Why deliberative polling?The government should be commended for wanting to consult the people about the difficult choice it faces on possibly suspending the Shin Kori nuclear power plants 5 and 6. This is not just a technical decision. It is also a contested political decision raising value questions. If technocrats were to decide by themselves, whose values would they use? In a democratic society they should not just apply their own. But if the people are to be consulted, how best should it be done?
Consulting the people is not as easy as it sounds. If you just ask them, in self-selected meetings or online forums, you will get unrepresentative responses. It will be just the people who feel very strongly or the lobbyists who will respond. If you do conventional public opinion polls with random samples, you will get mostly a mere impression of sound bites and headlines. A great deal of social science shows that most people are “rationally ignorant” about most policy questions. If I have one vote in millions why should I spend a lot of time and effort studying the intricacies of the arguments for and against a nuclear power plant? My individual voice on the matter, or my individual vote (if it were somehow taken to a vote), would not make any difference among those of millions of other citizens. It is more rational for me to focus on my job and my family and all the things where I can make more of a difference.
Not only do conventional polls measure opinions that are mostly uninformed, they may not represent opinions that actually exist. If I am asked about an issue I have not thought about, I may well select an answer more or less at random rather than admit that I don’t know. Hence many Americans famously offered their “opinions” in polls about the “Public Affairs Act of 1975”, but there was no such law. It was fictional. Their answers were phantom opinions or “non-attitudes.”
Hence the case for what I call deliberative polling. It is designed to answer the question: What would the people think, under good conditions for thinking about the issue? The idea is simple. We need a survey of a good sample of the population, both before and after it has had a chance to deliberate in depth about the issue. So criteria for success concern whether a good sample participates and whether they are engaged in good conditions for really thinking in depth about the issue. The government reports that they intend to survey 20,000 from which to draw 500 to deliberate. Hence they need to have questions that permit a comparison of the 20,000 with the 500 in key attitudes about the issue as well as demographics. A successful Deliberative Poll (DP) will show few significant differences between the two.
Our account of the good conditions for deliberation are also straightforward:
Briefing materials on the issue that show the pros and cons of the policy options. These briefing materials should be vetted by an advisory group for balance and accuracy.
Moderators trained to facilitate discussions among the sample divided randomly into small groups of 12-15. The moderators impart no knowledge themselves but ensure a mutually respectful, civil discussion. The small groups clarify key questions directed to experts when all the small groups gather for plenary sessions with competing experts.
The process proceeds with alternating small group discussions and plenary sessions. The latter have experts who do not give speeches, but only answer the questions determined by each small group. The experts should represent different points of view on the issue.
The data is collected both with an extensive questionnaire applied both before and deliberation and also with the taping of the small group discussions for qualitative analyses. The questionnaire should include both the policy options and explanatory variables that would help explain why people support or oppose the policy options. A control group taken from the original 20,000 and interviewed at the time of the weekend would also be desirable.
This design permits the DP to put the whole country in one room under conditions where it can really think about the issues. It allows an assessment of how opinions change — and why. This process has been applied in 107 cases in 27 countries on six continents. It has helped bring wind power to Texas, a decision on nuclear power in Japan, a process of constitutional change in Mongolia, a decision on the press law in Macau, de-segregation of the Roma schools in Bulgaria, and many other policy changes on complex issues. I hope it will be used successfully in South Korea.
*The author is the director of the Center for Deliberative Democracy at Stanford University where he holds the Janet M. Peck Chair of International Communication.