can correct the injustice by acquitting. In other words, we reasoned that the target of typical outrage appraisals might be the law/prosecution, not the defendant, for jurors whose attitudes make them prone to take the defendant’s side. We tested the possibility that these emotional and cognitive reactions mediate the effect of attitudes on verdicts. We also replicated the moderating role of nullification instructions – instructions that explicitly inform jurors they are allowed to rely on their own conscience in reaching a verdict, regardless of the law (Peter-Hagene & Bottoms, 2017).

Should jurors receive nullification instructions?

Jury nullification is a controversial issue among legal scholars and legal professionals. Some legal scholars support the jury’s nullification power, reasoning that the legal system should always reflect the community’s morality (Rubenstein, 2006; Scheflin & Van Dyke, 1980) and that, when laws are perceived as unjust, jurors (as representatives of the community) should have the power to bypass them. From this perspective, instructing jurors about nullification is not problematic – it merely informs them about an existing power that is already given to them by the legal system. Yet, others find nullification problematic because it allows jurors to act in potentially irrational, irresponsible ways and to follow their incorrect intuitions or personal prejudices (Finkel, 2000). Furthermore, if jurors disregard the law, the legal system has no recourse against their decisions, because jury acquittals are not reversible.

Due to these concerns about jurors’ inability to place the law above their personal attitudes, courts have historically ruled that jurors need not be explicitly informed about nullification (Sparf v the United States; United States v Dougherty). Thus, although courts have to abide by jurors’ nullification decisions when these occur naturally, they do not encourage jury nullification by providing explicit instructions about it. In some jurisdictions, attorneys are allowed to mention nullification power during their opening and closing arguments, but only at the discretion of the presiding judge (People v Weinberg). State legislatures recently enacted initiatives in New Hampshire, Oregon, New York and Massachusetts to mandate the inclusion of nullification instructions delivered by the judge with unequivocal, strong language (National Center for State Courts, 2015; Volokh, 2016). Although these efforts have been historically unsuccessful (ie they did not result in nullification language being included in jury instructions), they illustrate society’s ongoing concern with the tension between biding jurors to the letter of the law or allowing them to rely on personal morality.

It is important, as this debate over nullification instructions continues, to study their effect in a systematic, empirical manner. Prior research has shown that telling jurors explicitly about nullification increases their tendency to acquit in morally ambiguous cases (eg euthanasia, Horowitz, 1985; 1988; Horowitz et al., 2006), but not in cases where the defendant had clear moral blame (eg driving drunk, Horowitz, 1988). Two studies have directly tested the hypothesis that nullification instructions exacerbate the effect of jurors’ pre-existing attitudes on verdicts by encouraging them to follow their conscience. Meissner et al. (2003) did not find a stronger relation between jurors’ euthanasia attitudes and their verdicts when jurors were informed of their ability to nullify the law – attitudes had a significant effect on verdicts regardless of the type of instructions. Furthermore, after jurors deliberated on the case, the effect of pre-trial attitudes was reduced only for jurors who heard nullification instructions. Peter-Hagene and Bottoms (2017), however, found that nullification instructions did exacerbate jurors’ reliance on their euthanasia attitudes when jurors were in a neutral emotional state (but not when they had been induced to feel anger). Given the conflicting results in existing literature, we wanted to test whether the moderating effect of nullification instructions on the relationship between jurors’ pre-existing attitudes and their trial verdicts would replicate, especially when studied in a more generalizable community sample (the prior studies were conducted with undergraduate samples). Further, we wanted to provide direct evidence that nullification instructions affect verdicts and moderate the effect of attitudes because jurors’ psychological responses to these instructions are feelings of increased freedom to rely on their conscience and decreased obligation to follow the law.

Study hypotheses and overview

Mock jurors participated online for monetary compensation. First, they reported their attitudes toward euthanasia, then they read a trial summary involving a hospital physician accused of killing his terminally ill patient at the patient’s request. The case facts left no reasonable doubt that the physician-administered drugs caused the patient’s death. Then, jurors received either standard or nullification instructions from the judge and the defendant’s lawyer and completed the study measures (ie verdict, moral outrage at the defendant, moral outrage at the law, manipulation checks and demographics).

First, we predicted a significant main effect of euthanasia attitudes: jurors who support euthanasia would be more likely to acquit than jurors who oppose it. Second, we predicted an interaction effect between instruction type and jurors’ euthanasia attitudes, such as that jurors who receive nullification (versus standard) instructions would be more likely to rely on their euthanasia attitudes. Jurors who support (versus oppose) euthanasia would be less likely to convict the doctor in the nullification, but not standard, experimental condition.

Third, we hypothesized that effect of euthanasia attitudes on verdicts would be mediated by moral outrage toward the defendant and moral outrage toward the law. Specifically, jurors who support (versus oppose) euthanasia would feel less moral outrage toward the defendant, but more moral outrage toward the law (a paths); in turn, moral outrage toward the defendant would result in higher likelihood of conviction, whereas moral outrage toward the law would result in higher likelihood of acquittal (b paths). Fourth, we predicted that nullification instructions encourage jurors to rely on their moral outrage reactions as legitimate sources of information to use in their decision-making process; standard jury instructions, in contrast, restrict jurors’ decision to case facts and the letter of the law. Thus, we hypothesized a moderated mediation effect where instructions moderate the b-paths (ie the effect of moral outrage toward the defendant and toward the law on verdicts). Specifically, the conditional indirect effect of euthanasia attitudes on verdicts through moral outrage would be significant in the nullification instruction condition, but not significant in the standard instructions condition.

Finally, we predicted that jurors’ measured perceived obligation to follow the law would perform exactly the same as the experimental manipulation of jury instructions, providing a conceptual replication and evidence of the psychological mechanism underlying the moderating effects of the instructions. Specifically, the effect of attitudes on verdicts would be stronger for jurors with high (versus low) perceived obligation to follow the law.

Method

Participants

Participants were 178 jury-eligible community members (ie older than 18 years of age and United States citizens) who participated online via Mechanical Turk. Mechanical Turk samples are increasingly common in the fields of social psychology and psychology and law. This method provides a national sample that is more like a jury venire than college students. Research suggests that this methodology can be reliable (eg Casler, Bickel, & Hackett, 2013; Rouse, 2015) and that participants take their tasks seriously (Mason & Suri, 2012). Twenty-one additional participants were dropped from analyses (19 claimed they had not received nullification instructions although they had, and 2 did not complete the dependent measures). The sample was 19–75 years old (M = 34.75, SD = 10.79); 55% women; 9% Asian, 6% Black, 8% Hispanic, 78% White and 3% of other or multiple ethnicities; and 7% Fundamentalist Christian, 16% Protestant, 20% Catholic, 1% Jewish, 4% Hindu, 2% Buddhist, 15% agnostic, 28% atheist and 7% other or multiple religions. Power analyses indicated that a sample size of n = 71 per experimental cell was sufficient to detect medium effects (R² = .15) with 80% power for the focal predictors, as well as for the simple slopes of euthanasia attitudes at each level of the experimental instructions manipulation.

Materials and measures

Attitudes toward euthanasia

Euthanasia attitudes were measured with six items of the Attitudes toward Euthanasia questionnaire (Wasserman, Clair, & Ritchey, 2005), answered on a 7-point scale ranging from Strongly disagree to Strongly agree, with a midpoint indicating Neutral (responses were scored from 1 to 7; see Table 1). Item examples include: ‘If a patient in severe pain requests it, a doctor should prescribe that patient enough medicine to end their life’; ‘Even if a doctor does not think that a patient will recover, it would be wrong for the doctor to end the life of a patient’. Roughly equal numbers of jurors endorsed pro-euthanasia (scores above 4, N = 89) and anti-euthanasia (scores below 4, N = 82) attitudes (the rest were neutral, N = 8). The scale was reliable in the original (Wasserman et al., 2005) and present samples, Cronbach’s αs = .91 and .82, respectively.

Table 1.

Scale descriptives and correlations for continuous measures.

						Correlations
	Mean	Mean 95% CI	SD	Skew	Kurtosis	1.	2.	3.
1. Euthanasia attitudes	3.96	[3.71, 4.20]	1.69	−.15	−.71	–	–	–
2. Moral outrage	1.36	[1.27, 1.46]	.67	2.17	4.72	−.36**	–	–
3. Moral outrage at the law	2.91	[2.71, 3.11]	1.34	−.04	−1.27	.55**	−.36**	–
4. Obligation to follow the law	3.15	[2.95, 3.34]	1.35	.14	−1.25	–.19**	.02	−.13

Trial stimulus

The written summary trial transcript was a modified version of the euthanasia case used by Horowitz and colleagues (2006) and a shortened 2-page version of the transcript used by Peter-Hagene and Bottoms (2017). The case involved a hospital physician, Dr Wood, accused of the murder of Henry Bates, a terminally ill cancer patient. Dr Wood performed euthanasia by administering drugs that were lethal to Henry Bates. Hospital staff testified that Dr Wood was upset by the patient’s suffering. The victim’s wife testified that she and Henry Bates both asked Dr Wood to end Henry’s suffering. Dr Wood confessed to administering the lethal drug. The trial stimuli were designed to (a) leave no room for reasonable doubt that Dr Wood intended to kill Henry Bates and that he was successful in doing so (ensuring that a not guilty verdict reflected true nullification) and (b) convince jurors that Dr Wood was motivated by his desire to end Bates’ suffering. Participants in the nullification instruction condition were informed of their power to nullify by the defense attorney in his closing statements (eg ‘Your feelings and beliefs in this case reflect the feelings and beliefs of the community you represent, and you have the power to find the defendant not guilty if you don’t believe he acted immorally’; ‘Our laws are clear: Nobody can deny jurors their right to vote according to their conscience, no matter what the facts of the case are, and no matter the evidence’). Our trial summary was based on an actual case and included appropriate charges, legally admissible evidence and actual pattern jury instructions. The transcript was reviewed by a trial attorney for legal realism.

Jury instructions

We used Illinois jury pattern instructions for first-degree murder containing elements of the crime of murder, admonitions to focus only on the evidence and reminders of burden of proof and reasonable doubt criteria (illinoiscourts.gov, Pattern Criminal Jury Instructions). As the nullification-instruction manipulation, half of the jurors also read the following paragraph developed by a legal scholar (Van Dyke, 1970) and used by several researchers in their prior investigations (Horowitz et al., 2006; Meissner et al., 2003):

While you must give respectful attention to the laws about which you have just been instructed, you have the final authority to decide whether or not to apply a given law to the acts of the defendant on trial. As jurors you represent the community and it is appropriate to bring into your deliberation the feelings of the community and your own feelings based on your conscience. You must respect the law, that is clear. However, regardless of your respect for the law nothing should stop you from acquitting the defendant if you feel that the law, as applied to the facts and situation in this case, would lead to an injustice.

Verdicts

Participants indicated whether they found the defendant Guilty or Not guilty.

Perceptions of obligation to obey the law

Participants rated, on scales from 1 (Not at all) to 5 (Very much), how free they felt to follow their conscience and how obligated they felt to follow the law about first-degree murder. The first item was reverse-coded, and the two items were combined into a single scale, with higher scores indicating increased feelings of obligation to follow the law (α = .70, Table 1).

Moral outrage scale

The 4-item scale (Skitka et al., 2004) assessed the emotional reactions (‘I feel morally outraged by the defendant’s actions’); action tendencies (ie desire to punish, ‘I feel the need to punish the defendant’; ‘I feel a desire to hurt the defendant’) and beliefs (‘I believe the defendant is evil to the core’) that form the construct of moral outrage. Item responses ranged from 1 (Not at all) to 5 (Very much). The scale had good reliability in the initial, α = .80, and our sample, α = .75 (Table 1).

Moral outrage at the law scale

Jurors’ moral outrage at the law and their desire to correct the injustice done to Dr Wood by the prosecution were measured with the following questions answered on scales from 1 (Not at all) to 5 (Very much): ‘I am morally outraged that the law makes me vote GUILTY in this case’; ‘I felt the need to correct the injustice done to Dr Wood by the prosecution’; ‘I believe a NOT GUILTY verdict would right the wrong done to Dr Wood by the unjust prosecution’. These were combined into one scale, α = .80 (see Table 1).

Manipulation checks

To ensure that participants noticed the nullification instruction manipulation, after completing the dependent measures, they answered the following question with yes or no: ‘Were you told verbally or in anything you read something like: You did not have to follow the law about first degree murder if you thought it was not right?’.

Demographic characteristics

To determine eligibility, participants indicated their age and citizenship status. They also indicated their gender, religious affiliation, political orientation and whether anyone close to them had suffered from painful terminal illness.

Procedure

Mechanical Turk workers received $2 to complete the 30-minute study online via a Qualtrics link. To participate, all workers had to have 30 minutes available to complete the study in one sitting, had to be older than 18 years and United States citizens, and could not sign up for the study twice (people took between 18 and 50 minutes to finish, M = 29 min). Participants were randomly assigned to the jury instructions conditions. First, they completed the euthanasia attitudes scale. We measured the attitudes before the trial for several reasons. First, we wanted to ensure that we could test the effect of these attitudes on verdicts, while also making sure we assessed verdicts immediately after the trial with no intervening measures. Second, like Meissner et al. (2003), we reasoned that measuring attitudes first ensured that case-specific factors would not influence people’s responses on the attitude scale (eg polarization). Third, we reasoned that asking jurors about their case-relevant attitudes before the trial was an ecologically valid way to mirror what happens in actual trials during voir dire. That is, in a trial such as this, jurors would likely answer questions about their general attitudes regarding physician-assisted suicide and euthanasia during voir dire – before they are exposed to the case evidence (eg Crocker & Kovera, 2010; Kovera, Dickinson, & Cutler, 2003). Third, we did not have the option of conducting the study in two separate sessions – that is, of collecting the attitude data weeks before the actual case. Given this methodological limitation, we reasoned that measuring attitudes first was more ecologically valid and also allowed us to test theoretical hypotheses about the effect of pre-trial attitudes on verdicts and moral outrage.

Then, they read the trial summary with each page displayed to allow ample reading time. Participants gave their verdicts and completed measures of moral outrage toward the law, moral outrage toward the defendant, manipulation checks and demographic characteristics. Although measuring the theoretical mediators – moral outrage toward the defendant and moral outrage toward the law – after the dependent variable might draw into question the causal chain of cause and effect, we reasoned this option was less problematic than the alternative: asking jurors about their moral outrage before they gave their verdicts. Specifically, as others have done before (eg Jay, Salerno, & Ross, 2018; Levett & Kovera, 2009), we prioritized measuring the primary legal outcome (ie verdicts) without drawing jurors’ attention to the intervening mechanisms in an artificial way. Had we asked jurors about their moral outrage at the defendant and at the law, we might have artificially encouraged them to take these feelings into account when deciding on their verdict. Instead, we arguably captured the relationship between the psychological mechanisms and the actual decision measure in a more conservative manner.

At the end of the study, participants were debriefed in keeping with IRB-approved methods and given a completion code necessary for the $2 compensation.

Results

Preliminary analyses

Preliminary analyses revealed that women (60%) were more likely than men (41%) to vote guilty, χ² (N = 178) = 6.33, p = .01, and were less pro-euthanasia (M = 3.68) than were men (M = 4.29), t(176) = 2.43, p = .02. Jurors who convicted were also more conservative (M = 1.88) than jurors who acquitted (M = 1.37), t(176) = −4.98, p < .001, and the more conservative the jurors, the less pro-euthanasia they were, r = −.48, p < .001. No other background factors were significantly associated with verdicts or euthanasia attitudes. Therefore, we controlled for gender and political orientation in further analyses, but the patterns of results were the same when controls were not included.

Although we measured euthanasia attitudes before the manipulation of jury instructions, we wanted to make sure that our focal predictor did not differ significantly between the two conditions by accident. Further, given our hypothesis that nullification instructions would moderate the b-path in our mediation model, the relationship between our proposed mediators (moral outrage toward the defendant and toward the law) and jurors’ verdicts, we also wanted to make sure these mediators were not themselves affected by the manipulation and were not significantly different by experimental condition. We performed independent samples t-tests to test the effect of jury instructions on (a) euthanasia attitudes, t(177) = .67, p = .50; (b) moral outrage toward the defendant, t(176) = .01, p = .99; and (c) moral outrage toward the law, t(176) = .46, p = .64. All tests were not significant (see Table 2 for means and standard deviations).

Table 2.

Scale descriptives and % guilty verdicts per experimental condition.

	Standard Instructions N = 105		Nullification instructions N = 73
	M [95% CI]	SD	M [95% CI]	SD
Euthanasia attitudes	4.02 [3.69, 4.36]	1.75	3.85 [3.48, 4.23]	1.61
Moral outrage	1.36 [1.23, 1.49]	.65	1.36 [1.20, 1.53]	.72
Moral outrage at law	2.96 [2.69, 3.22]	1.38	2.86 [2.56, 3.16]	1.29
Obligation to follow the law	3.53 [3.26, 3.81]	1.42	2.59 [2.36, 2.82]	.97
	% Guilty		% Guilty
Verdict	64%	–	34%	–

Finally, we included an additional test to ensure that the nullification instructions were not only understood at a semantic level (ie participants remembering what they were told for the manipulation check), but that they had the intended psychological effect of ‘freeing’ jurors from the perceived obligation to follow the law for first-degree murder. An independent-samples t-test revealed that jurors felt significantly less bound to follow the law in the nullification, compared to standard instructions condition, t(176) = 4.93, p < .001 (Table 2).

Euthanasia attitudes and jury instructions effects on verdicts

Overall, although 89% of participants believed legal guilt had been established beyond reasonable doubt, 48% of jurors acquitted the defendant, reflecting true nullification: 36% in the standard and 66% in the nullification instructions condition. Results were similar from analyses with and without the 11 participants who did not believe the reasonable bar standard had been met, so results reported herein include the full sample. Of note, the trial stimulus resulted in an approximately even split in verdicts, which allowed us to test our hypotheses unencumbered by ceiling or floor effects.

We used hierarchical logistic regression to test our first two hypotheses about the main effect of euthanasia attitudes and the moderating effect of jury instructions. We found support for our first hypothesis that jurors who support euthanasia would be less likely to convict than jurors who oppose it: analyses revealed negative and significant main effects of attitudes and nullification instructions on dichotomous verdicts in Step 1, indicating that pro- (versus anti-) euthanasia jurors and jurors who heard nullification (versus standard) instructions were less likely to convict and endorsed a lower degree of guilt. Overall, the Step 1 model was useful in predicting verdicts, Nagelkerke R² = .42. In Step 2, which tested the moderation hypothesis, the two-way interaction between nullification instructions and attitudes was significant, and adding the interaction terms improved the overall model for verdicts, Nagelkerke R² = .48 (all individual predictor statistics are found in Table 3). Follow-up analyses of simple slopes for the two-way interaction between jury instructions and attitudes revealed that, when mock jurors received standard instructions, their attitudes had no significant effect on verdicts (B = −.37, Wald = 2.83, p = .10, OR = .69, 95% CI[.45, 1.06]). In contrast, when jurors received nullification instructions, attitudes significantly predicted verdicts (B = −1.44, Wald = 13.79, p < .001, OR = .24, 95% CI[.11, .51]) (Figure 1). Again, this pattern of results was in line with our second hypothesis.

Table 3.

Hierarchical logistic regression results for moderating effects of attitudes and instructions on dichotomous verdicts.

	B	SE	OR	OR 95% CI	Wald	p
Step 1
Instructions	−1.71**	.40	.18	[.09, .39]	18.32	<.001
Attitudes	−.55**	.14	.58	[.39, .64]	15.30	<.001
Step 2
Instructions	−1.97**	.47	.11	[.03, .34]	10.53	<.001
Attitudes	−.27	.16	.69	[.40, .90]	2.83	.088
Instructions * attitudes	1.07**	.38	.35	[.17, .75]	7.40	.005

Notes. Standardized odds ratios (OR) represent OR when standardized euthanasia attitudes scores are used. *p < .05, **p < .01.

Figure 1.

Simple slopes for euthanasia attitudes effects on dichotomous verdicts for each manipulated jury instructions condition (standard, nullification). *p < .05. The graph illustrates that the effect of euthanasia attitudes on guilty verdicts was significant for jurors who read nullification instructions, with pro-euthanasia jurors less likely to convict compared to anti-euthanasia jurors. This difference was not significant for jurors who read standard jury instructions.

Moral outrage at the defendant and moral outrage at the law

Because the predicted effect of attitudes and the interactive effect with jury instruction type were significant, we tested a more complex model that incorporated our proposed mediators: moral outrage toward the defendant and moral outrage toward the law. In all analyses, the same control variables described above were included – and the pattern of results was the same with and without these controls. To test our third hypothesis (moral outrage toward the defendant and toward the law would mediate the effect of attitudes on verdicts), we first ran a mediation model using Hayes’ PROCESS Macro Model 4 (Hayes, 2013) that included the jury instructions variable, but not its interaction with the two proposed mediators (moral outrage at the defendant and moral outrage at the law). The path coefficients are reported in Figure 2. In this preliminary model, the indirect effect of attitudes on verdicts through moral outrage at the defendant was not significant, B = −.04, S.E. = .07, 95% CI = [−.24, .05], but the indirect effect through moral outrage at the law was significant, B = −.12, S.E. = .07, 95% CI = [−.27, −.01].

Figure 2.

Mediation results with euthanasia attitudes as independent variable, moral outrage and reactions to the law scale (i.e. moral outrage toward the law, desire to correct the injustice of prosecuting the defendant, belief the injustice can be corrected by a not guilty verdict) as mediators, and dichotomous verdicts as dependent variable. Numbers represent non-standardized regression coefficients. *p <.05, **p <.01.

Next, we tested our fourth, predicted moderated mediation hypothesis that jury instructions would moderate the indirect effects of euthanasia attitudes through the two mediators. Specifically, we tested whether the conditional indirect effects would be significant when jurors receive nullification instructions (because these instructions would allow them to rely on their own reactions of moral outrage toward the defendant or toward the law, if the law is perceived as unjust) but not when jurors receive standard jury instructions. We again used Hayes’ SPSS Process Macro, this time testing Model 14, in which the proposed moderator moderates the b-paths (ie the path from mediators to the dependent variable). Overall, the model was successful in predicting jurors’ verdicts, Nagelkerke R² = .50. The direct effect of attitudes on verdicts was significant, B = −.43, S.E. = .16, 95% CI = [−.74, −.12], p = .006. The interaction between jury instructions and the indirect effect of attitudes through moral outrage toward the defendant was also significant, B = 1.91, S.E. = .68, 95% CI = [.58, 3.24], p = .005. As predicted, the conditional indirect effect of attitudes on verdicts via moral outrage was not significant in the standard instructions condition, B = .08, S.E. = .09, 95% CI = [−.06, .24], but significant in the nullification instructions condition B = −14, S.E. = .25, 95% CI = [−.74, −.03]. The interaction between jury instructions and the indirect effect of attitudes through moral outrage toward the law was not significant, however, B = .37, S.E. = .34, 95% CI = [−.30, 1.04], p = .283; moral outrage toward the law mediated the effect of attitudes on verdicts regardless of jury instructions. All path coefficients are represented in Figure 3.

Figure 3.

Moderated mediation results with euthanasia attitudes as independent variable, moral outrage and reactions to the law scale as mediators, jury instructions as moderator of the b-path, and dichotomous verdicts as dependent variable. Numbers represent non-standardized regression coefficients. *p <.05, **p <.01.

Perceived obligation to follow the law

Finally, we tested our fifth hypothesis to establish that jurors’ subjective, perceived obligation to follow the law would mirror the moderating effect of the experimental manipulation of jury instructions. As a conceptual replication of the instructions effect, we conducted the same logistic regression analyses, but instead of the instructions manipulation, we introduced the centered continuous measure of jurors’ perceived obligation to follow the law. The main effect of this variable was significant, indicating that jurors who felt more obligated to follow the law were more likely to convict than jurors who felt less so (B = 1.89, Wald = 42.39, p < .001, OR = 6.64, 95% CI [3.76, 11.75]). Overall, the Step 1 model was useful in predicting verdicts, Nagelkerke R² = .72. In Step 2, the two-way interaction between obligation to follow the law and attitudes was significant (B = .71, Wald = 9.56, p = .002, OR = 2.03, 95% CI [1.30, 3.18]), and adding the interaction terms improved the overall model for verdicts, Nagelkerke R² = .77. Simple slopes analyses revealed that, when mock jurors felt less bound by the law, their attitudes had a significant effect on verdicts (B = −1.47, Wald = 9.19, p = .002, OR = .23, 95% CI [.09, .60]). In contrast, when jurors felt more bound to follow the law, the effect of attitudes was not significant (B = −.18, Wald = .95, p = .331, OR = .83, 95% CI [.58, 1.20]).

Overall, jurors’ reactions to the law (ie their moral outrage at the law, perceptions of the law as unjust and desire to correct the injustice toward the defendant) mediated the effect of attitudes on verdicts regardless of nullification instructions. In contrast, jurors’ moral outrage toward the defendant only mediated the effect of euthanasia attitudes on verdicts when jurors were explicitly encouraged to rely on their conscience by the nullification instructions.

Discussion

Euthanasia attitudes, nullification instructions and verdicts

Our findings are only partially consistent with the literature on moral convictions, which would predict that moral attitudes guide people’s judgments regardless of legal standards and instructions. According to Skitka et al.’s (2009) authority independence hypothesis, when people hold strong moral beliefs about something (ie hold a ‘moral stake’ in a particular outcome), their perceptions of legal outcomes are shaped primarily by these moral beliefs and are somewhat independent of the legitimacy of, and desire to comply with, a legal authority. On the one hand, mock jurors could not easily divorce legal judgments (ie guilt) from the inherent moral judgments (ie blame): although nearly all mock jurors agreed that the prosecution had proven guilt beyond reasonable doubt, roughly half still chose to acquit the defendant, thereby nullifying the law. Of note, even of those jurors who did not learn about nullification (ie read standard jury instructions), a third still chose to acquit the defendant. Hence, evidence of legal guilt was clearly not the only factor shaping these jurors’ verdict decisions – because the evidence indicated legal guilt beyond reasonable doubt, including the defendant’s confession, on the stand, that he had acted with the intent to kill the victim.

On the other hand, however, neither were jurors driven only by their pre-existing attitudes toward the moral wrongfulness of euthanasia. Attitudes influenced jurors’ verdicts only when jurors learned about their ability to nullify the law. That is, we would be wrong to conclude that jurors throw legal instructions and evidence to the wind and rely only on their attitudes. Instead, when we accounted for the moderating effect of nullification instructions, we found that euthanasia attitudes predicted verdicts only when jurors learned about the option to nullify. Granted, Skitka and colleagues measured the strength of people’s moral convictions about a topic and not merely the strength of their attitudes. This might be an important distinction if jurors who, for example, strongly supported euthanasia, did so for practical reasons and with only moderate moral conviction. Moral convictions are more than just strong attitudes related to moral issues; and in our study, we measured attitudes, whereas their authority independence hypothesis relied on measures of moral conviction.

Although others have shown that pre-deliberation attitudes influence verdicts regardless of instruction type (Meissner et al., 2003), our findings are in line with prior research showing that, for attitudes to influence verdicts, some other factor must be present to ‘free’ jurors to rely on their beliefs. Non-experimental research also suggests that pre-trial attitudes (death penalty, insanity defense, etc.) matter less than trial-relevant judgments (eg expert witness evaluations, defendant’s mental status) when it comes to jurors’ verdicts (Poulson, Braithwaite, Brondino, & Wuensch, 1997; Poulson, Brondino, Brown, & Braithwaite, 1998). For example, both nullification instructions and incidentally induced anger (ie anger elicited before the trial by a completely unrelated manipulation) increased jurors’ reliance on their pre-trial attitudes – the former because it gave jurors explicit permission to do so, the latter because anger tends to increase people’s reliance on heuristics and their certainty that their beliefs are right (Peter-Hagene & Bottoms, 2017). We can only speculate about the reasons why our results and the results from the recent Peter-Hagene and Bottoms study differ from the results observed by Meissner and colleagues. In their study, attitudes were strongly correlated with pre-deliberation verdicts regardless of instruction type, and even though the strongest correlation was observed in the nullification-instructions condition (r = .58), it was not significantly stronger than in the standard-instructions condition (r = .53). It is possible that the video stimulus used by Meissner et al. was more emotionally evocative than our trial summary, which might have facilitated reliance on attitudes – which was also the case for incidental anger in the Peter-Hagene and Bottoms study – regardless of nullification instructions.

In addition to relying on the experimental manipulation of explicit nullification instructions, we tested the psychological process involved in nullification by measuring jurors’ perceived obligation to follow the letter of the law. We found that jurors who read nullification (versus standard) instructions felt less bound by the law and more liberated to follow their conscience. Thus, we provided empirical support for the long-standing assumption that nullification instructions have the psychological effect of freeing jurors from their perceived obligation to follow the letter of the law.

Moral outrage as mediator of euthanasia attitudes effects on verdicts

The intersection of morality and law has received increased empirical attention from psychology and law scholars, who have started to incorporate psychological theory regarding the moral reasoning of jurors (Miller & Borgida, 2016; Votruba, Braver, Ellman, & Fabricius, 2014). Several studies have focused on the role moral outrage plays in jurors’ tendency to convict defendants (eg Peter-Hagene & Bottoms, 2017; Salerno & Peter-Hagene, 2013; Wiley & Bottoms, 2009), to accept self-defense as a reason for reduced charges (Salerno et al., 2015) and to determine liability in civil cases (McCracken & Stevenson, 2017). Yet, in our study, moral outrage toward the defendant was not a significant mediator of attitudes’ effect on jurors’ verdicts. We suspect this might be due to jurors’ overall low levels of moral outrage toward a sympathetic defendant motivated by a desire to help his patient die with dignity. More importantly, the evidence and legal instructions in this case were skewed toward guilt – meaning that a completely dispassionate juror who relied only on the defendant’s actions and the legal standards for first-degree murder would convict as much as a juror in the throes of moral outrage at the defendant.

We found support for our prediction that jurors can also become morally outraged at the law when it violates their personal beliefs. Moral outrage at the law mediated the effect of jurors’ euthanasia attitudes on verdicts – the more outraged at the law, the less likely jurors were to convict Dr Wood. Furthermore, this effect was more robust than the mediating effect of moral outrage at the defendant. Specifically, (a) a parallel mediation model revealed that only moral outrage at the law, but not moral outrage at the defendant, significantly mediated the effect of attitudes on verdicts, and (b) moderated mediation analyses revealed that moral outrage at the law predicted verdicts regardless of jury instruction type, but moral outrage at the defendant only predicted verdicts when jurors were encouraged to rely on their feelings (ie when they read nullification instructions).

Our findings are consistent with prior evidence demonstrating that people experience moral outrage when they learn of unjust political and institutional practices and policies (eg Antonetti & Maklan, 2016). Moral outrage underlies people’s recommendations for punishment against corporations who treat their employees poorly (eg sweatshop workers, Rothschild & Keefer, 2017) and against tyrannical political regimes (eg Saddam Hussein’s regime; Pagano & Huo, 2007). Furthermore, moral outrage at corporate unethical behavior can motivate direct punishment from consumers – such as boycotting (Lindenmeier et al., 2012).

Yet, moral outrage can also be a prosocial reaction to social injustice, serving to motivate support for victims in an effort to remedy the injustice (Darley & Pittman, 2003). This might be particularly likely in the context of institutional misconduct, where direct punishment against a particular transgressor is not always possible. For example, moral outrage at corporate injustice increased people’s willingness to help the disadvantaged (Wakslak, Jost, Tyler, & Chen, 2007). Moral outrage at the continued shortage of clean drinkable water in developing countries increased people’s commitment to take action to reduce this crisis (Thomas & McGarty, 2009). Similarly, jurors who experience moral outrage at a perceived unjust prosecution would not be able to endorse their punitive drives (ie they cannot punish the prosecutor or judge) – meaning the only way available to them to restore justice is through acquittal.

Implications for law, practice and public policy

Although jury nullification has received relatively little empirical attention in recent years, cases involving nullification reflect the direct influence of public morality on the legal system via jurors. For example, following the implementation of a New Hampshire law that requires courts to inform juries of their right ‘to judge the application of the law in relationship to the facts in controversy’, a jury acquitted 59-year-old Doug Darrel in a felony marijuana case in 2012 – despite clear evidence that Mr Darrel had, indeed, grown the marijuana illegally (Newman, 2012). With 62% of Americans now supporting the legalization of marijuana (Hartig & Geiger, 2018), attitudes have clearly changed, and there is no longer significant moral opposition to its use or cultivation in the general public. Thus, the language in the nullification instructions (ie ‘even if you find that the state has proved all of the elements of the offense charged beyond a reasonable doubt, you may still find that based upon the facts of this case a guilty verdict will yield an unjust result, and you may find the defendant not guilty’) might have encouraged jurors to rely on their own, more permissive values and attitudes about marijuana. There has also been speculation that those involved in the 2016 standoff at the Malheur National Wildlife Refuge in Oregon were acquitted of all charges due to an apparent campaign going on outside of the courthouse to raise awareness about jury nullification (Banzhaf, 2016). In fact, jurors’ knowledge about nullification might increasingly come from efforts to raise awareness about it through social media, for example. Jurors might not need to learn about nullification through explicit instructions – which are within the courts’ control to offer or to withhold. Thus, given the increased public awareness of nullification, it is important to understand the potential impacts of yielding this power in the legal arena. Future research on the effects of jury nullification information should focus on how jurors respond to different legal and extralegal sources of information about jury nullification.

In addition, our study shows – for the first time – that when laws or decisions to prosecute appear unfair, jurors experience more than disagreement – they are morally outraged at the law and the prosecution. This is important above and beyond negative attitudes, because stronger (versus weaker) emotional experiences are more likely to result in active attempts to correct the injustice (Wakslak et al., 2007). Furthermore, mock jurors’ perceptions of the fairness or ‘moral credibility’ of one law or legal decision can not only affect their likelihood of complying with that specific law, but can generalize and affect their likelihood of compliance with other laws in general (Nadler, 2005).

Our work also has policy implications by providing insight into people’s reactions to euthanasia. Numerous court cases have been filed by patients and physicians requesting legal permission to die with dignity (eg Bartling v Superior Court; Bouvia v Superior Court; Quill v Vacco), and physician-assisted suicide is already legal in a number of other countries and some states, a states’ right upheld by the US Supreme Court in Gonzales v Oregon. Although assisted suicide is still illegal in most jurisdictions, and people receive murder convictions for it (eg People v Cleaves; Michigan v Kevorkian), the public is becoming more accepting of individuals’ right to decide the circumstances of their death, including receiving assistance from physicians to end their lives (Moulton, Hill, & Burdette, 2006). Groups such as the Final Exit Network openly help terminally ill patients take their own lives, by providing tutorials, materials and emotional support. Yet, with one exception (Chanen, 2016), charges against Final Exit members have been dismissed or their trials have resulted in acquittals or hung juries (Baenen, 2013; Greene, 2011).

The case against nurse Barbara Mancini, accused of helping her terminally ill father commit suicide by providing him with morphine, was dismissed by a Philadelphia judge in 2014 (Vitez, 2014). This global shift in attitudes is also evident in other countries where physician-assisted suicide remains illegal, with an Australian jury finding Peter John Nixon not guilty of assisting in the suicide of his father (Carson, 2017) and an Irish jury finding Gail O’Rorke not guilty of attempting to assist in her friend’s suicide (Hayes, 2015). Our study provides empirical evidence that assessing potential jurors’ moral attitudes about the permissibility of euthanasia is crucial even when legal guilt is established.

Limitations and future directions

The research had several limitations. As in most juror decision-making studies, our mock jurors were not chosen after voir dire. Although we ensured all participants were over 18 and United States citizens, we did not ask them about criminal history – which means some participants might not have been jury eligible due to felony records. Mechanical Turk jurors participate in several studies and are therefore more familiar with research paradigms than the general population (Chandler, Mueller, & Paolacci, 2014). Although attention checks can help ensure that online participants are paying attention, it is difficult to assess participants’ engagement and effort. Yet, our patterns of results resembled those found in laboratory studies (eg Peter-Hagene & Bottoms, 2017), reducing concerns over whether participants read the trial transcript and answered questions carefully.

Mock jurors read a trial transcript, but they were not exposed to actual live witnesses or attorneys. Future studies would include audio and video stimuli, either derived from actual trials or, when the manipulations include language used by attorneys, with actors playing the roles of attorneys and witnesses, in order to increase the realism of the trial experience. Another significant limitation was the lack of deliberation. Jurors rendered individual verdicts without deliberating as a group. Although there is much research about jurors that does not involve deliberation because of the practical difficulties associated with group research, we recognize this as a principal limitation of our study. Specifically, deliberation might attenuate (Shaw & Skolnick, 1995), amplify (Haegerich, Salerno, & Bottoms, 2013; Moscovici & Zavalloni, 1969) or not affect at all (Kalven & Zeisel, 1966; MacCoun & Kerr, 1988; Sandys & Dillehay, 1995) individual jurors’ biases as compared to the average of pre-deliberation verdict preferences. Meissner et al. (2003) found that deliberation reduced the effect of pre-trial attitudes on verdicts (especially when jurors were instructed to nullify), suggesting that perhaps the effects found in the present study might be attenuated after deliberation.

As in other research involving individual juror decisions, we measured verdicts immediately after the trial and instructions – because we wanted to capture mock jurors’ decisions without focusing their attention on our proposed mediators. That is, we wanted to avoid asking jurors about their moral outrage at the defendant and at the law before we asked them to endorse a verdict decision, because we did not want them to be artificially guided by our theoretical mechanisms by priming them to think about their feelings of moral outrage when they would not have naturally done so. Yet, for any mediation argument, the choice of measuring the outcome before the mediator is a limitation: how can we be sure that the verdict decision did not drive jurors’ moral outrage reactions (eg Simon, Snow, & Read, 2004)? We argue that, theoretically, the psychological process (ie moral outrage) began while jurors were reading the trial, along with other natural psychological reactions (eg emotions, attributions of blame, ratings of witness credibility etc.) – and not after the verdict decision, when we measured the hypothesized mediators. Yet, future studies could replicate these effects by reversing the order of outcome variables.

Conclusion

Over 20 years ago, Finkel et al. (1993) linked pro-euthanasia attitudes to nullification, concluding, ‘In this life-and-death matter, black letter law and common-sense justice were not only far apart, but, in the eyes of some [jurors], irreconcilably so’ (p. 487). The legal system regards jurors as fact finders, tasked with applying the letter of the law to the evidence and rendering a factual judgment of guilt. Yet, our research confirms that many legal judgments are also moral judgments: when deciding whether a law has been broken, jurors also assess the underlying moral transgression and make judgments of moral, in addition to legal, blame, especially when reminded of their power to nullify the law.

Ethical standards

Declaration of conflicts of interest

Liana C. Peter-Hagene has declared no conflicts of interest.

Chasity L. Ratliff has declared no conflicts of interest.

Ethical approval

All procedures performed in the study involving human participants were in accordance with the ethical standards of the institutional Southern Illinois University Human Subjects 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.