Jury Deliberation

Jury deliberation begins when a trial ends and the jury moves to a secluded location to discuss the evidence and arrive at a decision. Understanding how juries reach their decisions is a prerequisite for answering the question of how well they serve their function in a democratic society. Jury deliberation has been studied empirically by social science researchers for more than 50 years now, but direct access to the jury room has been always greatly limited due to a concern that any “external” presence in the deliberation room could influence the jury’s decision. As a result, researchers have relied on two other methodologies to study jury deliberation: experimental studies with mock juries and posttrial reconstructions of actual deliberation via surveys of and/or interviews with former jurors.

This entry summarizes what has been learned about jury deliberation using these two methodologies. The picture that emerges is one in which juries take their task seriously and generally do a good job of reviewing the evidence, although they often struggle with their instructions. Influence in juries is a function of information exchange and the pressure to conform. An excellent predictor of the jury’s final verdict turns out to be the distribution of verdict preferences on the jury’s first vote; the verdict favored by the initial majority ends up being the jury’s final verdict about 90% of the time. However, the initial majority does not always prevail, and these “reversals” represent some of the most interesting products of jury decision making.

A General Model of Jury Deliberation

A simple stage model provides a reasonable framework for thinking about jury behavior in many, if not most, deliberations. In the first stage, jurors get oriented toward each other and their task. They settle in, introduce themselves, select a foreperson, and discuss how they will do things. In the second stage, which takes up the bulk of the time spent in deliberation, jurors discuss the evidence, take opinion polls (or “votes”), and confront disagreement among members. During this stage, conflict often surfaces, efforts are made to persuade other jurors to change their minds, and the jury moves toward a consensus. In the final stage, the jury achieves sufficient agreement (usually unanimity) to reach a verdict, followed by attempts to smooth ruffled feathers, reconcile individuals to the group’s decision, and help everyone feel good about the collective verdict. Jury deliberations vary considerably in length, but most last somewhere between 2 and 4 hours.

The Foreperson

Juries typically receive little instruction regarding how they ought to deliberate, but one thing they are all told to do is choose a foreperson. Forepersons are usually selected early, often immediately after members assemble in the jury room but occasionally not until the end of deliberation. Most of the time, the selection process is brief and even perfunctory, particularly in criminal juries. Forepersons tend to be White, male, better educated, seated at the end of the table, and the first to speak or the first to identify the need to select a foreperson. This profile suggests that juries tend to eschew confrontation and rely on stereotypes and subtle interpersonal cues (such as seating and order of speaking) to identify a “natural leader.” Only two task-relevant characteristics have been found to be associated with forepersons: They tend to have previous experience as jurors, and they sometimes have a relevant occupational background in civil settings (e.g., an accountant being chosen in a trial featuring complex financial transactions). Much of the research on forepersons was conducted before 1970, however, and it is unclear if the findings related to a foreperson’s demographic characteristics still hold in the wake of systematic efforts to increase the demographic diversity of juries as well as significant changes in societal attitudes over the last 40 years.

Speaking and Discussion Content

In general, research has shown that speaking during deliberation is not spread evenly across jurors—several individuals tend to do the lion’s share of the talking, while a few members typically say little or nothing (especially in larger juries). As might be expected, forepersons generally speak more than the typical juror. Although there has been a pervasive fear that juries will get sidetracked easily and end up spending much of their time talking about irrelevant topics, this apprehension appears to be unfounded. A robust finding has been that juries (both real and mock) take their jobs very seriously and try to stay focused on the task at hand. Indeed, most of the deliberation time is taken up with discussion of legally relevant topics, such as the evidence (around 75%) and the jury’s instructions (around 20%). Juries also generally do a good job of correcting inaccurate statements made by their members, exhibiting good collective recall. Furthermore, when individual jurors introduce legally irrelevant considerations (e.g., the similarities between the current trial and a recent movie), this is often noted and sanctioned (corrected) by other members.

Initial Distribution of Verdict Preferences

One of the strongest and most reliable findings about juries is that the distribution of verdict preferences at the beginning of deliberation is a very good predictor of the jury’s ultimate verdict. Specifically, the verdict option favored by the initial majority tends to be the jury’s final verdict about 90% of the time. This finding is based on extensive research with mock juries where individuals were asked to provide an explicit verdict preference prior to deliberation, as well as several field studies where the first vote was reconstructed later and used as a proxy for the initial preference distribution in real trials. This robust phenomenon has often been referred to as the “majority effect.”

A good deal of research on “social decision schemes” in the 1970s and 1980s aimed to identify the specific probabilities associated with the various verdicts that occurred given every possible initial distribution of juror verdict preferences. Early studies suggested that factions within the jury would usually “succeed” in getting their preferred verdict in the end if they began with a strong majority or a higher share of the vote (defined as two-thirds or more). If no two-thirds majority existed initially, juries were theorized to acquit or hang with high probability. Subsequent work provided support for the majority effect but also identified an asymmetry in the probabilities working in favor of the defendant in criminal trials—a so-called leniency effect. The leniency effect essentially corresponds to an increase in the likelihood of a prodefense verdict for any given preference distribution relative to what would be expected if the nature of the verdict had no impact on the majority’s chance of succeeding. For example, a weak majority of 7 people in a 12-person jury is considerably more likely to succeed if it favors acquittal as opposed to conviction. This prodefense shift is consistent with the explicit value placed on giving the defendant the benefit of the doubt in criminal trials and the strict standard of proof needed to convict (i.e., “beyond a reasonable doubt”).

Consistent with both the majority and the leniency effects noted above, a recent meta-analysis of studies measuring early-verdict preference distributions (i.e., predeliberation or first vote) and final verdicts in criminal trials identified two critical thresholds of member support related to potential verdicts. When there existed a strong early majority favoring conviction (i.e., 75% or more of the jury), a “guilty” verdict was the usual result. Conversely, if 50% or less of the jury initially supported conviction, a “not guilty” verdict was extremely likely. The only time the jury’s final verdict could not be forecasted correctly was when a weak initial majority favored conviction, in which case “guilty,” “not guilty,” and “hung” verdicts were all relatively common.

The Dynamics of Consensus

In most deliberations, it becomes obvious at some point that all members do not favor the same verdict and some jurors will have to change their minds if the jury is to accomplish its task. A number of studies have examined how juries go from an initial nonunanimous distribution of preferences to a consensus verdict, and two general sources of influence have been identified: informational and normative. Informational influence is associated with the content of the jury’s discussion; it stems from the articulation of case “facts,” interpretations, and arguments supporting a particular verdict. Normative influence comes about from a desire by individuals to fit in with others and not be seen as deviant, incorrect, or even disagreeable. In essence, normative influence represents a pressure to conform that is rooted in the desire to avoid the social costs associated with standing apart from the collective.

Research on the dynamics of deliberation has shown that majority factions exert both informational and normative influence on minority (i.e., dissenting) jurors. In particular, several different theoretical models of majority influence converge on the conclusion that the degree of influence exerted is proportional to the size of the majority faction. In addition to exerting normative influence in direct proportion to their size, majority factions exert informational influence as well. In this regard, majority factions have a “sampling advantage” over minority factions during deliberation in that they have more members to draw on for evidence-related recollections, observations, and views consistent with their preferred verdict. Conversely, in keeping with their smaller size, minority factions are forced to rely primarily on informational influence. Research has also shown that juries tend to move away from their initial distribution slowly at first but with increasing speed as the majority faction grows and its influence increases in a snowball-like fashion. In essence, as the majority is given time to deploy its normative and informational advantages, it becomes harder and harder for the minority to prevail. Interestingly, though, a few studies have observed a momentum effect as well: Once the jury begins moving in either direction toward a verdict (whichever it happens to be), it rarely stops and changes direction. This suggests that a key event underlying a “reversal” of the early majority is when the first juror to change votes joins the minority faction.

Opinion polling is one procedural variable that may influence the dynamics of deliberation. Polling during deliberation can vary on several dimensions, including timing (e.g., early vs. late) and secrecy with regard to individual votes. In particular, early opinion polls conducted in a “public” fashion (where members can observe how other jurors vote) may bring considerable normative pressure to bear on members of the minority faction to change their vote. In one clever series of studies using mock juries with known verdict preference distributions (i.e., an even 3:3 split or a 4:2 weak majority), polling was structured so that all members of one faction voted first in sequence. Consistent with the research on conformity, the first member of the second faction was much more likely than chance to switch his or her vote and join the faction that voted first. This was particularly so when the vote was taken early and when the first faction voted “not guilty.” Of note, some research suggests that minority factions may have more influence if subgroups emerge (or are specifically formed) before a general collective discussion and the minority faction finds itself having a “local majority” in one or more of the subgroups. Taken together, these findings support the notion that the order in which things are done may distort the influence of faction size by affecting how obvious the factions are.

Another stream of research on the dynamics of deliberation has identified two general deliberation “styles” that juries may adopt as they go about their task: evidence driven and verdict driven. Evidence-driven juries work toward the goal of establishing the “facts” of the case as they see them before any discussion of the appropriate verdict. They spend a great deal of time in reviewing the evidence and take their first (and sometimes final) vote relatively late in the process. On the other hand, verdict-driven juries take an early first vote to get a sense of their members’ standing and then organize their discussion around the verdict options and which one seems more appropriate given what seems to be generally accepted by the group. Research suggests that a substantial number of juries (perhaps 33-50%) adopt a verdict-driven style, and studies in which deliberation style was manipulated suggest that its influence on the final verdict may depend on other variables, such as the required legal elements for the available verdicts.

Sometimes the influence process fails to produce a critical mass of opinion, however, and the jury cannot reach a verdict (i.e., it “hangs”), resulting in a mistrial. In a classic early study in the 1950s, the frequency of hung juries was found to be about 5%. More recent work with samples of juries drawn from several large metropolitan areas suggests that the frequency rate varies considerably across jurisdictions and may be as high as 15% to 20% in some places, although the observed frequency rate depends on whether hanging is defined at the level of the defendant as a whole or the specific charge. A primary determinant of hanging is the ambiguity of the evidence, with trials involving moderately strong evidence producing initial preference distributions without strong majorities. Small minority factions—especially those composed of only one individual—are rarely able to hold out.

The Dynamics of Monetary Award Decisions

Research on civil juries has increased dramatically since the 1980s, but the study of how juries arrive at damage awards is still in its infancy. Nonetheless, experimental research using mock juries has established a number of influences on both the likelihood and the amount of compensatory damage awards, including whether or not the trial is bifurcated (divided into liability and award phases), the characteristics of the two parties (e.g., the number of plaintiffs and the variability of their injuries), trial complexity, the explicit mention of “appropriate” damage awards by the attorneys, and the actual severity of the plaintiff’s injury. However, relatively little attention has been devoted thus far to the issue of how juries arrive at a specific figure, and at least three models are possible: (1) juries begin with a salient benchmark figure explicitly identified by one party or the other during the trial and modify it based on the content of the deliberation; (2) juries break down the award into components, assign dollar values to the various elements of the claim through discussion, and then sum the component values to achieve a total value; and (3) juries simply choose an amount that seems appropriate in a holistic fashion. Distinct from the issue of when these models might apply is the question of whether they would be executed by individual jurors, the jury as a whole, or both in a mixed fashion. Research examining the relationship between the amounts preferred by individual jurors and the collective jury award has shown that jury awards tend to be (a) larger than the central tendency of their constituent individual members and (b) less variable than individual award preferences (particularly in 12-person juries as opposed to 6-person juries). Related to the first point, some research suggests that the median of the individual award preferences (as opposed to the mean or mode) is the best predictor of the actual collective award.

Jury Deliberation Quality

Much of the research on juries can be viewed as a search for variables that influence the nature of the verdicts observed (i.e., guilty vs. not guilty). Relatively little attention has been directed at the question of how well juries deliberate and the extent to which this influences the nature of their decision, but several notable empirical findings bear on the issue. First, considerable research on the impact of jury size and assigned decision rule (e.g., unanimous, critical majority, simple majority) has generally shown these procedural variables to have little influence on verdict distributions. However, larger juries and those required to reach unanimity tend to take more time, generate a greater exchange of information, produce more satisfied members, and show less variability in their damage awards—all characteristics that might be taken to indicate a higher quality of deliberation. Second, evidence-driven deliberation styles tend to be associated with longer deliberations, a more thorough examination of the evidence, and less normative influence. In other words, juries that are larger, evidence driven, and/or required to be unanimous may be more likely to have higher-quality deliberations than smaller, nonunanimous, and verdict-driven juries.

Another question related to the notion of deliberation quality is whether the deliberation process amplifies or suppresses the biases of individual jurors. An early, well-known hypothesis was that the biases of individual jurors would tend to manifest themselves when the evidence was ambiguous and not clearly supporting any verdict, but the empirical research has been mixed. In some situations, deliberation seems to reduce the biases of its members, but in other situations it appears to magnify them. Most likely, there is no simple, straightforward answer to this question—whether deliberation accentuates or attenuates individual bias probably depends on the strength of the evidence as well as how that bias is distributed across jurors (and particularly whether it is concentrated in an early majority).

Finally, it is worth noting that a variety of initiatives intended to improve the functioning of juries have been identified, implemented, and occasionally examined in the past 25 years or so. These initiatives include rewriting and simplifying legal instructions, providing copies of the instructions to jurors, allowing jurors to take notes and/or ask questions through the attorneys, and allowing jurors access to exhibits and/or transcripts during deliberation. Most recently, jurors in some jurisdictions have been allowed to discuss the evidence prior to deliberation so long as all members are present. Most of these initiatives have been aimed at making jurors’ jobs easier by reducing their cognitive burden and removing any constraints on the process. While these changes have generally not been found to systematically influence the nature of observed verdicts and jurors almost always react positively to them, there have been few efforts to evaluate how the deliberation process itself is affected.

A Final Thought

An unfortunate consequence of an early (and well-known) likening of jury deliberation to the development of a predetermined photographic image appears to have been a squelching of scholarly interest in the dynamics of the deliberation process. After all, what value is there in studying what juries do if the final outcome is so reliably predicted by the preference of the initial majority? However, even though reversals of the initial majority preference are relatively rare, their absolute occurrence is still considerable given the thousands of jury trials held each year. In a free and democratic society, and with so much at stake for the individual participants as well as the community, it is clearly important that we understand how juries go about making their decisions. Furthermore, in a large number of those trials, deliberation itself is a critical determinant of what those decisions are.


  1. Davis, J. H., Kameda, T., Parks, C., Stasson, M., & Zimmerman, S. (1989). Some social mechanics of group decision making: The distribution of opinion, polling sequence, and implications for consensus. Journal of Personality and Social Psychology, 57, 1000-1012.
  2. Devine, D. J., Clayton, L. D., Dunford, B. B., Seying, R., & Pryce, J. (2001). Jury decision making: 45 years of empirical research on deliberating groups. Psychology, Public Policy, and Law, 7, 622-727.
  3. Ellsworth, P. C. (1989). Are twelve heads better than one? Law and Contemporary Problems, 52, 207-224.
  4. Hastie, R., Penrod, S. D., Pennington, N. (1983). Inside the jury. Cambridge, MA: Harvard University Press.
  5. Kalven, H., & Zeisel, H. (1966). The American jury. Chicago: University of Chicago Press.
  6. Tanford, S., & Penrod, S. (1986). Jury deliberations: Discussion content and influence processes in jury decision making. Journal of Applied Social Psychology, 16, 322-347.

Return to the overview of Trial Consulting in Forensic Psychology.