Deliberative Polling and the inevitability of consensus: the evidence

In her argument that Deliberative Polling inevitably leads to consensus, Catherine Bannister makes the sweeping generalization that the Political Science discipline has learned nothing from Social Psychology, particularly in the area of Deliberative Polling. In support of her argument, Bannister extensively quotes social psychology research from the 1930s and 1940s. In so doing, Bannister falls into the very traps of human decision–making that Deliberative Polling is designed to combat: namely, jumping to conclusions based on sound bites and headlines. A more comprehensive search and analysis of information about Deliberative Polling uncovers a whole body of research, informed not just by social psychology, but a range of other disciplines spanning the entire twentieth century, including organizational, educational and cognitive psychology, sociology, organizational behaviour and management – just to name a few.

Bannister’s main argument is that Deliberative Polling inevitably leads to consensus. Leaving aside the debate about whether consensus is a good or bad outcome, if we extensively explore the research on individual and group decision-making it becomes obvious that many aspects of the process of Deliberative Polling actually counter the inherent flaws in human, particularly group, decision-making. Further, comprehensive analysis of any of the 20 Deliberative Polls conducted internationally (including the two in Australia) reveals that consensus did not occur as an omnipresent inevitability.

First, the evidence from the Deliberative Polls. For all 20 Deliberative Polls conducted thus far, both diversity of opinion and consensus are evident in post-deliberation opinions. Deliberative Polls are designed to solicit informed opinion on a range of sub-components of an issue, often involving more than 40 questions. The media tend to report and even sensationalize, the results on particular sub-components of an issue where consensus has occurred, often focusing on just one or two questions from the whole battery.

In Australia, the Deliberative Poll on Reconciliation is a typical Deliberative Poll where consensus was reached on some aspects of Reconciliation, but not on others. A majority of the Representative Australians did move toward an apology to members of the so-called ‘stolen generation’. Having been exposed to competing arguments about the very plausibility of that term, participants also explored competing arguments about whether or not to pay compensation to those who claim to have been forcibly removed from their Indigenous families. A majority of participants did agree with an official apology, but an apology bounded by strict caveats of proof of forced removal and consequent hardship. In contrast, for other aspects of Reconciliation, for example, a treaty, the Representative Australians remained relatively divided. Other examples abound, both for the Australian Deliberative Polls, as well as those from Denmark, Great Britain and the USA.

Bannister specifically argues that consensus reached on one question for Australia Deliberates: A Republic – Yes or No? (the move toward a 'Yes' vote for the referendum question), was a natural outcome of powerful and persuasive advocates for the Parliamentary Appointment model. Bannister gives too much power to the Yes advocates and not enough to the individual decision-makers themselves. Deliberation occurs over a period of several weeks. From the moment a citizen commits to attend the weekend of dialogue with peers, experts and advocates, they commence the deliberation process – they read more, they converse more, they conduct their own research. The citizens are sent a Briefing Document, for this Poll, a synopsis of the competing arguments for and against the referendum question. Both the Yes and No Committees for the republic referendum, particularly Kerry Jones and Malcolm Turnbull, had equal input into the Briefing Document and after some twenty iterations, agreed that it was a balanced synopsis.

Construction of the Deliberative Agenda involves similar consultation and iterations with lobby groups. As with the Briefing Document, the goal is to ensure relatively equal representation of experts and advocates from the diversity of perspectives reflected in the academic literature. The Republic Deliberative Poll was an unusual case where an unambiguous Yes/No dichotomy existed. This is not so for most Deliberative Polls, especially Reconciliation, where arguments covered the whole political spectrum from left to right. For the Republic Poll, Bannister argues that the presence of a more persuasive member of the "Direct Elect camp", such as Phil Cleary, would have changed the outcome.

As a member of the Advisory Group for Australia Deliberates- A Republic – Yes or No? Phil had a voice in both the Briefing Document and the Agenda. More importantly, renowned Direct Elect advocates, Jocelynne Scutt and Ted Mack, did attend and argued the Direct Elect perspective. The informed citizens rejected it. Further, evidence from nine regional Deliberative Polls conducted in Texas negates the view that individual advocates actually make a difference. The nine Texan Polls examined the various alternatives for the provision of energy in Texas. Results provide unambiguous repeated evidence that regardless of individual advocates, slight differences in Briefing Documents and Agenda, or different random samples of the general population, changes in attitudes on specific questions are amazingly consistent over time and region.

Finally, for every Deliberative Poll, post-deliberation questions are always asked about perceived balance of materials, attempts at persuasion by moderators and facilitators, and opportunities to voice dissenting opinion. For both Australian Deliberative Polls, the participants themselves maintained that equality, openness and balance reigned.

To address Bannister’s more general claim that Political Science, particularly Deliberative Polling, ignores lessons from social psychology: Deliberative Polling methodology is premised on countering a multitude of decision-making traps, but particularly the tendency to base conclusions on a ‘bounded’, limited search for information. When expressing an opinion about a specific issue most citizens have never thought extensively about the issue, or the questions being asked. Most do not want to appear ignorant or disappoint the interviewer, so an answer is made up on the spot.

Philip Converse (1964) labelled such survey responses as ‘NON-ATTITUDES’. In the United States, these ‘non-attitudes’ were clearly demonstrated in the now famous research about the Public Affairs Act of 1975. Large proportions of random samples of the American population supported or opposed this Act. Yet the Public Affairs Act of 1975 does not exist! More recently, on the 20^th ‘non-anniversary’ of that Act, a follow-up study examined public opinion about the repeal of the Act. The random samples were divided into those who were told that President Clinton had repealed the Act, and those who were told that the Republican Congress repealed the Act. Results supporting or opposing the Repeal of the entirely Fictional Public Affairs Act strongly correlated with political affiliation.

It is a well documented fact that most citizens possess little knowledge about political issues. The knowledge they do possess is often gained from eight-second sound bites gleaned during 30 minute news summaries, 30 second political advertisements, or newspaper headlines. Thus, on many issues, few people have truly well informed positions. Why? People may not pay attention; they may not be motivated; they may not have the time or resources to focus on the issue. They are, using Anthony Downs label, ‘RATIONALLY IGNORANT’. It is in an individual’s interest to economize on seeking and storing information about public policy issues. For busy lives, shortcuts and heuristics save time. We pay attention only to what matters, when it matters.

Any number of the different kinds of heuristics - "representativeness", "recency", "insensitivity to base rates", "misconceptions of chance", the "conjunction fallacy" - could be operating on conclusions drawn about public policy issues. The whole body of research examining heuristics and their impact on human decision-making has informed debate among political scientists about the extent to which such heuristics affect voting behaviour.