LEGISLATIVE PANELS

In effect, the priority panels set the agenda for legislative panels (box 6). To move this hypothetical example from the national level to the state level, let us say, for example, that the Ohio secretary of state has already convened priority panels on the state legislature. The secretary of state would then use a nonproflt polling flrm to select a new set of ten representative legislative panels, each of which would consist of flfty citizens.

Although a large sample would be more representative, it would also prove quite costly, and I think an argument can be made for smaller groups. (Those who cannot stomach this idea should read on and simply substitute larger samples as necessary.) The point in convening flfty citizens is to ensure that the sample includes the full range, if not the exact distribution, of viewpoints in the larger population. Panels with just flfty participants cannot perfectly represent a large population in the statistical sense, but quotas can make the sample sufficiently representative. By combining demographic quotas (for age, ethnicity, etc.) and possibly attitudinal quotas (for party affiliation, basic opinions, etc.), each panel of flfty citizens can be made a microcosm of the larger state population.^[23]

DELIBERATION ON LEGISLATION

The ten legislative panels would come together to Columbus, Ohio, but they would deliberate and vote independently. Each panel would be assigned to a single bill, and the panels could either deliberate simultaneously (but separately) or in sequence. The latter might prove less chaotic, and it could permit serial media coverage of the bills during the election season.^[24]

Prior to arrival in Columbus, citizens would receive information about the bill assigned to them. To promote continuity in judgments over time, panelists would also receive copies of the judgments of past panels who deliberated upon related legislation. Undoubtedly, many of these citizens would discuss their bills with friends and family beforehand, and such discussion could only beneflt the overall process. Indirectly, it would involve many more people in the process, and it would have a positive average effect on citizens' knowledge and the quality of

---

subsequent panel deliberation. Because the citizens are geographically dispersed prior to arriving for panel deliberations, this initial period would give them the chance to hear the quiet voices of average citizens from across the state. Some panelists will undoubtedly make reference to those interactions during panel deliberation.

To ensure that citizen panelists do share a common knowledge base, however, the first day of deliberation would consist of testimony from witnesses chosen by the highest-ranking members of the legislature in favor of and opposed to the bill that each panel considers. The structure for such testimony, cross-examination, and question-and-answer could take many possible forms, drawing on the examples set by the random sample forums described in chapter 6. At the National Issues Convention, pro and con experts each made statements and then took turns answering questions from the audience. At the New Mexico Citizen Conferences, each of five or six expert witnesses made a brief statement, then citizens asked questions, which experts answered in a free-for-all format. Citizen juries have used many discussion methods, including a straightforward succession of witnesses, each of whom devotes half of his or her time to presentation and half to a question-and-answer session. A new approach taken in a recent citizen forum used a panel of witnesses in which one speaker would make a presentation, then the others would give balanced critiques of the speaker's arguments.^[25]

This process of cross-examination is important because it exposes witnesses to professional scrutiny. Those testifying before the legislative citizen panel will be more reluctant to make false statements if they see it as a high-risk strategy. At a minimum, witnesses risk the erosion of their credibility if cross-examination reveals that they have used fabricated survey data, lied about their prior knowledge, or otherwise sought to deceive the citizen panel. Following the courtroom analogy, one might even want make these witnesses "subject to penalties for perjury"^[26] I if cross-examination "reveals that a witness made false statements.would stop short of placing a judge in the panel room, but witnesses guilty of perjury might be subject to civil lawsuit after the fact.

Whatever particular format is used, it is important that the number of witnesses remain manageable, that the citizens be able to guide the direction of the discussion, and that either a paid moderator or a citizen-chair maintain decorum and keep the panel on schedule. My own preference is for a paid, professional moderator to manage the meetings under the direction of the citizen panelists. Following the example of the citizen juries, the moderator would be responsible for establishing and

---

maintaining the ground rules for the discussion and keeping the event on track. Moderator duties would include stopping quarrels among witnesses, helping panelists clarify questions and get direct answers, and cutting off witness answers that run too long. The panelists would first meet the moderator during a brief orientation session, during which the moderator would go over the discussion guidelines sent to participants. This brief training would be more advanced than typical jury training, in that the moderator would openly discuss common group problems, such as inattention, communication style differences, and the formation of cliques. During panel deliberation and the question-and-answer sessions with witnesses, moderators would also keep track of speaker time and periodically invite quiet participants to speak up if participation rates became greatly unbalanced.^[27]

After the first day of testimony, citizens would have a day to deliberate on their bill. The course of the deliberation itself is impossible to predict, but it could be structured to ensure full participation and conformity to the preset schedule. A trained moderator could have the responsibility of periodically reminding a panel how much time it had and what it needed to do. The moderator could also intervene if verbal conflict moved toward physical confrontation, which, incidentally, has been a rarity at random sample forums conducted in the past. The moderator would also call for three or four "round-robins" at roughly preset intervals. In a round-robin, each participant has a chance to speak for a minute, and speaking turns move from left to right without interruption. Its purpose is to draw out quieter participants, and it can cause an otherwise silent voice to shift the emphasis or direction of deliberation.

After deliberating through most of the second day, the citizen panelists would write what they considered to be the strongest arguments for and against the bill. This set of arguments would be as inclusive as possible, and it would provide a succinct written record of the facts, ideas, and considerations influencing in their deliberation. Aside from this listing of arguments, the panel would not take anything resembling a preliminary vote or straw poll on the issue.^[28]

On the third day, expert witnesses would return and have the opportunity to respond to the arguments of citizens for and against the bill. Again, this process could follow many formats, but during this stage, it would be important that the lead witnesses be able to direct discussion toward their most pressing concerns. This day would be their last chance to bolster arguments, correct factual errors, and remove confusion or misunderstandings.

---

As citizens undertake their final deliberations on the fourth and flfth days, the debate they have heard on the previous day should give the panelists more confldence in their final votes. Citizens will have had a chance to be second-guessed by outside observers prior to reaching their final decision. By the end of the fourth day, the moderator will require the panelists to take their first formal vote by secret ballot.^[29] If twothirds or more of the panelists agree on the issue (with abstentions counting as "no" votes), they will then divide into pro and con subgroups to write one-paragraph rationales for and against the bill. If there is not a two-thirds majority for or against the bill, deliberation will continue until the middle of the flfth and final day, with the moderator calling for periodic votes. If still deadlocked, the citizens will divide into pro and con groups to write rationales, but their verdict would be recorded as "undecided.^[30]

Just as a priority panel would have guidelines for its deliberation on the legislative agenda, so legislative panels would have preset goals. Their task would be to reach agreement (two-thirds majority or greater) on the bill before them. Also, they would be asked—but not required— not to take a solid "yea or nay" position on the bill until their vote on the fourth day of deliberation. In the case of a statewide legislative panel, citizens would also be encouraged to try to base their vote upon their final judgment of what is in the interest of the state as a whole. Prior to arriving at that point, however, participants would be invited to contribute their own unique perspectives and experiences. Ultimately, citizen panelists would manage to combine genuine self-expression with the pursuit of a larger public judgment, but as Lynn Sanders points out, it is also important "to ensure that those who are usually left out of public discussions learn to speak whether their perspectives are common or not."^[31] Most civil and criminal jurors manage to take on the role of the "neutral judge" briefly, and most past participants in past random sample forums have taken their jobs very seriously.^[32] So might the participants in a legislative panel.

IMPLICATIONS FOR ELECTIONS

Because they involve substantive judgments about speciflc bills, the legislative panels could have far more powerful electoral implications than the priority panels. When secretaries of state present these panel results to voters, they would show candidates' votes on key legislation, but they would also permit voters to compare each candidate's vote with a

---

democratic public voice—the verdicts of a representative, deliberative, and articulate body of fellow citizens (see box 7). This would make it possible for voters unfamiliar with issues to judge candidate votes more critically by moving past their own surface-level judgments and using those of the legislative panels. In effect, the voter guides distributed after legislative panels would permit a voter to simulate deliberating about ten national issues and comparing final deliberative judgments to candidate positions.

Some might wish to stop at this point, having provided interested voters with detailed printed and on-line information about candidates for the Ohio state legislature and the views of a deliberative sample of the state's population. I suggest going one step farther by providing a very simple piece of information that reduces the detail of the voter guide to a single numeric symbol—the percentage of agreement between the votes of each candidate and those of the legislative panels. The rationale behind this number is that past voting research has found that the typical voter prefers to choose candidates using only simple bits of information, as explained in chapter 3. Voters "tend to make use only of information that is incidental" or available information "for^[33] which there is virtually no cost, such as names and parties on ballots. Even when voters lack reliable clues that permit confldent deduction of which candidate would best represent them, they can flnd highly suggestive "cues" that tell them how to vote.^[34]

In candidate selection, as in most mental activities, people are cognitive misers. Considering the likelihood of one vote deciding an election, plus the finite difference between a typical pair of candidates, citizens attribute little value to making a correct voting choice. Consequently, voters give low priority to candidate selection relative to the more immediate concerns they must attend to in their daily lives. This may be truer of less sophisticated voters, but when cues such as a candidate's party affiliation are present, even more involved voters can overlook complex candidate characteristics and make decisions solely on the basis of simple heuristics.^[35]

In low-visibility elections, where the average voter has heard little about the candidates, many voters lack solid data upon which to base their choice, so they make their decision using simple facts, such as the candidate's party and sex.^[36] When voters know nothing about candidates prior to entering the voting booth, they must either make random choices or respond to the scant information that appears on the ballot itself. Assuming even the slightest sense of responsibility, voters are likely to take the latter course by voting based upon at least some information, no matter how mundane. Sometimes this means voting based on party, although that cue is often unavailable. Other times, voters choose candidates because their first name sounds male or female or their last name suggests a particular ethnic background.^[37]

The political scientist Arthur Lupia considered this problem in a study of insurance reform elections in California. As others have observed in

---

national elections, Lupia found that some unsophisticated voters were able to make complex judgments by relying on cues from like-minded elites. Transferring the lesson of that election to the larger political process, Lupia recommends improving the quality of cues rather than pursuing wholesale civic education: "While scholars and pundits propose that we educate the public about politics in order to lessen the impact of uninformed votes on the responsiveness of democratic decision-making institutions … directing our efforts into the provision of credible and widely accessible ‘signals' may be a more effective and cost-efficient way to ensure the responsiveness of electoral outcomes.^[38]

I suggest providing voters with a new cue—a candidate rating that shows the extent to which each candidate's unofficial and official votes correspond to those of the legislative panels (see Box 8).^[39] The cue has the same virtue of other ballot cues, in that it is simple and readily available to even the least active voter. What sets it apart from all other ballot cues, except sometimes political party, is that it carries a tremendous amount of information. Particularly when one candidate has an agreement percentage (or "rating") well above the other candidates, a voter can use that cue to decide that the highest-rated candidate stands a much better chance of representing the public's interests. The cue is not flawless, but it provides uninformed, confused, or undecided voters with one simple fact to consider before marking their ballots.

---

Finally, county clerks can use the candidate ratings to order the names of candidates on the ballot. Many candidates believe that the order of names on the ballots influences the outcome of elections by a small percentage. A rigorous study of this problem found name-order effects in 48 percent of 118 Ohio elections. In races where there was an ordering effect, the candidate listed first almost always beneflted, and the average impact was 2.5 percent. Partisan races and elections in counties with higher levels of political information were less likely to demonstrate ordering effects.^[40] It is discouraging to imagine that innumerable close races have been decided by the order of candidates' names on the ballot. Even if one doubted the necessity of convening citizen panels for all elections, one might still see wisdom in obtaining panel judgments in competitive races just to remove this problem. By listing candidates in descending order, with the highest-rated candidate first, clerks can assign this advantage in a manner that is less arbitrary than random ordering.