Deliberative polling methods will inevitably lead to consensus. So what?

Tim Dunlop's piece on deliberative polling demonstrates how little the discipline of political science has learned from social psychology. Social psychology is the study of how people are influenced within groups, and as such is vitally important to understanding, critiquing, and predicting the ramifications of different political systems and ideologies. Most public relations officers and successful politicians have a gut feel for it.

Social psychology has found that people will conform to group opinion, even in contradiction to their senses. In a classic 1936 experiment, social researcher Sherif demonstrated the development of spurious group beliefs using an optical illusion called the autokinetic effect. If a pinpoint of light is shown in a darkened room for a short time, it appears to oscillate. The distance it seems to move is erratic, varying across trials and observers. Sherif asked his subjects to estimate the distance over which the light was moving. The subjects initially did this in isolation, and subsequently were brought together in groups of two or three. When they were tested alone, they saw the light moving a distance that was completely individual, differing widely across the trials and participants. However, when the subjects were brought together and nominated the distance aloud, their answers converged until finally they were in agreement. When the subjects performed the task individually afterwards, this agreed value, or group norm, persisted.

This was the first time it was demonstrated that a group consensus could be reached which bore absolutely no relationship to anything in the outside world. The conclusion is remarkably robust, having been tested a myriad of times over various experiments in the intervening years. Groups just naturally do it.

This is precisely what would have been happening in Tim Dunlop's example of the deliberative poll on the Republic issue. Contrary to Dunlop's argument, this was not a bunch of individuals each independently coming to a resounding agreement. It was a group coming to a group decision. We know absolutely nothing about the composition of this group, other than that it was "ordinary Australians". Who was the most influential in the group? Who was marginalised? Who spoke loudest? Who asked the most questions? If someone as persuasive as Phil Cleary or Peter Reith had been in the group, that group would almost certainly not have reached the same conclusion. I suggest that there were a core of very persuasive members of that group leading the others. The reason that I can suggest this is that social psychology has found that it's true of every group of humans. It's very probable that even by just giving answers, the expert panel played leaders.

Through the 1930s and 1940s, a fellow called Newcomb ran the famous Bennington study, a longitudinal study of an American college. Bennington was a small private college with a strong liberal political ethos. It recruited largely from conservative upper middle-class families. Newcomb tested their political beliefs through their time at the college and demonstrated that while they went in conservative, they came out liberal. These were intelligent students who were taught to think independently, but even so they switched allegiances in line with the college norm. I'll cite just one example to make the point.

"During the study the 1936 US presidential election occurred and within the college a mock election was also to be held. From the first-year students, who had only been at the college a month or two, there was a solid majority for the conservative Republican candidate over the more liberal Roosevelt (62 per cent vs 29 per cent). This was entirely consistent with their conservative family backgrounds. However, the third- and fourth-year students, whose families were no less reactionary, voted 54 per cent to 19 per cent in favour of Roosevelt, and fully 30 per cent of them voted for the Communist/Socialist candidates! (versus only 8 per cent of the first years). The impact of Bennington's liberal norms on its students seems clear." --Rupert Brown, Group Processes: Dynamics within and between Groups, 1988.

It might be possible to argue that, perhaps, the more liberal point of view is more logical, and that as the students were attending a university which encouraged free thought, they reached their conclusions through individual reasoning. But clearly this is not the case, as many intelligent people are conservative.

This finding, that people’s ideas are influenced by the groups to which they belong, has been repeated in many different situations. There was a study of members of the British fascist group the National Front. When acting individually these subjects didn't bear any grudges against blacks and even professed non-violence. When, however, they thought of themselves as group members, they were prepared to participate in riots, race intimidation and attacks.

Then there's the Moscovici study, much beloved of psychology tutors: the subjects were asked to call out the colour of an unambiguously blue slide. In each group, there were just two planted confederates who called out green. Fully 32 per cent of the subjects called out green at least once. The slide was really obviously blue, but just the presence of dissenters caused some students to doubt their own senses. Not one of the control group (where there were no confederates) ever saw the slide as green. When these subjects were then tested individually on their blue-green threshold (or the colour at which blue turns to green) the ones who had been more influenced were more likely to see green slides, indicating that it actually had changed their perception. Thus even a loud intra-group minority can influence the group. This list is not at all exhaustive.

We all like to think of ourselves as independent thinkers, and on a personal level that’s what we are likely to be. But we have to derive our moral standards, values, ideals and premises from somewhere, and that's the community at large and every social group to which we belong. To a huge extent our beliefs are driven by our church groups, our work places, our lecturers, our political affiliations, our friends and our neighbours. Only the most self-deceptive could claim immunity. Latterly even television has become a de facto social group, which is why many people instinctively worry that the American content on our screens leads to cultural imperialism.

Tim Dunlop's piece displays a number of erroneous assumptions. First, it assumes that there is only one right conclusion to be drawn, and that if only people are allowed access to information without influence that they will reach this conclusion. This is a sort of absolutism. Second, as a corollary to the first, that information can be given without influence. Third, there is the assumption that just because these "ordinary Australian" opinions lined up with "elitist" opinions in that particular case, that in a deliberative poll they would do so on every issue. Fourth, that coddling "ordinary Australians" into a different belief is any different from any other propaganda session. Last, and I think underlying the whole argument, is a belief that a perfect democracy will give rise to a perfect society in which people would be so politically exhilarated that they would not resort to racism or other antisocial behaviour.

Dunlop has since gone on to develop the theory, concluding that "leaders" as such are not necessary if the population is "engaged politically" (on Margo Kingston's Webdiary, "Two Nations Tragedy", www.smh.com.au/news/webdiary). This is just ideologically driven tripe.

The study of public policy needs to take the 75 years' of research by social psychology into account. What they are missing is rudimentary stuff. Social psychology has developed more recent and esoteric explanations like social identity theory and self-categorisation which can add to the criticisms that I have made. It seems to me that there is a desperate need for some cross-disciplinary knowledge sharing.