Negative Emotions Felt During Trial: the Effect of Fear, Anger, and Sadness on JurorDecision Making

NARINA NUÑEZ1*, KIMBERLY SCHWEITZER1, CHRISTOPHER A. CHAI1 and BRYAN MYERS2

1University of Wyoming, Laramie, USA2University of North Carolina at Wilmington, Wilmington, USA

Summary: During trial, jurors may experience a variety of emotions, many of which are negative. The current study examined theeffects the negative emotions anger, fear, and sadness had on jurors’ sentencing decisions and explored whether the cognitive ap-praisal theory or the intuitive prosecutor model could explain these effects. Jurors viewed the sentencing phase of a capital murdertrial and were asked to sentence the defendant. Results indicated that after viewing the trial, jurors reported increased anger andsadness, but not fear. However, only change in anger affected jurors’ sentences. Jurors who reported a greater change in angerwere more likely to sentence the defendant to death. This effect was mediated by the level of importance that jurors placed on theprosecution’s evidence and argument. Consistent with the intuitive prosecutor model, increased anger led to higher ratings of theimportance of the aggravating evidence and an increase in death sentences. Implications are discussed. Copyright © 2015 JohnWiley & Sons, Ltd.

INTRODUCTION

The role of emotions in legal decision making has beendebated extensively (e.g., refer to Bandes, 1996), with thegenerally held opinion by legal scholars that emotionslargely impede rational decisions and sound judgments(Feigenson, 2000). Because jurors are required to assess cul-pability in the guilt phase of the trial and blameworthiness inthe sentencing phase for capital trials, the role that emotionsplay in judgments of responsibility and blame warrants anal-ysis. Emotional influence on legal decision making may oc-cur in a number of ways (refer to Feigenson & Park, 2006). Itmay influence decisions by serving a heuristic functionwhere the emotion or mood directly informs the decisionmaker. This affect-as-information model predicts that weuse our emotions as a replacement for more detailed infor-mation processing (Schwartz, 1990). Alternatively, emotionsmay influence decisions indirectly by either promoting de-tailed information processing strategies (e.g., Forgas, 2003)in the case of negative emotions or promoting more heuristicinformation processing (e.g., Bless et al., 1996) in the case ofpositive emotions.

Directional processing theories of emotions and judgment,according to Feigenson and Park (2006), suggest that somedecisions arise that are consistent with the valence of the par-ticular emotion. In the case of mood congruency effects,emotions may cue the decision maker to attend to particularaspects of the information that are consistent with that emo-tion. For example, the target is evaluated more negativelywhen the decision maker is experiencing negative affect(Bower, 1981; Forgas & Bower, 1987). Two other examplesof directional processing models suggest that decisionsreflect the valence of the emotion, and these decisions maybe later bolstered with supporting cognition. Both Haidt(2001) and Alicke (2000) suggest that immediate affectivereactions may arise during the act of evaluating another’s

blameworthiness, and these emotions can influence judg-ments of responsibility and blame. Importantly, both themoral intuitionist model (Haidt, 2001) and the culpable con-trol model (Alicke, 2000) suggest that the judgments arisingfrom these emotions are largely irrational, and rational judg-ment may follow later to provide post-hoc support for thejudgment (refer to Feigenson & Park, 2006, for a detailed ex-planation of this process).As Salerno and Bottoms (2009) have noted, much of the

research investigating legally relevant decisions hasemployed jury simulation methods that have focused ontwo primary mechanisms of introducing emotion-evokingevidence into the trial: gruesome crime scene photographsand victim impact statements. Investigators who haveexamined these evocative stimuli have sought to determinethe stimuli’s role in eliciting emotions and their effects onguilt judgments (Bright & Goodman-Delahunty, 2006;Douglas, Lyon, & Ogloff, 1997; Kassin & Garfield, 1991),sentencing decisions (Georges, Wiener, & Keller, 2013;Paternoster & Deise, 2011), and judgments relevant to civiltrials (Bright & Goodman-Delahunty, 2011).For example, Bright and Goodman-Delahunty (2006) pre-

sented mock jurors with a murder trial transcript and variedthe presence and gruesomeness of crime scene photographs.The more gruesome photographs led to a greater tendency tovote for conviction and rate the prosecution’s case as stron-ger. Importantly, self-reported anger ratings mediated thisrelationship between the evidence and subsequent judgmentsabout the case. Paternoster and Deise (2011) examined sen-tencing judgments following assigning mock jurors to eithera penalty phase containing a victim impact statement or a no-victim impact control condition. They found that negativeaffect partially mediated the relationship between the victimimpact statement and sentencing judgments.In a recent study, Georges et al. (2013) examined the ex-

tent to which 13 different emotions were aroused during adeath penalty case. They measured emotions prior to and atdifferent points during the trial and found that three negativeemotions increased in magnitude during the trial: disgust,frustration, and anger. At each point in the trial at which they

*Correspondence to: Narina Nuñez, University of Wyoming, 1000 E Univer-sity Ave, #3415, Laramie, Wyoming 82071, USA.E-mail: narina@uwyo.edu

Copyright © 2015 John Wiley & Sons, Ltd.

Applied Cognitive Psychology, Appl. Cognit. Psychol. 29: 200–209 (2015)Published online 7 January 2015 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/acp.3094

measured emotions (e.g., after opening arguments during theguilt phase and after opening arguments during the penaltyphase), disgust, frustration, and anger were significantlyhigher than pretrial measures. Although Georges et al.(2013) did not measure fear and sadness, we propose thatthese emotions would also increase during trial. Further,there is theoretical and empirical support to predict that thearousal of fear and sadness would lead to different decisionoutcomes than the arousal of anger.However, as Salerno and Bottoms (2009) have noted,

researchers have not tested particular theories as to how emo-tions influence legal judgments (refer to Georges et al., 2013,for an exception). The present study will explore two modelsas predictors of sentencing: the cognitive appraisal theoryand intuitive prosecutor theory. We therefore next revieweach of these theories in more detail.

Emotions and decision making: appraisal theories

Appraisal theories, such as Tiedens and Linton’s (2001) cog-nitive appraisal theory, posit that emotions affect decisionsbecause they elicit specific processing styles. For example,anger is thought to lead to a narrowing information process-ing style, which then influences subsequent decisions. Ac-cording to this approach, specific negative (and positive)emotions are associated with a unique configuration ofcognition, or appraisals, about the perceived cause and na-ture of an emotional experience. Although there is somedebate about the causal influence of these cognitive ap-praisals, there is consistent evidence that certain thoughtsand perceptions are accessible when a person experiences aparticular emotion.Anger and sadness are the emotions that have been shown

to lead to different processing and decision styles. To illus-trate, Keltner, Ellsworth, and Edwards (1993) induced sadand angry moods in individuals and asked them to evaluatehypothetical outcomes. They found that angry individualstended to make more dispositional attribution, whereas sadindividuals tended toward situational attribution. Moreover,people tended to feel angry when they perceived that anattainable goal had been blocked by the actions of anotherperson (Lerner & Tiedens, 2006). As a result, anger is asso-ciated with a general sense of certainty about the cause and astrong motivation to approach desired ends (Harmon-Jones,Sigelman, Bohlig, & Harmon-Jones, 2003). In contrast, peo-ple tended to feel sad when they perceived that some negativeself-relevant event had occurred as the result of an uncontrol-lable circumstance or situation (Keltner et al., 1993). Oncearoused, sadness is associated with a general sense of uncer-tainty (Smith & Ellsworth, 1985) about the cause and a morecautious mindset (Forgas, 1999). Thus, emotional states areassociated with quite distinct cognitive appraisal patterns.In the first of a series of studies, Tiedens and Linton

(2001) induced either disgust or fear in participants and thenasked them to describe how certain they were (e.g., aboutfuture events and about what would happen next) whenexperiencing the emotion. They found that disgust was asso-ciated with certainty and fear was associated with uncer-tainty. In a subsequent task where participants were askedto predict events that would happen the following year and

to rate their certainty that the event would occur, participantsin the disgust condition were significantly more confidentabout their predictions than people in the fear condition.Across four studies, they found that participants who recalledexperiencing emotions that increase certainty appraisals(including anger in one study) were more likely to use stereo-types in their judgments and to be persuaded by weak argu-ments delivered by a credible source (i.e., peripheral cue).In contrast, participants recalling emotions that elicit uncer-tainty (including sadness in one study) showed the oppositeeffect. Based on these results, Tiedens and Linton (2001)concluded that emotions that prime certainty (e.g., anger) in-duce shallow, heuristic information processing, while emo-tions that prime uncertainty (e.g., sadness) promote morethorough and systematic information processing.

Evidence also suggests that sadness and anger can havediverging effects on juror information processing. Semmlerand Brewer (2002) randomly assigned mock jurors to hearone version of a vehicular manslaughter case that had beendeveloped to evoke either sadness or a nonemotional state(i.e., neutral mood control). They also manipulated the pres-ence of factual and logical inconsistencies in the testimonyof witnesses. Mood did not influence the number of inconsis-tencies detected, but, consistent with the cognitive appraisaltheory, jurors who heard the sad version of the trial weremore accurate in recalling the details of the inconsistenciescompared with participants in the neutral condition. Al-though anger was not manipulated in the study, the re-searchers were able to examine the influence of this emotionon information processing using participant ratings from amood manipulation check. As self-reported anger increased,the detection of inconsistencies and accuracy of the inconsis-tencies detected decreased (Semmler & Brewer, 2002).Thus, sadness appeared to facilitate juror information pro-cessing, while anger tended to impede it.

Research indicates that fear, much like sadness, elicits un-certainty (Lerner & Keltner, 2000; Raghunathan & Pham,1999) and leads to more thorough information processing(Kligyte, Connelly, Thiel, & Devenport, 2014; Lerner &Keltner, 2000; Parker & Isbell, 2010). Parker and Isbell(2010) examined differences between fear and anger on po-litical information processing. After inducing fear or anger,participants were given the names of two mock candidatesthat were running against each other in a democratic pri-mary. Participants were allowed to access a website contain-ing information about the candidates and the positions thatthey held on a variety of issues. Parker and Isbell (2010)found that, as compared with the angry condition, partici-pants in the fear condition spent more time searching thewebsite for information and were subsequently more likelyto choose a candidate whose positions matched their own.In sum, fear, similar to sadness, appears to lead to more care-ful information processing and more informed decisions.

Appraisal theories suggest that emotions might affect de-cisions because, once aroused, emotions alter the depth of in-formation processing. Anger leads to greater certainty andshallower processing, while fear and sadness lead to greateruncertainty and facilitate processing. Thus, in a capital trial,jurors who become angry might narrow their processing ofinformation and might pay less attention to the details

Negative emotions and decision making 201

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

presented. They may also be more certain that their sentencingchoice was correct (i.e., have greater confidence in their sen-tencing decision). On the other hand, jurors who become fear-ful or sad may pay more attention to trial details and be lesscertain that they have made the correct sentencing choice.

In the present study, we varied the strength of the mitigat-ing factors in the case. Half of the jurors received weak andhalf received strong mitigating evidence (explained in fur-ther detail later in this paper). We later asked jurors to recallthe mitigating evidence presented after watching the trial,and each juror received a recall score. We predicted that tothe extent that the cognitive appraisal theory was correct,jurors who became sad and fearful during trial would be asfollows: (i) better able to recall details from the trial and(ii) more sensitive to the weak versus strong mitigating evi-dence. In other words, if jurors were in the strong condition,they would hear mitigating evidence that argued for leniencyand would therefore be more likely to give a life sentence,but in the weak condition, there would not be compelling ev-idence to argue for leniency, and jurors would be more likelyto give a death sentence. On the other hand, jurors who be-came angry were expected to use a more heuristic style ofprocessing, and they would subsequently recall fewer detailsof the evidence from the trial and not be as sensitive to theweak and strong mitigating evidence manipulation. Consis-tent with the cognitive appraisal theory, if any of theemotions (anger, fear, or sadness) significantly predictedsentencing decisions, recall scores should mediate the rela-tionship. Finally, if the appraisal theory is correct, angershould be associated with greater confidence in one’s sen-tencing decisions, while sadness and fear should result inless confidence in one’s verdict. Consequently, we wouldpredict that confidence in verdict (i.e., certainty) as well asthe amount of trial information recalled should mediate rela-tionships between emotion and sentencing decisions.

Emotions and decision making: the intuitive prosecutor

A slightly different view of the effect of emotions (and par-ticularly the role of anger) on decision making has beenproposed by Tetlock and colleagues (Goldberg, Lerner, &Tetlock, 1999; Rucker, Polifroni, Tetlock, & Scott, 2004;Tetlock et al., 2007), Darley (Darley, 2001; Darley &Pittman, 2003; Solan & Darley, 2001), and their colleagues.In their view, anger comes from moral outrage that onemight feel at trial, and this moral outrage requires some ret-ribution or punishment. Although they acknowledge the ap-praisal framework, they posit that decisions reached whileangry will be punitive. That is, in their view, it is the desireto mete out justice that is the driving force for angry jurors’decisions (i.e., there is an increased motivation to punishrather than a shift in information processing, as appraisaltheorists might suggest), and so, this is consistent with otheraffect valence models of emotion and decision making (referto Feigenson & Park, 2006), and quite different from thecognitive appraisal theory. Tetlock and colleagues (Goldberget al., 1999; Tetlock et al., 2007) refer to people who areangry as ‘intuitive prosecutors’.

In one study, Tetlock and colleagues (Goldberg et al., 1999)induced anger in participants and then had participants read

four crime vignettes and rate the extent to which the defen-dant should be punished, blamed, and so on. Importantly, inthis study, the anger manipulation (a video of a man beatinga helpless teenager) was followed by justice feedback.Following the video, participants were told that either thewrongdoer was caught and punished (punished condition),he was caught but escaped punishment because of a techni-cality (unpunished condition), or there was no informationabout the fate of the wrongdoer (ambiguous condition).Results indicated that subsequent ratings on the crimevignettes by participants were influenced by the justicefeedback that they received. As predicted, only jurors in theunpunished condition were more punitive in their ratings ofthe crime vignettes. Thus, the anger induction spilled overto the next task when participants believed that an injusticehad occurred in the crime that they previously watched.According to Goldberg et al. (1999), appraisal theories wouldpredict that anger would lead to an increasing willingnessto blame individuals for wrongdoing, but appraisaltheories would not account for why only anger in the unpun-ished condition would do so. Instead, they suggest thatpeople are motivated to maintain social order and that angeraffects decision making by arousing the ‘intuitive prosecutor’in jurors.If true, we would make slightly different predictions about

the role of anger, fear, and sadness on jurors’ decisions. Wewould predict that only anger would affect decisions and thatincreased anger would be associated with increased punitive-ness. Sadness and fear should not lead to punitive behaviorbecause they do not elicit moral outrage or the intuitive pros-ecutor. However, if anger arouses an intuitive prosecutor, wewould expect that jurors would consider the prosecution’scase to be more influential to their decisions, and this influ-ence would moderate the relationship between anger anddecisions. That is, in the present study, we asked participantsto rate the importance of crucial pieces of evidence in theirdecision making. If the intuitive prosecutor model is correct,angry jurors should rate evidence that favors the prosecutoras more influential than evidence from the defense, and theinfluence of the prosecution would moderate any relation-ship between anger and decisions.It should be noted that we are adopting the intuitive

prosecutor model of emotion and decision making to com-pare with the cognitive appraisal theory, but the intuitiveprosecutor model is highly consistent with other directionalmodels such as Haidt’s (2001) moral intuitionist model andAlicke’s (2000) culpable control model, and in a number ofrespects, the affect-as-information model by Schwarz(1990) would yield similar predictions as the intuitive prose-cutor model.

The present study

The purposes of the present study were to examine whatemotions are aroused at a capital trial and to test how thoseemotions impacted jurors’ decisions. Mock jurors watcheda video of the sentencing phase of a capital trial and wererandomly assigned to one of two trial conditions. In the weakmitigating evidence condition (n=80), jurors were told thatthe defendant’s intelligence quotient (IQ) was 80, the

202 N. Nuñez et al.

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

defendant was possibly physically abused as a child, and hehad been previously diagnosed with depression. In the strongmitigating evidence condition (n=79), jurors were told thatthe defendant’s IQ was 69, the defendant had experiencedextensive childhood sexual abuse, and he had been previ-ously diagnosed with schizophrenia.

Hypotheses

1. After watching the trial video, participants will report in-creases in anger, fear, and sadness.

2. Based on appraisal theories and the intuitive prosecutormodel, we predict that increases in anger would resultin more death penalty decisions.

3. If appraisal theory is correct,

(a) Sadness and fear should not affect sentence directly.Instead, in the weak mitigating condition, sadnessand fear should be associated with death sentences,but in the strong mitigating condition, fear and sad-ness should lead to increases in life sentences.

(b) Mitigating evidence recall scores should be higherwhen there are increases in sadness and fear andlower with increases in anger.

(c) Recall scores would mediate significant relationshipsbetween emotions and sentencing decisions.

(d) Increases in anger should be associated with greaterconfidence in verdict choice, while increases in sadnessshould lead to less confidence in verdict choice. If so,then verdict confidence should also mediate the signif-icant relationships between emotions and sentencingdecisions.

4. If the intuitive prosecutor model (e.g., Tetlock et al.,2007) is correct, only anger should be related to sentenc-ing decisions, and increasing anger should lead toincreases in death decisions. In addition,

(a) Anger, but not fear and sadness, would lead partici-pants to become ‘intuitive prosecutors’, rating aggra-vating evidence as more important to their verdictsthan mitigating evidence.

(b) Ratings of aggravating evidence would mediate therelationship between anger and sentencing decisions.

METHOD

Participants and design

One hundred fifty-nine death-qualified participants (68 maleand 91 female subjects) ranging in age from 18 to 66years(M=24.67, SD=9.73) participated. Participants were re-cruited through undergraduate psychology research pools attwo local universities (n=96), as well as Amazon’s Mechani-cal Turk (mTurk; n=63). University participants received par-tial course credit for their participation, andmTurk participantswere paid $8. Participants were randomly assigned to one oftwo conditions: strong or weak mitigating evidence. The pri-mary dependent variables were sentence (life in prison withoutthe possibility of parole or death) and confidence in one’s sen-tencing decision.

Materials and procedure

All participants completed the death qualification question-naire, signed a consent form, completed the Positive and Neg-ative Affect Scale—Expanded Form (PANAS-X), viewed avideo reenactment of the sentencing phase of a capital murdertrial, and completed a posttrial questionnaire. University par-ticipants completed the study in a research laboratory andresponded to the questionnaires via paper and pencil. mTurkparticipants completed the entire study online, at a locationof their choosing, after accessing the study link from themTurk website. The online study was created using the surveytool Survey Gizmo. All three subsamples completed the samematerials, with one exception. To ensure that the mTurk partic-ipants were paying attention, as we had no way of monitoringthis, three attention-check questions were asked after the trialvideo. Any participants who did not answer the questions cor-rectly were not allowed to continue with the study. Resultsindicated no significant differences between the three samples.As such, all analyses were collapsed across samples.

Death qualification questionnairePotential participants completed a three-item death qualifica-tion questionnaire to qualify them to participate. They wereasked to respond to death qualification questions consistentwith the Wainwright v. Witt (1985) and Witherspoon v.Illinois (1968) criteria. The questions assessed whether pos-sible participants were willing to give the death penalty andwhether they would give the death penalty no matter the cir-cumstances. Participants who responded that their attitudetoward the death penalty would seriously affect their abilityto perform their juror duties, that they would not find the de-fendant guilty to avoid the possibility of giving the deathsentence, or that they would either always or never givethe death penalty were excluded from participating and con-sidered not death qualified.

Rating of aggravating evidenceTwo questions assessed the importance of the prosecutionevidence. Participants were asked to rate how important eachof the following was to reaching their sentence: (i) The de-fendant committed the crime while he was on parole for an-other felony and (ii) the defendant committed the crimewhile in the commission of another felony. Mock jurorsrated the questions on a nine-point Likert scale (1= totallyunimportant, 9 = totally important). Answers on the twoquestions were highly correlated (r= .81), and an aggravat-ing evidence variable was constructed as the mean of thetwo questions.

Rating of mitigating evidenceUsing the same scale as in the previous measures, three ques-tions assessed the importance of defense evidence in deci-sions. Mock jurors rated how important each of thefollowing was in their sentencing decisions: (i) The defen-dant has a low IQ; (2) the defendant has mental health prob-lems; and (iii) the defendant has a difficult childhood.Responses on the three items were highly related (α= .84),and a mitigating importance variable was created as themean of the three items.

Negative emotions and decision making 203

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

PANAS-XParticipants completed a portion of the PANAS-X (Watson& Clark, 1994) both before and after watching the sentenc-ing phase of a reenacted capital murder trial. The PANAS-X examined participants’ affect on five domains: fear, anger,sadness, joviality, and attentiveness. Fear, hostility, and sad-ness represented basic negative emotions and were the focusof the present study; measures of attentiveness and jovialitywere included to mask our interest in negative emotions. Par-ticipants were given a list of 29 negative and positive emo-tion words and were asked to rate their current level ofeach emotion on a five-point Likert scale (1 = very slightlyor not at all, 5 = extremely). Six words comprised the fearsubscale (afraid, scared, frightened, nervous, jittery, andshaky), six words comprised the anger subscale (angry, hos-tile, irritable, scornful, disgusted, and loathing), and fivewords comprised the sadness subscale (sad, blue, down-hearted, alone, and lonely). Scores on each subscale weresummed, allowing fear and anger scores to range from 0 to30 and sadness scores to range from 0 to 25.

Trial videoAll participants viewed a reenactment of the sentencingphase of a capital murder trial based off of actual trial tran-scripts. The trial lasted approximately 45min and describedthe murder of Jeff Mason. Participants were told that thedefendant, Anthony Prentice, shot and killed Jeff Masonwhile attempting to rob the local convenience store whereJeff Mason worked as a clerk. They were told that as fol-lows: (i) The defendant demanded that Mr. Mason openthe store safe, and when Mr. Mason was unable to openthe safe, the defendant repeatedly beat Mr. Mason on thehead with the pistol and then shot him in the head; (ii) therewas a witness in the store who saw the crime and identifiedthe defendant; (iii) the defendant was found eight blocksfrom the crime, and his T-shirt was splattered with blood thatwas later matched to Mr. Mason; (iv) the police found thegun used in the crime in the bushes where the defendantwas found, fingerprints on the gun matched Anthony Pren-tice’s, and the blood on the gun matched Mr. Mason’s. Mockjurors were informed that the jury found Mr. Prentice guiltyof first-degree murder and that they would serve as mockjurors for the penalty phase of the trial.

The videotaped trial started with the opening commentsfrom the judge and the attorneys’ opening arguments. Then,a parole officer presented aggravating evidence stating thatthe defendant was on parole at the time that he committedthe crime and that he committed the crime while committinganother felony, meeting the prosecution’s standard of prov-ing beyond a reasonable doubt at least one aggravatingfactor. The next witness for the prosecution was the victim’swife. She talked about how the death of her husband affectedher and her children. For the defense, a clinical psychologistpresented mitigating evidence. In the strong mitigatingevidence condition, the defendant’s IQ was said to be 69,the defendant had experienced extensive childhood sexualabuse, and he was previously diagnosed with schizophrenia.In the weak mitigating evidence condition, the defendant’sIQ was said to be 80, the defendant was physically abused

as a child, and he was previously diagnosed with depression.Lastly, participants saw the attorneys’ closing arguments andreceived the judge’s jury instructions.The mitigating evidence chosen for the trial were items that

Barnett, Brodsky, and Davis (2004) reported to significantlyaffect life and death decisions of mock jurors. Barnett et al.(2004) gave mock jurors 10 different vignettes, some of whichcontained different types of mitigating evidence. Low IQ,mental illness, and child abuse were all associated with morelife sentences. In the current study, we developed weak andstrong mitigating evidence conditions by varying the severityof each of these items. The strength of mitigating evidencewas pilot tested with mTurk participants (N=20). Participantswere told that ‘at capital trials the defense often providesexpert testimony to convince a jury to show mercy by givingthe defendant life rather death’. They read each part of themitigating evidence (i.e., IQ, family background, and mentalillness) and were asked to rate on a seven-point Likert scale(1=not at all convinced that life in prison without parole isa good sentence, 7 = totally convinced that life in prison with-out parole is a good sentence) whether that piece of evidenceconvinced them that life in prison without parole was the bestsentence. The evidence was presented in the same order as thevideo trial. Although there may have been an effect of order,we wanted to mimic what would later be presented at trial.A repeated-measures ANOVA with mitigating strength

(weak or strong) as the independent variable and the ratingsas the dependent variable revealed a significant effect of mit-igating evidence, F(1, 18) = 6.86, p= .02, η2 = 0.28. Theweak mitigating condition was rated significantly lower(M=3.77, SE=0.48) than the strong mitigating condition(M=5.53, SE=0.48) indicating that participants were moreconvinced that life was the best penalty in the strong ratherthan the weak mitigating condition.

Posttrial questionnaireAfter viewing the trial recording, participants were asked tocomplete the PANAS-X for a second time and to sentencethe defendant to either life in prison without the possibilityof parole or death, as well as rate their confidence in their de-cision (counterbalanced across participants). Participantswere asked to rate how confident they were that they chosethe right sentence on a 10-point Likert scale (1 = totally notconfident, 10 = totally confident). Participants then com-pleted a series of questions regarding what mitigating evi-dence they recalled and how the aggravating and mitigatingevidence impacted their sentence.

RESULTS

Data preparation

Recall of mitigating informationUndergraduate research assistants who were blind to thestudy hypotheses coded answers to the free recall. Mock ju-rors were instructed to ‘describe all of the mitigating factorsthat you remember in this case’ as an open-ended question.Each of the four mitigators (three manipulated mitigators,IQ, history of childhood abuse, and mental illness; one con-stant mitigator, school difficulties and failure) received a

204 N. Nuñez et al.

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

score of 0, 1, or 2. Responses were given a score of 0 if themitigator was not mentioned, a score of 1 if the mitigatorwas mentioned but no other details were provided (e.g.,low IQ), and a score of 2 if the mitigator was mentionedand details were provided (e.g., an IQ of 69 in the strongmit-igating condition or 80 in the weak condition). Two codersscored each response, and raters achieved adequate interraterreliability (κ=0.88). Discrepancies in scoring werediscussed and settled by the primary authors. Recall scoresranged from 0 to 8 (M=3.68, SD=1.80).

General hypotheses

Hypothesis 1: After watching the video, participants will re-port increases in anger, fear, and sadness. In order to assesschanges in fear, anger, and sadness, we ran repeated-measuresANOVAs with scores from the PANAS-X (before and aftertrial) as the repeated measure (assumptions of sphericity weremet, and no correctionwas applied).We found that fear did notsignificantly increase after viewing the trial video, F(1, 158)=0.21, p= .62, ηp2 =0.001, and as such, fear was not includedin any further analyses. Sadness and anger, however, did in-crease after viewing the trial, F(1, 158)=77.63, p< .001,ηp2 =0.33 and F(1, 158)=80.73, p< .001, ηp2 =0.34, respec-tively. As can be seen in Table 1, mock jurors reportedincreases in sadness and anger after viewing the trial. In orderto examine those changes in subsequent analyses, we com-puted change scores for anger and sadness by subtracting theTime 1 PANAS-X score from the Time 2 PANAS-X score(ΔAnger=PANAS-X2Anger�PANAS-X1Anger; ΔSad=PANAS-X2Sad�PANAS-X1Sad).

Hypothesis 2: Increases in anger would result in more deathpenalty decisions. A logistic regression was run with sen-tencing (life or death) as the dependent variable and ΔAngeras the independent variable. As predicted, ΔAnger influenceddecisions, X2(1) = 4.34, p= .04, b=0.09, and as can be seen inTable 2, participants who chose a death sentence had higherΔAnger scores than those who chose a life sentence.

Appraisal theory hypotheses

Hypothesis 3a: In the weak mitigating condition, sadnesswould be associated with death sentences, but in the strongmitigating condition, fear and sadness would lead to in-creases in life sentences. The variation of weak and strongmitigation would not be significant for changes in anger. Inorder to test this hypothesis, we ran two logistic regressions.In the first, we examined whether ΔAnger scores interacted

with mitigating condition (strong or weak) to affect sentenc-ing decisions. The analysis using a model that includedΔAnger, mitigating condition, and the interaction betweenΔAnger andmitigating condition yielded only the main effectfor ΔAnger, X2(1) = 4.34, p= .04, b=0.09. Mitigating condi-tion and the interaction between ΔAnger and mitigating con-dition were not significant, X2(1) = 0, p= .99 and X2(1) = 0.95,p= .33, respectively (Table 2). The second logistic regressionwas the same as the first, except with ΔSad scores as theindependent variable. Contrary to expectations, none of thevariables reached conventional levels of significance (ΔSadscores, X2(1) = 0.58, p= .45; mitigating condition, X2(1)= 0.13, p= .71; ΔSad by mitigating condition interaction,X2(1) = 2.13, p= .15) although the predicted interactionapproached significance (Table 2).

Hypothesis 3b: Recall scores would be higher when thereare increases in sadness and lower with increases in anger,and recall scores would mediate significant relationships be-tween emotions and sentencing decisions. A linear regres-sion analysis was performed to examine whether emotionspredicted recall scores. The omnibus test for the regressionwas not significant, F(2, 154) = 1.07, p= .35, and importantlyneither ΔAnger nor ΔSad predicted recall, (β =�.13,t=�1.31, p= .19 and β = .02, t= .20, p= .42, respectively).Although the analysis did not reach conventional levels ofsignificance, the pattern of results suggests that as ΔAngerincreased recall decreased.

Hypothesis 3c: Increases in anger would be associated withgreater confidence in verdict choice, while increases insadness should lead to less confidence in verdict choice. Aregression analysis was conducted with verdict confidenceas the dependent variable and ΔAnger and ΔSad scores asthe independent variable. Contrary to expectations, the modeldid not reveal a significant effect, F(1, 157) =1.99, p= .12,R2 = .01. ΔAnger did not significantly predict confidence(β = .09, t=0.93, p= .35); however, ΔSad scores weremarginally associated with certainty, (β =�.16, t=�1.69,p= .09), and the slope was in the predicted direction.Increases in sadness were associated with decreases in sen-tencing decision confidence.

Table 1. Changes in emotions from pretrial to posttrial

Emotion Mean SE 95% CI lower 95% CI upper

FearPretrial 7.43 0.18 7.07 7.78Posttrial 7.50 0.22 7.07 7.93

SadnessPretrial 6.52 0.20 6.11 6.92Posttrial 8.40 0.27 7.88 8.93AngerPretrial 6.87 0.15 6.58 7.12Posttrial 9.53 0.31 8.92 10.13

Table 2. The changes in emotion and effects of mitigating evidencefor participants who selected life and selected death sentences

Life sentence Death sentence

M SE M SE

ΔAnger* 2.10 0.37 3.57 0.47ΔSad 1.83 0.27 1.98 0.35ΔAngerWeak mitigation 2.10 0.52 4.56 0.64Strong mitigation 2.10 0.52 2.43 0.67

ΔSadWeak mitigation 1.46 0.39 2.25 0.48Strong mitigation 2.18 0.38 1.68 .48

*p< .05

Negative emotions and decision making 205

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

Intuitive prosecutor hypotheses

Hypothesis 4a: Anger, but not fear and sadness, would leadparticipants to become ‘intuitive prosecutors’, and partici-pants who became angry when watching the trial would rateaggravating evidence as more important to their verdictsthan mitigating evidence. The importance of mitigatingevidence was examined with a linear regression analysis withthe mitigating importance score as the dependent variableand ΔAnger scores and ΔSad scores as the independent vari-ables. The omnibus test was not significant, F(2, 156) = 0.13,p= .88. Also, neither ΔSad (β= .05, t=0.48, p= .63) norΔAnger (β=�.04, t=�0.43, p= .67) predicted mitigatingimportance. A second linear regression examined if ΔAngeror ΔSad predicted aggravating evidence scores. The analysisyielded a significant overall effect, F(1, 162) = 5.94, p= .008.ΔAnger (β= .24, t=2.52, p= .01), but not ΔSad (β = .01,t=0.13, p= .90), related to aggravating evidence. As pre-dicted, as anger increased, the importance of the aggravatingevidence increased.

Hypothesis 4b: Ratings of aggravating evidence wouldmediate the relationship between anger and sentencingdecisions. Multiple regression analyses were conducted toevaluate each component of the proposed mediation model.First, results indicated that ΔAnger was positively associatedwith sentence, β = .11, t(157) = 2.34, p= .05. The results alsosuggested that ΔAnger was positively related to aggravatingevidence, β= .11, t(157) = 3.15, p= .01. Finally, resultsindicated that the mediator, aggravating evidence, was posi-tively associated with sentence, β = .30, t(157) =2.48, p= .05.After concluding that the a-path and b-path in this modelwere significant, a mediation analysis was tested using thebootstrapping method with bias-corrected confidence esti-mates (MacKinnon, Lockwood, & Williams, 2004; Preacher& Hayes, 2008). In the current study, the 95% confidenceinterval of the indirect effects was obtained using 1000bootstrap resamples (Preacher & Hayes, 2008). Results ofthe mediation analysis confirmed the mediating role ofaggravating evidence in the relationship between ΔAngerand sentence, β= .34, CI [0.01, 0.09]. Lastly, resultsindicated that the direct effect of ΔAnger on sentence be-came nonsignificant, β = .08, t(157) = 1.69, p= .09, whencontrolling for aggravating evidence, thus suggesting fullmediation (Figure 1).

DISCUSSION

One of the goals of the present study was to examine if neg-ative emotions are aroused during the sentencing phase of acapital trial, and our hypothesis was partially supported. Wemeasured fear, anger, and sadness pretrial and posttrial andfound that mock jurors experienced significant increases inlevels of anger and sadness while watching the trial. Therewas no significant increase in fear.

Our predictions about how emotions would affect sentenc-ing relied on two slightly different theories of emotion anddecision making: appraisal theories (e.g., Tiedens & Linton,2001) and the intuitive prosecutor model (e.g., Tetlock et al.,2007). As previously mentioned, both models predicted sim-ilar effects of anger on decision making but posit different

mechanisms to account for decisions. For example, withregard to sadness, appraisal theories would predict that sad-ness would affect verdicts, but the intuitive prosecutor modelwould not. Also, differences between anger and sadness ondecisions are hypothesized to be the result of differentstyles of processing according to appraisal theorists, whilethe differences are hypothesized to be the result of a prose-cution bias according the intuitive prosecutor model. Ourresults more fully supported predictions made using theintuitive prosecutor model and offered little support forappraisal theories.According to appraisal theories, anger and sadness could

both influence decisions because they lead to different stylesof information processing. Because sadness leads to a deeperlevel of processing, we predicted that increases in sadnesswould lead to more death sentences when the mitigating ev-idence was weak and fewer death sentences when mitigatingevidence was strong. We also predicted that increases in sad-ness would lead to better recall of mitigating information andthat increases in anger would lead to decreases in the recallof mitigating information. Finally, we predicted that angerwould lead to more confidence in sentencing decisions andsadness would lead to less certainty. None of our predictionsusing the appraisal framework reached conventional levelsof significance. However, the pattern of results was oftenconsistent with appraisal theory predictions. For example,we found that sadness led to less confidence in sentencingdecision, and increased anger led to lower recall scores.Overall, though, the results did not consistently mesh withappraisal theory predictions. With regard to recall, it maybe argued that the lack of differences in recall across emo-tions may be caused by the information being too easy torecall. However, there were no ceiling effects present forrecall. The mean recall score indicated that participantsrecalled slightly less than half of the mitigating evidencedetails presented.On the other hand, we found consistent support for the intu-

itive prosecutor model. As predicted, anger, but not sadness,influenced verdicts. Increases in anger were associated with in-creases in death sentence decisions. We also found thatincreases in anger resulted in more death sentencesregardless of whether the mitigating evidence was strong orweak (i.e., there was no interaction between anger scores andmitigating evidence strength). Finally, as predicted by the intu-itive prosecutor model, mock jurors who became angry rated the

Figure 1. Indirect effect of ΔAnger on sentence through prosecu-tion importance. Note: *p< .05, **p .01

206 N. Nuñez et al.

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

importance of the aggravating evidence higher, and the impor-tance of the aggravating evidence mediated the relationship be-tween anger and sentencing decisions. This is an important as-pect of the findings because it suggests that the level of angercovaried with the support for aggravating evidence—a findingthat would not be predicted by the appraisal model.Nevertheless, it is important to note that because we did

not manipulate anger, we cannot draw a causal connectionbetween anger and evidence ratings. Moreover, the relationbetween evidence ratings and emotions would be consistentwith other directional models, and not just the IntuitiveProsecutor Model. Both Alicke’s (2000) model as well asHaidt’s (2001) suggest that emotions drive the judgmentson a largely automatic and irrational level, and then post-hoc rationalizations for the judgments soon follow. Conse-quently, rating the aggravating evidence higher in impor-tance when angered would be predicted by any of thesedirectional models, including the affect-as-informationmodel (Schwarz, 1990), but would not be predicted by thecognitive appraisal theory. Thus, neither these findings areuniquely tied to the intuitive prosecutor model nor is thetendency for jurors to assign greater importance to evidencethat is related to their emotions. Indeed, Bright andGoodman-Delahunty (2004) also found that anger impactedthe weight that jurors assigned to the prosecution evidence.One interesting side note about the present findings is that,

as described in the introduction, Goldberg et al. (1999) con-ducted a study that provided evidence that anger led to aprosecutorial mindset. In their study, they used an incidentalemotion paradigm. They exposed participants to an angerinduction task and found that incidental anger led people tobecome more punitive. In our study, emotions were integralto the task. Mock jurors became angry or sad because of thecase circumstances. The fact that we found support for theintuitive prosecutor model in our study suggests that angerleads to a prosecutorial mindset if it is either incidental or in-tegral to the task.Our findings not only provide support for the intuitive

prosecutor model (Tetlock et al., 2007) but also are relevantto concerns that other scholars have voiced about emotionaljurors. We found that anger affected decisions and that theintuitive prosecutor model provided the best explanationfor the effect. However, future research should explore thistheory in more depth. For example, according to the model,anger leads people to want to punish, but it is unclearwhether the anger needs to be focused at a specific target(e.g., the defendant) or if free-floating anger (e.g., from amood induction) could result in greater punitiveness.Goldberg et al. (1999), described previously, induced angerby showing participants a video beating of a teenager andlater told participants that the perpetrator got away with thecrime (unpunished condition). Participants who were in theunpunished condition were more punitive when later givenother case vignettes. However, the mood induction in thisstudy was highly similar to the subsequent task. Both wereexamples of criminal behavior, and the desire to punish theperpetrator of the teen beating could have spilled over toother wrongdoers. It would be interesting to see if angermood inductions not as closely tied to the subsequent deci-sion task would lead to similar results as those obtained by

Goldberg et al. (1999). Future research could also be directedtoward investigating the role of emotions in judgments forother types of criminal cases, especially cases in which jurorscan choose from multiple possible verdicts (e.g., first-degreeor second-degree murder, manslaughter, or not guilty).

Future research should also examine which aspects ofcriminal trials are likely to arouse anger. In our trial,there were many pieces of evidence that could havearoused anger. However, we only assessed anger at twopoints in the trial (before and after) and do not knowwhat pieces of evidence aroused anger or if it was thetrial in its totality that increased anger. Repeated mea-sures of anger during crucial points in a trial could an-swer this question, as well as measure whether there isa tipping point at which anger affects decisions.

One of the problems with assessing emotions at variouspoints in a trial is that repeated assessment might alter mockjurors’ emotional states and disrupt the flow of the trial,which might also affect jurors’ emotional states. A previ-ously mentioned study by Georges et al. (2013) used a cleverdesign to surmount this potential problem. They used abetween-subjects design and assessed jurors’ emotions priorto trial and then at one other point in time (e.g., after openingarguments during the guilt phase or after opening argumentsduring the penalty phase). Using this design, they showedhow emotions, such as anger, varied during different por-tions of a capital trial.

Limitations

We did not find an increase in fear in our trial, and while itwould be tempting to say that fear is not relevant to jurors’decisions, that may not be true for jurors in actual trials.Our jurors knew that they were participating in a study,and although they watched a realistic reenactment of a trial,they were not sitting close to or looking at a live defendant.They did not see pictures of the crime scene, which actual ju-rors could have been exposed to. They did not hear details ofthe actual murder. Thus, although our mock jurors did notexperience an increase in fear, one could reasonably proposecircumstances in which fear might increase during an actualtrial. Because there was no increase in fear in the presentstudy, we could not assess how fear affected decisions.Researchers need to devise a paradigm in which they couldstudy the effects of fear to determine if and how it affectsjuror decision making.

We did not find any evidence that emotions had an effecton the free recall of evidence, as would be predicted byappraisal theories. As previously mentioned, one potentialexplanation was that perhaps the recall measure was tooeasy. However, the range in recall scores was quite good,suggesting that there was not a ceiling effect on recall.Participants were asked to write down the mitigating evi-dence, and scores were based on what the mock jurorswrote down. Perhaps, recall scores reflected not only whatmock jurors remembered but also their willingness to fullywrite out what they remembered. A follow-up interview tofully explore mock jurors’ memory of evidence could haverevealed if the recall measure was a valid and reliable mea-sure of what mock jurors remembered.

Negative emotions and decision making 207

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

CONCLUSIONS

In the present study, we showed that mock jurors reportedincreases in sadness and anger after watching the sentenc-ing phase of a capital trial. We also showed that anger, butnot sadness, affected sentencing decisions. As anger in-creased, so did the likelihood that a juror would give adeath, rather than life, sentence. Further, we found thatas anger increased, jurors were more likely to say thatthe aggravating evidence was important to their decisions.The importance of mitigating evidence did not vary withchanges in anger. A mediation analysis demonstrated thatthe importance of aggravating evidence mediated the rela-tionship between anger and sentencing, lending supportfor the intuitive prosecutor model. There is an accumulat-ing body of evidence that suggests that emotions experi-enced at trial, particularly anger, can seep into thedecision making process. Perhaps, the next step is to seeif there are ways to mitigate the impact.

ACKNOWLEDGEMENTS

This research was funded by the National Science Founda-tion grant to the first and fourth authors, SES-0851004.

REFERENCES

Alicke, M. D. (2000). Culpable control and the psychology of blame.Psychological Bulletin, 126, 556–574.

Bandes, S. (1996). Empathy, narrative, and victim impact statements.University of Chicago Law Review, 63, 361–412.

Barnett, M. E., Brodsky, S. L., & Davis, C. M. (2004). When mitigatingevidence makes a difference: Effects of psychological mitigating evi-dence on sentencing decisions at capital trials. Behavioral Sciences andthe Law, 22, 751–770.

Bless, H., Schwartz, N., Clore, G., Golisano, V., Rabe, C., & Wolk, M.(1996). Mood and the use of scripts: Does a happy mood really lead tomindlessness? Journal of Personality and Social Psychology, 71,665–679.

Bower, G. (1981). Mood and memory. American Psychologist, 36,129–148.

Bright, D. A., & Goodman-Delahunty, J. (2004). The influence of gruesomeverbal evidence on mock juror verdicts. Psychology, Psychiatry, andLaw, 11, 154–166. doi: 10.1375/1321871041335984

Bright, D. A., & Goodman-Delahunty, J. (2006). Gruesome evidence andemotion: Anger, blame, and jury decision-making. Law and Human Be-havior, 30, 183–202. doi: 10.1007/s10979006-9027-y

Bright, D. A., & Goodman-Delahunty, J. (2011). Mock juror decision mak-ing in a civil negligence trial: The impact of gruesome evidence, injuryseverity, and information processing route. Psychiatry, Psychology andLaw, 18(3), 439–459. doi: 10.1080/13218719.2010.492095

Darley, J. M. (2001). Citizens’ sense of justice and the legal system. CurrentDirections in Psychological Science, 10, 10–13.

Darley, J. M., & Pittman, T. S. (2003). The psychology of compensatory andretributive justice. Personality and Social Psychology Review, 7,324–336.

Douglas, K. S., Lyon, D. R., & Ogloff, J. R. P. (1997). The impact ofgraphic photographic evidence on mock juror decisions in a murder trial:Probative or prejudicial. Law and Human Behavior, 21, 485–501.

Feigenson, N. (2000). Legal blame: How jurors think and talk aboutaccidents. Washington D.C.: American Psychological Association.

Feigenson, N., & Park, J. (2006). Emotions and attributions of legal respon-sibility and blame: A research review. Law and Human Behavior, 30,143–161.

Forgas, J. P. (1999). On feeling good and being rude: Affective influenceson language use and request formulations. Journal of Personality andSocial Psychology, 76, 928–939.

Forgas, J. P. (2003). Affective influences on attitudes and moodjudgments. In R. Davidson, K. Scherer, & H. Goldsmith (Eds.), Hand-book of affective sciences (pp. 596–618).Oxford: Oxford UniversityPress.

Forgas, J., & Bower, G. (1987). Mood effects on person perception judg-ments. Journal of Personality and Social Psychology, 53, 53–60.

Georges, L. C., Wiener, R. L., & Keller, S. R. (2013). The angry juror: Sen-tencing decisions in 1st degree murder. Applied Cognitive Psychology,27, 156–166.

Goldberg, J. H., Lerner, J. S., & Tetlock, P. E. (1999). Rage and reason: Thepsychology of the intuitive prosecutor. European Journal of Social Psy-chology, 29, 781–795.

Haidt, J. (2001). The emotional dog and its rational tail: a social intuitionistapproach to moral judgment. Psychological Review, 108, 814–834. doi:10.1037/0033-295X.108.4.814

Harmon-Jones, E., Sigelman, J. D., Bohlig, A., & Harmon-Jones, C.(2003). Anger, coping and frontal cortical activity: The effects of cop-ing potential on anger- induced left frontal activity. Cognition andEmotion, 17, 1–24.

Kassin, S. M., & Garfield, D. A. (1991). Blood and guts: General and trialspecific effects of videotaped crime scenes on mock jurors. Journal ofApplied Social Psychology, 21, 1456–1472.

Keltner, D., Ellsworth, P. C., & Edwards, K. (1993). Beyond simple pessi-mism: Effects of sadness and anger on social perception. Journal of Per-sonality and Social Psychology, 64, 740–752.

Kligyte, V., Connelly, S., Thiel, C., & Devenport, L. (2014). The influenceof anger, fear, and emotion regulation on ethical decision making. HumanPerformance, 26, 297–326.

Lerner, J. S., & Keltner, D. (2000). Beyond valence: Toward a model ofemotion-specific influences on judgment and choice. Cognition and Emo-tion, 14, 473–493.

Lerner, J. S., & Tiedens, L. Z. (2006). Portrait of the angry decision maker:How appraisal tendencies shape anger’s influence on cognition. Journalof Behavioral Decision Making, 19, 115–137.

MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidencelimits for the indirect effect: Distribution of the product and resamplingmethods. Multivariate Behavioral Research, 39, 99–128.

Parker, M. T., & Isbell, L. M. (2010). How I vote depends on how I feel:The differential impact of anger and fear on political information process-ing. Psychological Science, 21, 548–550.

Paternoster, R., & Deise, J. (2011). A heavy thumb on the scale: The effectof victim impact evidence on capital decision making. Criminology, 49,129–161.

Preacher, K. J., & Hayes, A. F. (2008). Contemporary approaches to assessmediation in communication research. In A. F. Hayes, M. D. Slater, & L.B. Snyder (Eds.), The Sage sourcebook of advanced data analysis forcommunication research. Thousand Oaks, CA: Sage Publications.

Raghunathan, R., & Pham, M. T. (1999). All negative moods are notequal: Motivational influences of anxiety and sadness on decisionmaking. Organizational Behavior and Human Decision Processes,79, 56–67.

Rucker, D. D., Polifroni, M., Tetlock, P. E., & Scott, A. L. (2004). On theassignment of punishment: The impact of general-societal threat and themoderating role of severity. Personality and Social Psychology Bulletin,30, 673–684.

Salerno, J. M., & Bottoms, B. L. (2009). Emotional evidence and jurors’judgments: The promise of neuroscience for informing psychology andlaw. Behavioral Sciences and the Law, 27, 273–296.

Schwarz, N. (1990). Feelings as information: Informational and motiva-tional functions of affective states. In E. T. Higgins & R. Sorentino(Eds.), Handbook of motivation and cognition (Vol. 2, pp. 527–561).New York: The Guilford Press.

Semmler, C., & Brewer, N. (2002). Effects of mood and emotion on jurorprocessing and judgments. Behavioral Sciences and the Law, 20,423–436.

Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cognitive ap-praisals in emotion. Journal of Personality and Social Psychology,48, 813–838.

208 N. Nuñez et al.

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)

Solan, L. M., & Darley, J. M. (2001). Causation, contribution, and legalliability: An empirical study. Law and Contemporary Problems, 64,265–298.

Tetlock, P. E., Visser, P. S., Singh, R., Polifroni, M., Scott, A., Elson, S. B.,… Rescober, P. (2007). People as intuitive prosecutors: The impact ofsocial-control goals on attributions of responsibility. Journal of Experi-mental Social Psychology, 43, 195–209.

Tiedens, L. Z., & Linton, S. (2001). Judgment under emotional certainty anduncertainty: The effects of specific emotions on information processing.Journal of Personality and Social Psychology, 81, 973–988.

Wainwright v. Witt, 469 U.S. 412 (1985).Watson, D., & Clark, L. A. (1994). The PANAS-X: Manual for the positive and

negative affect schedule-Expanded Form. Iowa City: University of Iowa.Witherspoon v. Illinois, 391 U.S. 510 (1968).

Negative emotions and decision making 209

Copyright © 2015 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 29: 200–209 (2015)