Competence of Randomly Selected Groups

From "The Trouble With Elections: Everything We Thought We Knew About Democracy is Wrong," Chapter 13.3

Are everyday people competent enough to understand policy implications, or do we actually need superior individuals — an elite — at the helm? As Hélène Landemore puts it:

“The traditional aristocratic prejudice against democracy stemmed from the belief that the people were neither wise nor virtuous enough to rule. The contemporary version of the same belief is that the people are irrational, uninformed, and apathetic.”

A few sortition advocates sidestep the concern by suggesting some sort of minimal competency test, perhaps comparable to the test immigrant applicants for USA citizenship are required to pass. Alternatively, simply allowing citizens to remove themselves from the sortition pool, or, as in ancient Athens, requiring them to proactively add themselves to the pool by volunteering, might remove many of those with no capacity or willingness to serve the community in this manner. I argue that such filtering approaches would do more harm than good by limiting diversity. Ordinary, diverse citizens within well-designed mini-publics are not only adequately competent as a group, (not each as an individual), but demonstrably more competent than relatively homogeneous groups of elites.

When examining mini-public competence it is important that we not conflate the incompetence of some individual members of a group with incompetence of the group as a whole. The sortition proposal seeks to have typical citizens act in groups, where the poor judgment of a few members with weak cognitive abilities will not (with proper institutional design) spoil the group decisions. Within deliberative groups, individuals who are significantly incompetent tend to be recognized as such by those working with them, and are not persuasive. Thus the real concern is not whether any incompetent individuals would be selected using a lottery system, but whether a ruinous quorum of a large group would be incompetent. The evidence from the study of court juries shows that this wouldn’t be a problem for even a relatively small mini-public. But beyond this limited corrective notion of a group’s competence or intelligence, the cognitive sciences recognize the existence of distributed intelligence within groups that can generate an emergent intelligence (referred to as collective intelligence in previous chapters) that surpasses the intelligence of the best individuals in the group. 

Two studies with 699 participants conducted by researchers from M.I.T. and other universities led by Prof. Thomas Malone found that groups of people exhibit a sort of “group intelligence,” which predicts their ability to successfully tackle a range of tasks. The interesting thing, as they reported in the October 1, 2010 issue of Science magazine, is that this group intelligence was not strongly correlated with either the average or maximum individual intelligence within the group. Instead, much to their surprise, the researchers found that it was more correlated with

“the average social sensitivity of group members, the equality in distribution of conversational turn-taking, and the proportion of females in the group.” As one of the researchers concluded, “Having a bunch of smart people in a group doesn’t necessarily make the group smart.”

So, what if a group of average citizens had the relevant information, opportunity, and incentives to tackle complex policy questions? Could they do as well, or better than politicians in Congress? The results of numerous studies conducted by the Program for Public Consultation (PPC), affiliated with the School of Public Policy at the University of Maryland are promising. Unlike a traditional “off the top of the head” polls, for decades they have presented a representative sample of about 2000 Americans with information on challenging policy issues, such as the budget deficit challenge,¹ but also give respondents key information and a wide range of arguments on the issue. Their goal is for respondents to have a deliberative experience like that of a policymaker. In one of these surveys, for example, through a combination of budget cuts (particularly to the defense department), and tax increases (such as income taxes on the wealthy and the estate tax on large estates) participants were able to dramatically reduce the deficit.² This consultative exercise did not involve deliberative discussions among participants, (though, they did engage in internal deliberation within their own minds), and there were no lobbyists present. The unanticipated consequences of their hypothetical proposals were unknown, so this experiment was too simplistic to be conclusive. Yet, Steven Kull, director of PPC observed that,

“when given information and a chance to sort through their options, most Americans do a pretty good job of dealing with America’s budget problems – better than most politicians.” 

A more comprehensive experiment that included small group face-to-face deliberation was conducted in June, 2010 with 3,500 participants at 57 sites across the U.S. The project, called “The People Speak: Our Budget, Our Economy,” funded by the MacArthur Foundation, was evaluated by researchers from Harvard and two California Universities. Organizers brought together a representative sample of Americans to evaluate a broad range of federal deficit reduction proposals. The groups’ recommendations were strikingly similar to those produced by the PPC project mentioned above. The researchers reported that

“Both liberals and conservatives appear to have moderated in their policy views regarding spending cuts and tax increases.”

Prof. James Fishkin, has conducted many experiments with “deliberative polling,” in which random samples of citizens engage in carefully structured deliberation on complex public policy issues. The insights he has gleaned about competence are worth noting. He points out that

“Whether or not ordinary citizens appear competent may well depend on whether they have reason to pay attention, whether they think their voice will matter, how discussions and interactions are conducted, and how any data about their views is collected.” [When The People Speak, page 68]

Fishkin concludes that

“The picture that emerges from Deliberative Polling and indeed from other deliberative consultations is that the public is indeed capable of dealing with complex issues, once it believes its voice matters, once it believes there is reason to spend time and effort in public discussion, listening to alternative points of view.”

After examining the dozens of citizen deliberative processes around the world, including citizen juries, citizen assemblies, planning cells, participatory budgeting, etc. Frank Fischer observes in Democracy and Expertise that

“they have shown that citizens are capable of much more involvement in thinking about and assessing technical questions than is conventionally presumed.”

He goes on to cite multiple studies by various researchers who studied the Citizens’ Assembly in British Columbia showing that the

“intellectual and deliberative capacities of the ordinary citizens to have been impressively high … Citizens not only made good, reasoned choices, but that they did at least as well as the experts on the topic.”

1. Note that this all assumed that government deficit spending was a serious problem, in contrast with the analysis of those economists who subscribe to modern monetary theory (MMT), which argues that deficit spending is a normal and useful thing for governments to do, since all money is created by fiat, so long as inflation is kept in check.

2. Individual members of Congress have unique interests, including employment by defense contractors within their district and campaign contributions, that conflict with these popular priorities. But even if most members didn’t, Congressional practices, such as log-rolling, where members trade votes benefiting favored special interests, effectively block such solutions.