Goffman on the Jury: Real Jurors’ Attention to the “Offstage” of Trials

Abstract

Social psychologist Erving Goffman, in his classic work The Presentation of Self in Everyday Life, provides a framework that explains why jurors may turn their attention at the courthouse to information not formally presented from the witness stand. We dub this “offstage observation,” a type of juror behavior that has not been systematically examined empirically. Analyzing a unique data source of 50 actual jury deliberations in civil trials, we find that jurors do look to the offstage in evaluating the claims of the parties. However, in contrast to predictions, these observations played a surprisingly minor role in the jury deliberation process.

Most empirical research on the jury comes from controlled experimental research (see especially Hastie, Penrod, & Pennington, 1983; Pennington & Hastie, 1986; Simon, Snow, & Read, 2004; for reviews of research, see Greene & Bornstein, 2003; Hans, 2000; Sunstein, Hastie, Payne, Schkade, & Viscusi, 2002). Laboratory experiments have developed a rich body of work that informs how jurors reach decisions; however, some issues cannot be addressed in the standard simulation. Even a well-designed simulation will typically present the trial only in formal ways: that is, the stimulus conveys only what would appear in a trial transcript (e.g., testimony, argument, instructions), or, if a videotaped presentation is used, only what the camera captures, typically the judge, the witness, and the attorney while he or she is questioning the witness (e.g., Diamond & Casper, 1992). Any other activity by trial participants or other individuals present during the trial is unlikely to be part of the case presentation.

As we will describe in more detail below, some commentators on the jury would consider this omission to be quite significant if one wishes to know how jurors reach their decisions. In this view, jurors are influenced not only by features of the evidence and formal case presentation, they are also affected by the information they glean from watching people who are not providing evidence or argument, especially if such people do not recognize that they are being watched. This assumption finds theoretical support in social psychological theory, especially work by Erving Goffman. Nonetheless, we have very little empirical evidence, and none of it the product of systematic analyses of jury deliberations, on jurors’ use of what we will refer to as “offstage observation” in reaching their verdicts. As a result, it has been difficult to say whether simulations that limit what can be seen in a trial do, in fact, omit significant material. The research reported here permits us to directly test this issue. We draw on a unique set of 50 real jury discussions and deliberations to systematically examine juror talk about behavior in this offstage region and thus to assess the extent to which simulations do omit crucial evidence that real jurors regularly draw on in reaching their decisions.

We analyze jurors’ comments about and responses to instances in which trial actors (the parties, the attorneys, judges, and others in court) were not engaged in the formal presentation of the trial. We find that, in one sense, offstage attention is commonplace in jury trials: the vast majority of cases included at least one offstage observation. Moreover, a substantial proportion of these observations, especially during deliberations, were “valenced,” that is, the conclusion or implication of the remark favored one party or another. Across these civil trials, the valenced remarks reflected several consistent themes. At the same time, through a close examination of deliberations, a number of indicators also reveal that offstage comments had little observable effect on a group’s decision making. We close by considering the implications of our findings for jury decision making in civil cases.

The “Offstage” at Trial

Judges tell jurors that their decisions are to be informed by the legal instructions and the evidence, which consists of testimony, exhibits, stipulations, or any fact they have been instructed to accept (see, e.g., Arizona Pattern Jury Instructions); additionally, the instructions tell jurors that the opening and closing statements, while not evidence, may help them understand the evidence. These elements, intended to guide decision making, typically take place in what, in theatrical terms, would be the “front stage” of the courtroom: the section of court from which formal case presentation occurs, including the witness stand, the judge’s bench, or the attorney’s podium.

However, in a courtroom, or even in an entire courthouse, few physical boundaries effectively wall off the front stage areas from everywhere else. Again to use theater terminology, a courtroom is more of a “theater in the round” than a venue with a traditional Proscenium stage. In a theater in the round, actors who are not, at the moment, part of the performance wait just offstage, usually in darkened portions of the stage or in the aisles. The lighting and design of a courtroom do not provide even this minimal shield. Thus, inside a courtroom, a variety of people—e.g., parties, attorneys, audience members—will be in full view of the jury while these trial actors, for example, listen to or react to others’ testimony. Further, mid-day breaks and trips through the courthouse or to nearby areas provide additional opportunities to run into courtroom actors. Although judges typically instruct jurors not to speak with any attorneys, parties, or witnesses when out in public, the classic work of Erving Goffman strongly suggests that interesting information from out-of-court encounters need not arise through oral conversations (see also DePaulo, Rosenthal, Eisenstat, Rogers, & Finkelstein, 1978; Ekman & Friesen, 1969).

In his essay, The Presentation of Self in Everyday Life, Goffman argued that social observers have a particular interest in how others act when such targets are not—or assume they are not—the central object of an audience’s attention. Goffman notes, for example, that a man engaged in a three-way conversation may “relax his usual constraints and tactful deceptions” when his other two partners are talking to one another; at this point, the man may regard himself as the “unobserved observer” (1959, p. 7). Being engaged in conversation, the two others may not notice any changes in the man’s expression; but an outside observer watching this trio may find clues to what the man is “actually feeling” toward the two others (p. 7). Indeed, Goffman suggests that social observers assume these moments are particularly diagnostic of people’s more authentic selves. If audiences suspect that a target of observation is aware of being watched and is attempting to “exploit this very possibility” (1959, p. 7)—for example, if a juror thought a defendant was feigning his reactions in the offstage areas—observers would likely look elsewhere for less controllable sources of information. Of note, observations of people who are slightly away from, but still proximate to, front stage activity are distinguishable from Goffman’s “backstage” area, which he describes as an area that is “bounded to some degree by barriers to perception” (1959, p. 106). In a true backstage (e.g., a dressing room in a theater, the kitchen area in a restaurant), targets of observation “can reliably expect that no member of the audience will intrude” (1959, p. 113) and therefore routinely believe they can stop performing. As we have suggested, courthouses have very few places in which someone can reasonably believe they are not exposed to potential scrutiny from others.

The Alleged Allure of the Offstage

Several lines of court opinion and commentary suggest that real jurors deciding cases do take notice of and are influenced by offstage activity. In criminal cases, the U.S. Supreme Court has ruled that some types of non-verbal, off-the-stand information (e.g., shackles on a defendant, armed sheriffs seated behind a defendant, a picture of a slain victim that non-testifying family members wore as a button in court) potentially prejudiced a jury (Carey v. Musladin, 2006; Estelle v. Williams, 1976; Holbrook v. Flynn, 1986). Courts have also struggled with the offstage moments of judges. Blanck, Rosenthal, and Cordell (1985) reviewed cases in which judges gave audible asides about a case (e.g., “hmmph”; State v. Larmond, 1976) or exhibited nonverbal reactions (e.g., a scowl or a smirk; Allen v. State, 1973), and appellate courts had to decide if these prejudiced the jury. In non-criminal cases, appellate courts have reviewed the opinions of lower court judges who openly write about considering various types of offstage information, including whether or not an allegedly injured plaintiff displayed signs of being in pain while listening to a hearing (Tyler v. Weinberger, 1976), the rude manner in which a party behaved during the testimony of others (Morgan v. Dep’t of Fin. & Prof’l Regulations, 2007), or some types of nonverbal displays that two people (e.g., a divorcing couple) may exhibit toward one another during a trial (Leslie v. Leslie, 1957; for a review of legal issues in non-criminal cases, see Rose & Diamond, 2009).

Laurie Levenson considered appellate rulings in criminal cases involving in-court “off-the-stand” activity, concluding that jurors scrutinize “every move” of a defendant sitting in court, “attaching deep importance to a quick glance or a passing remark – details a nonjuror might consider insignificant” (2008, p. 575). According to jury consulting manuals, scrutiny allegedly extends to attorneys. One manual not only devotes nearly 40 pages to clothing, jewelry, hair, make-up and color considerations for both male and female attorneys, but also gives advice on such details as how to behave toward other attorneys during breaks in the case (e.g., plaintiffs and criminal defense attorneys should rebuff any attempts at friendliness from the other side, while prosecutors and civil defense attorneys should try to show that there is “nothing personal” between the parties; Smith & Malandro, 1985). According to another text, careful attention to impressions includes areas beyond the courtroom doors:

You never know when a juror may be observing you. If you are going to impress the jurors, you have to impress them in every way. And for those of you who own Cadillacs, Maserattis, and Mercedes, give those cars back to your spouses on the day you go to court. Go to court in a Chevy. (Stern, 1991, p. 31).

Despite such claims, little systematic empirical data exist on offstage observation. Eisenberg, Garvey, and Wells (1998) examined data from the South Carolina section of the Capital Jury Project, which included interviews with 153 jurors from 41 capital cases. Jurors who rated the defendant as looking “bored” during trial—an assessment that depends upon actions in the offstage region—were significantly less likely to see the defendant as remorseful; perceived remorse, in turn, modestly predicted sentences, at least in cases rated as low on “viciousness.” Although suggestive, this is the lone systematic study of the topic, and its interview methodology may not fully capture jurors’ attention to the offstage, as respondents may not admit or even accurately recall what they observed in the offstage. Significantly, no prior research has studied offstage observations in the context of deliberating jurors; hence, we know little about how the group reacts to observations by individual jurors—that is, whether a comment from a single juror leads to changes in the group’s discussions and how, if at all, the jury responds to offstage remarks. Only by examining when and in what ways offstage statements come up during deliberations is it possible to assess their possible effect on verdicts and the deliberation process. As we next explain, an unusual opportunity to observe the actual deliberations of 50 civil juries allowed us to fill in gaps in existing knowledge about jurors’ use of the offstage during deliberations.

Observing Offstage Activity Through the Arizona Jury Project

As we have described in more detail elsewhere (Diamond, Vidmar, Rose, Ellis, & Murphy, 2003), the Arizona Supreme Court sanctioned a videotaping project in Pima County (Tucson is the major city) as a way to examine the effects of a jury innovation that allows jurors to discuss evidence prior to deliberation. As part of the research design, a random subset of cases was assigned to the discussion condition and jurors were told they could discuss the case (n = 37 cases), whereas discussions during trial were prohibited for the remaining trials (n = 13; see Diamond et al., 2003, for details on case assignment and comparisons across conditions). The research required elaborate permissions and security measures, and a court order ensures strict confidentiality and limits the use of the tapes (i.e., to research undertaken by the project investigators). Participation in the research was voluntary both for jurors and parties. Among jurors, 95% of those approached agreed to be involved for a total of 402 juror participants in the taped trials. Roughly half (53%) were female; 70% were white and 30% were minorities (primarily Hispanic). Fourteen percent of the sample reported annual household incomes of less than $20,000, 42% between $20,000 and $49,000, and 44% $50,000 or more.

Parties were less willing to participate: 22% approached agreed to do so. Despite this lower response, the resulting distribution of cases in the sample closely matched that for Pima County as a whole during a similar time period (see Diamond et al., 2003, for more detail). In the sample 52% (n = 26) of trials involved motor vehicle cases, 34% (n = 17) alleged a wrongful injury not involving a motor vehicle, 8% (n = 4) medical malpractice cases, and 6% (n = 3) contract cases. Cases varied from common rear-end collisions with claims of soft tissue injury to cases involving severe and permanent injury or death.

Trial testimony, arguments and instructions were videotaped via a camera focused on the witness box. When cameras malfunctioned or were not turned on, we obtained the trial transcript. From the tapes and transcripts, we created “roadmaps” of all elements of the trial, including who testified, when in the course of the trial the witness appeared, as well as a detailed summary of the content of the testimony. This allowed us to know when people were engaged in front-stage activity. To record discussion and deliberation in the jury rooms, two unobtrusive cameras were mounted at the ceiling level in opposite corners of the deliberation rooms. Unobtrusive ceiling microphones recorded the discussions. An on-site technician was instructed to tape the conversations in the jury room whenever at least two jurors were present. From the videotapes, we created quasi-transcripts of the discussions that occurred during breaks in the trial and verbatim transcripts of all deliberations. (The “quasi-transcripts” helped us to expedite a report on the effects of discussion, as reported in Diamond et al., 2003. These transcripts capture the substance of what occurred during discussions; although not fully verbatim, they are highly detailed as to speaker and content. For this paper, we examined the original tape to check the actual language in any instance in which the transcript indicated an offstage observation, but the content was not clear.) In all, we coded 2,502 pages of discussion quasi-transcripts and 5,276 pages of deliberation transcripts for the 50 trials.

Coding Remarks About Offstage Behavior

We defined an offstage comment as a reference to any behavior not occurring at the front stage of the courtroom. For parties, fact witnesses, and expert witnesses, these references reflected behavior the individual engaged in when not testifying on the witness stand. All comments about non-witness observers (e.g., family members, friends, or others who did not testify in the case) were coded as offstage remarks. For attorneys, offstage behavior was anything other than that occurring when an attorney was speaking to the jury during jury selection, arguing to the jury, or asking questions of witnesses. For judges, offstage activity reflected comments that were unrelated to judicial behavior occurring when the judge was instructing the jury or ruling on objections (e.g., the judge’s behavior while a witness was testifying). Unless the coder could be sure that a comment pertained to offstage activity, it was not coded. This rule meant we coded comments about appearance, for example clothing and attire, only when they pertained to a non-testifying person or clearly occurred on a day when a witness did not testify (per the trial roadmaps). For each comment, we tracked the physical area in which the observation took place, including inside the courtroom or in various out-of-court locations (e.g., in the courthouse but not in courtroom, such as in an elevator or bathroom, or outside of the courthouse, such as in a parking lot or nearby restaurant).

Data from a separate project that assesses the balance and extremity of positions jurors adopt during deliberation (Diamond, Rose, Murphy, & Krebel, 2007) provided a record of the valence of each comment, that is, whether the observation favored one of the parties. A remark was considered valenced if one of the parties would, and the opposing party would not, want a juror to make that remark.

Reliability of Coding

Following an initial training session on five transcripts (read independently by the first and third authors), a single coder (the third author) identified offstage remarks in the remaining 45 transcripts. To test reliability, another person then read five different transcripts, selected to include cases that did and did not have offstage remarks, as well as to vary the range in the number of remarks within a transcript. These two coders obtained perfect agreement on whether or not the transcripts contained any offstage remarks; further, of the total number of remarks the third author identified, the independent coder also identified 88% of them (missing just two remarks). Using the index described by Smith (2000) for qualitative text—twice the number of agreements on a category divided by the sum of the frequency that each rater used the category—the reliability was .86. To further assure reliability, all authors reviewed every coded remark to confirm that it was an unambiguous instance of offstage discussion.

To measure reliability of the valence coding, two other coders read three transcripts in common, which were selected to represent different types of civil claims. Each transcript contained numerous conversations turns (there are over 78,000 turns in all of the deliberations), making the index recommended by Smith (2000, discussed above) highly precise for each individual transcript. Comments valenced in favor of the defendant had indices of .82, .85, and .79 (mean agreement across transcripts = .82); the indices for comments valenced for the plaintiff were .62, .85, and .92 (mean agreement = .80). (In a few instances, a conversation turn could be coded as valenced for both parties; however, the offstage remarks within each turn were always valenced for one party or the other.) Once all remarks were coded, we summarized when the remarks occurred and whom they concerned, identified themes that ran through valenced comments, and looked in detail at how remarks were used during deliberations. We turn now to these results.

Descriptive Information

Table 1 provides a profile of the frequency of offstage remarks across different phases of the case. It bears emphasizing that the number of offstage remarks does not indicate how often offstage activity occurs during a trial or how often jurors observe it, but rather how often a juror decided to mention an observation to others. Observers may not notice some activity, they may not regard something noticed as significant, or they may not feel the information is worth sharing with others.

Table 1 Profile of offstage comments (n = 50 cases)

In the majority of trials in our sample—n = 40 cases, or 80%—jurors made at least one offstage remark. In all, we coded 303 offstage remarks about targets (when the same comment referenced two different targets, e.g., an attorney and a client, we counted this as an observation about each of the targets; 10% of observations reflected multiple targets). The majority of all remarks (n = 200, 66%) emerged during pre-deliberation discussion periods. Not depicted in the table is the fact that nearly all discussion comments came primarily from cases assigned to the Discuss condition: Just 6 of the 13 “No Discuss” cases produced any offstage remarks during discussion periods (n = 28 comments); of these, only two remarks (7%) were valenced. By contrast, in the entire sample, 49% of all remarks (n = 148) were valenced.

As Table 1 shows, the majority of comments referred to behavior that occurred in the courtroom (74%) rather than in other locations (i.e., in hallways, elevators, or outside the courthouse; 26%). This pattern is consistent with the fact that jurors spend more of their day inside the courtroom than out. Moreover, the opportunity to see people outside of court is largely a product of chance, and jurors also receive a special instruction about out-of-court contact (i.e., it is to be avoided if possible), which may reduce their inclination to share an out-of-court observation with others.

The profile of offstage comments differed in several ways across discussion and deliberation periods. Ninety percent of comments made during deliberations concerned in-court rather than out-of-court observation, whereas approximately 70% of comments from pre-deliberation and post-deliberation periods were from in court (see Table 1), χ ²(2, N = 303) = 12.73, p < .01. Table 1 also shows that comments were far more likely to be valenced during deliberations than during other periods, χ ²(2, N = 303) = 77.80, p < .0001 (in addition to the significant overall χ ², deliberation differs individually from each of the other periods at p < .0001). Parties were also far more likely to be the target of offstage remarks during deliberations (65% of remarks were about parties) than during discussions (31%) χ ²(1, N = 272) = 26.75, p < .0001 (deliberations and post-deliberations do not differ, p = .69). Further, among valenced comments, deliberation remarks were more likely to favor the defense rather than the plaintiff compared to discussions, χ ²(1, N = 132) = 7.10, p < .01, or post-deliberation periods, χ ²(1, N = 83) = 5.72, p < .05. As the bottom row of Table 1 indicates, when comments concerned the parties, deliberation remarks were particularly likely to be valenced for the defense, χ ²(2, N = 97) = 9.60, p < .01 (deliberation differs individually from each of the other periods by at least p < .05).

We looked for patterns that might distinguish jurors who offered offstage remarks from those who did not; according to t-tests, no significant demographic factors emerged. During deliberations (which had the greatest proportion of valenced comments), remarks were as likely to be offered by men as women (p = .71), whites as non-whites (p = .91), professionals and non-professionals (p = .76), and older versus younger jurors (p = .23). Talk time was the sole distinguishing feature: People who offered offstage remarks also tended to speak a greater percentage of total words during deliberation (16%) than those who did not (12%), t(399) = 2.75 (p < .01, d = .28). As a reviewer noted, offstage observation may also vary with case-level factors. In particular, jurors may be more likely to look to the offstage when front-stage evidence is more ambiguous. Using strength of evidence ratings obtained from the judge in each case, we tested this possibility but found no significant differences between the strength of evidence ratings across (a) cases with and without any offstage observations, t(48) = 1.49 (p = .14, d = .43); (b) cases with or without valenced observations, t(48) = 1.76 (p = .08, d = .51); or (c) cases with and without valenced comments during active deliberations, t(48) = 0.52 (p = .60, d = .15).

In sum, offstage remarks were commonplace across cases, with 80% of the cases in our sample generating at least one comment, more often arising from in-court observation than from chance out-of-court encounters. Just under half of all of these remarks had valenced implications for one of the parties. At the same time, talk about the offstage was sparse: We uncovered just over 300 comments, only 72 of which came from deliberations. The nature of offstage remarks also changed across the decision-making context (mid-trial discussions versus deliberations), with deliberation remarks more likely to be valenced and to be disproportionately critical of the plaintiff’s case. The next section describes the themes that ran through valenced¹ observations offered either during discussions or deliberations; following that, we consider more closely the role of valenced comments during the deliberations as the juries reached their verdicts.

The Substance of Valenced Offstage Remarks

Valenced offstage remarks fell into three general categories. Two independent coders rating the deliberation remarks exhibited high levels of agreement when placing remarks into each of these categories (kappa = .91.) (In describing the theme of offstage remarks, we offer some representative quotations. Note that within the longer quotations, parenthetical remarks refer to descriptive information a transcriber included, e.g., to note interruptions or non-verbal behavior; brackets indicate material we have added for clarification or instances in which we edited material to preserve case confidentiality. Ellipses (…) indicate deleted material.)

Plaintiffs’ Level of Functioning

The offstage offers jurors an opportunity to observe the current health of the plaintiff. We found that comments about a plaintiff’s level of injury occurred across several different trials (n = 35 total comments in 18 trials), many emerging in cases in which the state of the plaintiff’s current physical or emotional state was a question of fact for the jury to decide. Specifically, there were 33 cases in which plaintiffs claimed that an injury was ongoing and therefore sought future damages; 16 of these cases (48%) had an offstage observation on this point. (Two additional cases also produced remarks about plaintiff’s health, even though the plaintiff did not claim ongoing injuries. In both cases the comments were brief asides during deliberations that the group did not take up further.)

When commenting on the plaintiff’s state of health, a few remarks offered general appraisals. Jurors variously noted that they were “surprised” that the plaintiff “walked pretty well” [#046]; that the plaintiff “walks around real good” [#001]; “has been able to get around” [#013]; and that, despite a back injury, the plaintiff “walked off the elevator just fine” [#027]. Other remarks talked about more specific facets of a plaintiff’s alleged injury. For example, when a plaintiff claimed to have problems lifting one of his arms, a juror in the case remarked the he had seen the plaintiff put his arm around his wife [#042]; in another, a juror said that the plaintiff claimed to be in pain but yet never grimaced during the trial [#040]. Comments concerned mental as well as physical health. One juror expressed disbelief that the plaintiff continued to suffer emotionally (as claimed in the case) because the plaintiff never cried when away from the witness stand; other jurors in the case, however, noted that the plaintiff cried “constantly” [#041].

A plaintiff’s ability to sit for long periods of time without showing pain caught the notice of three juries, although notably in two of these cases, the plaintiff or counsel invited that scrutiny and critique by saying that the alleged injury required that the plaintiff get up and walk around during trial. In both instances, jurors pointed out that the plaintiff never got up. In one case, a juror noted that the plaintiff “sat like a rock” and “never got out of his chair” [#029], a fact the juror carefully documented by taking notes on the plaintiff’s seating patterns: “Starting on Tuesday, I can tell you how long that man sat in his chair” [#029]. Other jurors raised additional questions about this plaintiff’s credibility based on other offstage behavior. One juror reported on being in the restroom when the plaintiff entered, and

• #1: … he is walking normal. And, we’re both right there and all of a sudden, I realize by the body reaction, he kind of looks at me and kind of like, he recognized me. So then I started washing my hands and all of a sudden, he put on an act and started, (the juror imitates someone limping), okay, I’m sorry, I have a problem with … (juror is interrupted).

A different juror confirmed that the plaintiff failed to limp when walking across the street outside of the courthouse, and later one person noted:

• #4: And here’s my Dad, and he’s older, granted, but it’s twelve years ago he had the back surgery and here’s my Dad getting out of a chair (she stands up slowly and groans).. … And here’s this guy [plaintiff] getting out of the chair (#4 stands up spryly and walks away).

Not all comments about a plaintiff’s physical condition reflected an absence of symptoms. Some jurors commented that the plaintiff engaged in offstage behavior that seemed designed to amplify an injury during the trial. For example, one juror claimed that on the last day of trial, the plaintiff “evidently, obviously” failed to take pain medication, and another juror on that case confirmed that the plaintiff’s eyes “seemed a lot clearer today” [#027].

Responses to the Trial Event

A second set of valenced comments discussed how people in the courtroom responded to trial events, how they responded to the fact of being involved in a trial, or the decorum they displayed during trial (n = 86 comments from 23 cases). These remarks took several different forms. Jurors in three different cases wondered whether the trial was putting stress on one of the parties [#028, #047, #048]. Jurors in three other cases claimed to have observed coaching of witnesses from offstage areas, in particular, through hand signals from an offstage attorney [#041] or because a plaintiff’s spouse allegedly mouthed answers [#031] or sent other signals [#030] from offstage.

Targets’ reactions to specific events struck a few jurors as highly notable. For example, upon hearing an expert testify that the plaintiff’s weight may have contributed to the plaintiff’s disability, a juror said he deliberately looked over to see the plaintiff’s response [#013]. In another case, two jurors saw the plaintiff cry after a key ruling by the judge undercut her case [#048]. The plaintiff, a person of modest means (facing a wealthy defendant), had emphasized during testimony that she and her co-plaintiff were pursuing the case largely on principle (i.e., not for money). In the jury’s view, the strongly emotional response to a setback in the case undermined the claimed dispassionate, noble motivation.

Jurors remarked, often critically, on people whose offstage reactions appeared aimed at sending signals to the jury. For example, a juror said she covered her face when an attorney looked over at the jury and rolled his eyes “like he was going to share an in-joke with me” [#013]. In another case, jurors offered a range of views about the wisdom (and intentionality) of an attorney’s expressions and reactions, with some deeming it as just part of the attorney’s job:

• #4: The [plaintiff’s] lawyer [gives name], she was making faces.

• #1: Watch her make faces when [the defendant] is talking.

• #4: She’s like a little kid. I like watching it.

• #9: But that’s not smart. They don’t realize we are watching them.

• #4: It was adorable. She’s putting on a show.

• #1: She even looked at me and rolled her eyes saying this guy is totally lying out his butt.. … She isn’t as obvious [as the defendant, who responded to others’ testimony by appearing disgusted (e.g., throwing a piece of paper onto the table) and mouthing words such as, “Liar” and “Fuck you.”]. She just rolls her eyes.

• #9: Yeah but you think about young school children who roll their eyes, so you think about the level of maturity she’s at.

• #4: She’s sending messages.

• #5: It’s all part of the tactics, being a lawyer [#010].

Although this last juror seemed resigned to the “show” from attorneys, jurors generally were less philosophical about, and more critical of, dramatic displays from other actors. Jurors in five separate cases commented negatively about defendants who smirked or seemed disdainful. For instance, during discussions a juror in one case deemed a defendant, the owner of a business where an accident occurred, to be “smug” based on how he responded to the testimony of others; another agreed:

• #7: Yeah, he’s very smug and he was laughing and shaking his head and going ‘How are you the expert’? .…

• #8: And he’s supposed to make a good impression.

• #7: That’s the problem, he’s not supposed to make an impression and the worst part of it is he reminds me of an old boss I had. I can’t think about that.

• #8: That’s your civic duty to put that away and just look at the facts.

• #7: You know, there he is just sitting there and I think that says something about him and something about how he runs his business. Maybe it doesn’t mean a lot but it means something. [#042]

Jurors in two cases [#004, #013] likewise remarked on some plaintiffs’ tendency to smile at the jury, with one juror saying he thought to himself, “Okay, stop that, I know you’re told to do that” [#013].

Apart from theatrics, jurors were also critical about targets who were inattentive to their appearance—for example, on one day of the trial a plaintiff wore a top that revealed too much of her cleavage [#030], and a defendant in another case mindlessly picked at calluses [#037]. A few jurors criticized people who joined them on elevators, suggesting that the parties or the attorneys should have waited for a different one [#012, #033]. Only a handful of offstage remarks were complimentary, with jurors praising a plaintiff [#017] or one of the attorneys [#001, #016, #039] for remaining composed and calm while listening to testimony.

Additional Details

In a final set of observations (n = 28 comments from 13 cases), offstage remarks contributed alleged background facts and additional details about targets, without remarking specifically on injuries or reactions to events at trial. For example, one juror said of an elderly defendant being sued in an auto accident case, “I hope he doesn’t drive anymore,” to which another juror reported having seen him walk out of the parking lot [#012]. Jurors in another case found it noteworthy that the plaintiff’s own family did not sit on the plaintiff’s “side” of the courtroom and also pointed out that the plaintiff’s daughter was seen talking to the defense attorney [#029]. One juror on a motor vehicle case constructed a small background story about the plaintiff:

• #5: Well, I don’t know if, how you guys felt about it, but when I seen her husband, it’s like she intimidated by him. She seems like a very shy, very intimidated little girl and he seemed like he was rough and I think she might have had an anxiety attack over the fact that, ‘wait till he finds out’ she had an-

• #2: Accident.

• #5: -accident. Maybe, I don’t know, maybe she’s not supposed to be running around … and she was out late at night (other jurors laugh). [#043]

Other remarks reflected simpler conclusions, with jurors using an offstage observation to note, for example, that a plaintiff was “cold” [#030], a witness’s family seemed “nice” [#023], or that, given the way a person walked up to the stand, a juror thought, “this man is a shyster” [#030].

The Offstage as Part of Deliberations

The previous section documented the substance of what individual jurors reported to others about offstage observations. Nonetheless, valenced offstage remarks—even those highly relevant to the issues being decided (e.g., the plaintiff’s current state of health)—may or may not influence a group’s final verdict. The jury may actively dispute the validity or relevance of the observation, or they may simply ignore it. In this section, we consider the role offstage observations played in deliberations about final verdicts.

In examining deliberations, it is important to distinguish what we can determine from these data from what we might wish, ideally, to know about decision making. Quite clearly, the Arizona data do not allow us to assess the direction of causal influence of some factor or exactly how an individual juror arrived at a conclusion (e.g., whether an offstage observation confirmed an already-emerging position or whether the same observation caused the juror to change his or her impression of a case). Instead, we know what jurors report to others regarding their beliefs about the case. Given interruptions and shifts in conversation, jurors’ reports to others can sometimes be disjointed and challenging for observers to pull together into a coherent whole (even if the position is coherent in the juror’s own mind; see Simon et al., 2004). Additionally, jurors themselves may not have access to the bases for their own verdict preferences (Nisbett & Wilson, 1977) and therefore may not be describing their decision processes accurately to others. These same limitations apply to knowing whether, once offered to the group, offstage comments influenced fellow jurors. Some individual jurors may have been influenced by the remark but may not have known they were affected, and others may have been persuaded but opted not to announce that fact.

At the same time, deliberation talk provides several indicators of how the group treats offstage observations. In particular, to the extent that juries attach special significance to offstage conduct, then we would expect them to devote a nontrivial amount of deliberation time to offstage discussion. Second, if influential on the group, then offstage remarks favoring one party should be significantly associated with verdicts favorable to that party. Finally, offstage remarks could have a less direct influence on deliberations if they provide information that alters the trajectory and tenor of conversation. By tracking the consistency between the valence of an offstage comment and remarks a juror has previously shared with the group, we can assess whether offstage observations seem merely to offer additional support for a juror’s previously expressed position or whether the comment presents a unique form alternative viewpoint. If the latter, the remark should also be linked to subsequent discussion topics or decisions the group goes on to make.

How Much Deliberation Time Do Jurors Spend on Offstage Topics?

For a remark to have an observable effect on deliberations, it must be mentioned during periods of active deliberation. By this standard, the vast majority of remarks from mid-trial discussions had no observable role in deliberations: Only 4 of the 64 valenced remarks that occurred during pre-deliberations discussions (from four separate cases) were mentioned again during deliberations. Although the remaining 60 valenced offstage observations from discussions may have remained on an individual juror’s mind, this did not translate into a reiteration of the comment during deliberations. In a similar vein, five deliberation remarks from two different cases occurred just prior to the start of formal deliberation, when all jurors were not-yet present, and the group was passing the time while waiting on the others to appear. None of these remarks were repeated or referred to again when the group formally began discussing the case. (Table 1 also shows that post-verdict discussions also contained a small number of valenced remarks; by definition, these remarks could not have shaped the way the group arrived at an outcome.)

During active deliberations there were a total of just 56 valenced offstage remarks from 21 separate cases (counting observations that included two targets at the same time, there were 62 remarks). To gauge time spent on offstage talk during active deliberations, we examined all 56 remarks and totaled the conversation turns devoted to presenting the information, including any responses that built upon the observation and therefore continued the same topic. Across cases, the number of turns devoted to offstage activity ranged from a low of 4 turns to a high of 139. The latter figure came from an outlier jury—which we will discuss in more detail below—in which the jury’s opportunity to view people in the offstage figured into damage calculations; without this case, the highest value was 48 turns. As context, the outlier case with 139 offstage turns had 6,110 total deliberation turns, meaning that offstage talked consumed just 2.3% of total turns. Across all cases with offstage remarks, the mean percent of total turns devoted to offstage discussion was just 1.5% (SD = 1.5%, range 0.3–6.4%). Of note, these turn counts included instances in which the jurors agreed with and supported the offstage remarks, as well as instances in which others disputed the relevance or legitimacy of the offstage. In the 21 cases with offstage remarks, there were 9 in which at least one juror explicitly disputed either the fact of the observation or its relevance to their discussions.

On the turn-counting measure, the case with the highest percentage of turns devoted to the offstage (6.4%) deserves mention. This is the group in which one of the jurors took notes on how long the plaintiff sat in his chair (after his attorney and a witness had said he could not sit for long periods of time), and jurors subsequently offered a series of negatively valenced observations about the plaintiff, including (as mentioned previously) that he was not limping when in bathroom and outside court (but seems to start limping when juror is seen); that he got up easily from chair; his own children sat on the “other side” of the court from plaintiff; and, further, one of these children was seen talking with the defense attorney. During the entirety of these discussions, there was only tacit “cross examination” of the information, with one juror suggesting, “if he does have pain, that may be the only comfortable way he can sit… [he] may not want to move around. …” The juror who made the original observation (#4) interrupted to assert that her offstage observations were less speculative than her fellow juror’s theory: “The testimony was that he had to get up, that was the testimony. I’m going by testimony not by how he feels” [#029].

Such remarks and the pattern of comments indicate that this jury viewed the offstage as a legitimate source of information that bears on what they must decide. But if these observations contributed measurably to outcomes in this or any other case, the tenor of offstage remarks should predict verdicts, and the offstage information should provide the group with unique insights on the case. We next consider evidence for these effects.

The Offstage and Verdicts

As we noted in our profile of offstage remarks, offstage comments during deliberation disproportionately referenced the plaintiff as a target, and the valence of remarks disproportionately favored the defense. Given this, one might expect anti-plaintiff comments to be clustered in cases that plaintiffs lost. However, this was not the case. Among the 19 cases with remarks valenced against the plaintiff or the plaintiff’s associates (e.g., witnesses, attorneys), just 4 (21%) resulted in a defense verdict (defense verdicts occurred in 35% of the cases in this sample). A defense verdict is a conservative measure of “loss,” however, because some cases are uncontested on liability (and therefore must result in a plaintiff award) and because even when plaintiffs “win,” their award may be so small (e.g., <$3,000) that it would probably not even cover litigation costs. If we count as a “loss” any outright defense verdict on liability, any instance of a minimal award (<$3,000), or any instance in which liability was uncontested and the jury awarded the amount the defense conceded as reasonable, then 8 of the cases that had remarks valenced against the plaintiff constituted losses (42%). Under this same definition, the base rate of a plaintiff loss in the entire dataset was 46%. Indeed, cases with defense verdicts had fewer anti-plaintiff offstage remarks (M = 0.61) compared to other cases (M = 1.33, t(48) = 1.47, p = .15).

We found only one instance in which we could link offstage observations directly to decision making and a final verdict. During the course of calculating damages, one jury [#048] noted that the plaintiffs were present each day in court, no matter whether they were testifying or not—as one juror noted, “to please the jury.” This jury reasoned that this time was an additional burden plaintiffs had to bear as a result of the alleged incident and that the plaintiffs should be compensated for the time they were “required” to be present and on display. There was conflict over this position, with one juror arguing that the plaintiffs made a choice to be in court on days they were not testifying:

• #7: If they, if they needed money, they wouldn’t have sat in that courtroom. They would have told that lawyer, ‘Look, we’re paying you. You do the job. I know it looks good, but we got to make money.’

But several others felt that the plaintiffs had little option but to attend, even if they had jobs or other responsibilities during that time. One of these jurors defended this position by pointing specifically to the offstage information the jury gleaned:

• #1: … I mean, because they were there, we got to sit there, look at them, see whether they’re smiling or laughing or crying or . … whatever it was. [#048]

The jury ultimately awarded a few hundred dollars a day for the plaintiffs’ time in court, a small increment to the overall award. This was a singular event. No other jury took the position that offstage time was compensable as damages.

Offstage Remarks as Unique and Decisive Information

Because many aspects of the trial are likely to influence outcomes, the verdict represents a crude measure. The context of comments and the course of discussion following a remark provide a more sensitive indicator of the influence of offstage observations. Consider, for example, the following exchange:

• #6: They’re looking up to six months but I think she’s looking up to right now, ‘cause she’s still in pain, don’t you think?

• #5: Uh…

• #2: Now, I’m not doubting that she’s in pain, but-

• #7: (interrupts) You know if she was in pain she shouldn’t be sitting down in court.

• #4: I sure, I [am] sure that she wasn’t in pain. She said she couldn’t sit for more than an hour. Well she sat for more than an hour.

• #7: She sat more than an hour, in one position.

• #4: And then um…

• #2: I, I moved six times. (Laughter).

• #7: I, I kept moving.

• #4a: She sat the same way the whole time. She was just sitting there. [#016]

In this instance, a remark’s influence may have been subtle, operating to cut off a potentially sympathetic view that called for a higher damage award—in this case a longer period of injury (“…she’s still in pain, don’t you think?”). If so, then prior to the comment, there should be indications that the juror making the remark was open to a more sympathetic view. Instead, for the above case, the jurors’ stated positions prior to the offstage remark were highly consistent with the tenor of the remarks, indicating that the offstage remarks reflected a firm position on the case (at least those stated to others). In particular, three jurors commented critically on how the plaintiff sat in her chair (jurors 4, 4a, and 7). Before the offstage discussion occurred, juror 7 had not made a single pro-plaintiff remark, but had made 35 pro-defense remarks; juror 4 had made 32 pro-defense remarks and just 2 comments favoring the plaintiff. Juror 4a was the exception: By the time of the offstage comment, juror 4a’s valenced remarks were evenly split for the plaintiff and defense; however, this juror had made only 4 previous valenced comments. (Juror 2, who added commentary to the observation, but did not report the experience independently, resembled jurors 4 and 7, having previously given 20 pro-defense comments and 4 pro-plaintiff comments.) In other words, before the offstage observations emerged, 95% of all valenced comments previously expressed by all of the offstage observers favored the defense (35 + 32 + 2 = 69) of (35 + 34 + 4 = 73). The average percentage of each individual’s pro-defense proportion was 61% (i.e., the average of 100, 94, and 50%, with the lower value reflecting the influence of juror 4a, who was more balanced but also offered few valenced points). These percentages suggest that the offstage remarks reflected a perspective that was unlikely, for a variety of reasons, to accept a sympathetic view of the plaintiff.

How does this profile compare to the other cases that included at least one offstage remark that was valenced against the plaintiff’s case? We focus here on valenced patterns preceding an anti-plaintiff offstage remark because we have so few cases in which offstage remarks disfavored the defendant during deliberations. (Even among the five separate cases in which remarks about the defendant or the defendant’s associates favored the plaintiff, all but one [#025] also had remarks about the plaintiff that favored the defense.) By either measure (the proportion of all valenced comments expressed by jurors making offstage remarks, or the average of each juror’s pro-defense proportions), fully 77% of valenced comments expressed prior to an anti-plaintiff offstage observation favored the defendant. (Note that the average of all jurors’ pro-defense remarks includes only jurors who have spoken before. In three cases [#029, #035, #040], the offstage remarks of all or some of the jurors were the first valenced comments the jurors made, e.g., said while the juror summarized his/her initial impression of the case. None of these offstage remarks was discussed further by the group. If we include these zero values in the average of juror’s pro-defense proportions, then two-thirds of remarks, 67%, were pro-defense.) Whatever measure used, the results offer strong evidence that during deliberations, offstage remarks tended to shore up positions jurors had already expressed. Indeed, this seems to be the case for the jury we cited previously [#029], in which several jurors contributed multiple offstage remarks (totaling 6.4% of deliberation time). Among the jurors who had made at least one valenced comment before the offstage remark, 90% of those comments were valenced against the plaintiff.

There was certainly variability in the 77% figure; in particular, six cases were below this mean value, i.e., remarks preceding offstage observations in these cases were more balanced. These below-mean cases are those most likely to represent instances in which an unflattering offstage comments about a plaintiff (or someone associated with plaintiff’s side) had the potential to shift the direction of deliberation, as the stance of the juror making the remark had previously been more ambiguous. For these cases, a close look at the timing of the remark and how the sentiment of the remark may have been reflected in subsequent conclusions the juries reached uncovers no evidence that the offstage comment had a detectable effect on the group’s deliberation.

For example, in one [#048, in which 36% of offstage observers’ remarks favored the defense], a juror (#1) responded to an anti-plaintiff offstage remark—the plaintiff was seen crying after a judge’s ruling, which undermined her claim that she was not pursuing the case for money—by suggesting the jury would not have believed the plaintiff’s testimony in any case: “Everybody, nobody’s naïve enough here to think that they’re here because, you know, they wanted to David and Goliath somebody. They were thinking three-story house with the pool” [#048]. Further, this is the same jury that opted to slightly supplement the plaintiffs’ damage award on the principle that the parties were required to be available during trial for offstage observation—hardly a jury that can be characterized as using the offstage as a means to undermine the plaintiff.

In two other cases with somewhat more balanced valenced remarks [#013 and #027], a remark about the plaintiff’s health occurred after the jury had already decided not to award any future damages, rendering any question about the observation’s influence moot. The context in another case also suggested a remark had little impact. During a lull, while jurors crafted a question for the judge, the group joked about having seen the plaintiff smile confidently at the jury—only to have one juror remark, “Oh, God, I feel bad for you.” Consistent with the balance in previous valenced remarks (57%), the group also made critical offstage remarks about the defendants (they rolled their eyes during others’ testimony and one had a bad hair piece [#004]). More significantly, none of these remarks—and no other offstage comment—was part of the jurors’ talk after the group submitted its question and returned to formal deliberations.

In two other cases, the most generous interpretation of the context suggests that remarks may have temporarily shifted the positions of one or more jurors. In one, jurors were setting up a damages discussion by generating a lengthy list of the problems the plaintiff claimed still affected her:

• #5: Her hip, when, she said, she walked.

• #8: Yeah?

• #5: She said when she was trying to walk…

• #1: (interrupts) She wasn’t limping. That’s why I was thinking, ‘cause usually when people have difficulty walking they….

• #3: (interrupting) You can see.

• #1: Yeah, so, question mark.

• #8: I’ll put ‘minor’ beside ‘hip’. [#023]

Although it appears that the offstage observation changed one element of the plaintiff’s claim about her injuries to “minor,” the list itself proved largely inconsequential because the jury opted not to award future damages. Likewise, in a damage discussion in another case [#042], individuals were going around the table and offering their preferred damage amounts. Nearly all rejected the plaintiff’s requested amount (offered during closing arguments), typically saying that half that amount seemed more appropriate. When one juror (juror #5) hesitated (saying that he was still thinking), another prompted him by asking, “How much does it mean to you to put your arm around your wife?” Other jurors laughed and went on:

• #7: His wife is just going to have to sit on his left side from now on.

• #2: This is a slight bit of information, but I was watching [the plaintiff], he favors his right hand when he does things, he drinks with his right hand,

• #7: Yeah.

• #2: which I thought was odd if he has had [this injury].

• #6: I thought I’d seen him with his wife and his arm around her, with his right hand … [His disability] is not going to do that much around here (juror demonstrates putting his right arm around a make-believe partner). It may do a little up here (juror raises his arm over his head).

• #2: I don’t know if he’s left-handed or right-handed.

• #4: I always wash my hair with one hand anyway.

• #7: He’s left-handed.

• #8: He’s left-handed because [explains comment from testimony].

• #7: They asked him if he could [do that].

• #8: So, [juror #5], what do you think would be reasonable, instead of [what the attorney requested]? [#042]

Juror #5 went on to offer an even lower amount than others had endorsed. Perhaps the discussion about the seemingly trivial impact of the injury provided juror #5 with a reason to go lower than his fellow jurors. If so, the effect was short lived; the juror later agreed to go along with the amount others had suggested.

In these latter two cases, in which offstage remarks resulted in, at best, a temporary shift in a position, it is important to reiterate that we do not observe any potential unstated influence on individuals. In case #042, the group ultimately settled on the amount of reimbursement most favored before any offstage talk began; thus, there is no indication that the offstage remarks ultimately altered the group’s final outcome. At the same time, the comments may explain, in part, why the jurors making the observations had already offered amounts that were below what the plaintiff requested for future damages. In case #023, it is conceivable that the absence of a limp when the plaintiff walked helps explain why people did not want to award future damages. As we have said, no observable evidence for these possibilities exists in these data. What was clear, however, was that no juror used or reiterated an offstage comment as an explicit argument in favor of his or her preferred award. Instead, offstage comments were used primarily to bolster opinions with grounding in other sources. As with the other cases in this dataset, we thus find little evidence that the offstage observations provided the group with a unique type of information that changed the course of deliberations.

Discussion

Consistent with social psychological theorists such as Erving Goffman, it is reasonable to expect jurors at trial to keep a close watch on the moves and expressions of people who are out of their “front-stage” roles. This prediction is consistent with the trial strategy advice of those who suggest that various forms of “offstage” observation can make or break the credibility and appeal of a party. If true, it suggests a major limitation in the way psychologists and others typically study the jury, since most experiments present only the formal testimony and argument a jury would hear in a trial. Yet whether jurors attend to areas away from the witness stand and the front sections of the courtroom has not been adequately considered in previous research. Through a unique look at the actual deliberations and discussions of 50 real civil juries we find that although individual jurors clearly talked about information observed in the offstage, claims about the effects of such juror observations on trial outcomes are probably overstated.

A majority (80%) of cases included at least one offstage remark, and nearly one-half of all these comments had positive or negative implications for one side or the other, particularly if offered during active deliberations as opposed to pre-trial discussions. The valenced remarks sometimes went to the heart of issues before the jury, such as whether a plaintiff remained injured or whether someone was credible or not. Jurors sometimes deliberately sought out information on how someone reacted to trial events, and they also offered commentary on weakness in presentation (e.g., picking at calluses). In a few instances, jurors used observations to offer background knowledge about people—for example, that the defendant might still be driving at an advanced age—or they offered brief character sketches of people (someone seems like a “shyster” or “nice”).

Valenced offstage observations during deliberations disproportionately discussed the plaintiff and were, more often than not, unflattering commentary on the plaintiff’s case—for example, jurors did not offer an observation such as, “You can see from how he walks that he is still hurting a great deal.” This result is consistent with the findings from interview data (e.g., Hans, 2000), which revealed that jurors are quite suspicious of plaintiffs. Precisely why jurors are so critical of plaintiffs merits further investigation. It is worth noting, however, that valenced observations about the defendant—although far less frequent—also typically highlighted flaws in presentation and therefore supported the other side, suggesting an even more broadly critical stance the jurors take toward trial participants, particularly the parties. Jurors typically dismissed what they saw as attempts by anyone to perform for the jury through displays of strong emotion or back-channel comments about witness’s testimony—as one juror noted about a defendant: “He’s not supposed to make an impression.” This awareness of performance and jurors’ critical stance toward the case offers a picture that is quite at odds with one presented by detractors of the civil jury (for a review, see Diamond, 2003). Far from being gullible dupes, these jurors showed that they were well aware that they were viewing an adversarial trial, in which “spin” and some amount of theater were to be expected.

Although we found that it was routine for jurors to supplement what they heard from the stand by searching for spontaneous and unguarded sources of information, our close analysis of deliberations reveals that these offstage sources played a subordinate role in the group’s decision-making process. Most basically, across 5,000 pages of transcript material, we uncovered just over 70 remarks about targets from deliberation periods. Counting generously, just 1.5% of deliberation turns were devoted to offstage activity—hardly indicative of groups focused on the missteps, contradictions, and potential clues they find from offstage regions. Even the jury that showed the most vigorous interest in the offstage—taking notes, and commenting on a plaintiff’s failure to limp when outside of court, among other things—devoted only 6.4% of their time to this discussion.

We did find that one jury attached enough significance to the region that they opted to award the plaintiffs a small amount of money to compensate the plaintiffs for the time they spent in court during the trial. The jurors assumed that the plaintiffs were in court because juries expect parties to be present and available for observation in the offstage, thus awarding money largely on principle, rather than because they identified a courtroom signal that the plaintiffs needed compensation for the time in court (indeed, other offstage remarks criticized the plaintiff’s claims to not be motivated by money). We could find no other instance that so directly traced a decision on liability or a plaintiff’s award to some perceived failure in the offstage. Indeed, the tenor of offstage comments was strongly foreshadowed by and largely consistent with previous remarks about front-stage evidence. On average, 77% of valenced comments from people who made anti-plaintiff offstage observations also expressed favorable positions toward the defense on other matters. Among the cases with less than 77% valenced comments, that is, those that expressed more balance in their positions, we could find no instance in which the trajectory of the deliberation changed following offstage information nor were offstage remarks reliably linked to significant decisions from the group (e.g., whether to award future damages).

According to other research, all decision makers, including jurors, attempt to create a coherent explanation of events. Whether this effort arises as part of the construction of a story that explains the events that led to the trial (Hastie et al., 1983; Pennington & Hastie, 1986), or comes about from an attempt to make the available pieces of evidence fit with an emerging decision (Simon, 2000; Simon et al., 2004), there is a tendency to create congruence among perceived elements of evidence. For that reason, we would expect jurors to call on offstage behaviors that are consistent with the material they find persuasive from the front-stage region of the trial (Simon, 2000). Our review of the conversations among the jurors does not indicate whether the Arizona jurors also observed inconsistent offstage signals that they simply did not report in their conversations. Nor do we know whether such inconsistent offstage signals simply did not occur. Two kinds of future studies are needed to directly test these possibilities. Experimenters might manipulate offstage signals that are consistent or inconsistent with front-stage evidence and test the impact of such occurrences during simulated deliberations. In addition, trained observers might systematically record any offstage behavior that occurs in the courtroom during actual trials to see whether courtrooms generally present a mixture of offstage signals. Ahead of such studies the current research shows, for the first time, that although jurors talk about notable offstage behavior, the remarks play a highly subordinate role in deliberations.

Finally, because many criminal defendants do not testify, conceivably jurors in criminal trials might focus on and rely more heavily upon observations they are able to gather from the offstage. This question awaits further analysis based on data from criminal cases. With respect to the types of civil trials these real jurors considered, however, there is little evidence that verdicts are the product of offstage data collection. Rather, we find that jurors keep the focus of discussions and deliberations squarely on the front stage of a trial.