pageok
pageok
pageok
Deliberation Day Experiment in Colorado.--

Over at the University of Chicago faculty blog, Cass Sunstein reports on an experiment on the effects of deliberation on views:

Deliberation Day and Political Extremism

A few months ago, David Schkade, Reid Hastie, and I helped to organize a kind of Deliberation Day in Colorado. (The events were sponsored and funded by ABC News, which should be broadcasting our experiment soon, in connection with a general discussion of political polarization in America.) Two cities were chosen: Boulder (a predominantly liberal area) and Colorado Springs (generally Bush country). About 60 citizens were brought together to explore three of the most controversial issues of the day: affirmative action, an international treaty to control global warming, and civil unions for same-sex couples.

People in Boulder deliberated with others from Boulder, and people from Colorado Springs deliberated with people from Colorado Springs. Thus people were generally sorted into groups of like-minded people. . . .

Our key question was this: What would be the effect of deliberation on people's views?

Here are our three major findings.

(1) Liberals, in Boulder, became distinctly more liberal on all three issues. Conservatives, in Colorado Springs, become distinctly more conservative on all three issues. The result of deliberation was to produce extremism — even though deliberation consisted of a brief (15 minute) exchange of facts and opinions! . . .

(2) The division between liberals and conservatives became much more pronounced. . . .

(3) Deliberation much decreased diversity among liberals; it also much decreased diversity among conservatives. After deliberation, members of nearly all groups showed, in their post-deliberation statements, far more uniformity than they did before deliberation.

Thomas Roland (mail):
That's damnably scarey! To rectify this situation and enforce moderation, we need a program of forced relocation. Please discuss.
2.3.2006 12:54am
SteveB:
Not very surprising. I'd like to think I was above such groupthink but I've noticed this phenomenon in myself when I've previously limited my sources of information to those biased in a particular way. I have to make a concerted effort at balance things out now.
2.3.2006 1:33am
Robert Schwartz (mail):
I guess we can take that idea off the table.
2.3.2006 2:16am
Splunge (mail):
So talking to your own cheerleaders gets you pumped. Golly, who'd have suspected?
2.3.2006 2:20am
therut (mail):
Was this some kind of Government Funded study that tells us all what we already knew? I realize it was not but it sure sounds like alot of studies our tax dollars go to.
2.3.2006 2:24am
Bruce Wilder (www):
I fail to see what is "deliberative" about exchanging views and opinions with people you already basically agree with.

The essence of deliberation is having to confront differing or opposing views, and to have one's own views challenged and criticized. Ultimately, logic is compelling and facts are objective, so it is possible for deliberation among people with opposing points of view to reach moderating compromises, superior to radical, but ill-considered positions.

Unchallenged and reinforced, of course people confuse their point of view with Truth. That's just human nature. Duh.
2.3.2006 2:51am
Wintermute (www):
Sounds like a follow-up to this ABC story I blogged a while back here.
2.3.2006 4:08am
Jack John (mail):
It confirms that one can create one's worldview by choosing one's associates. Does one want exposure to diversity, then? What if one wants to become a libertarian through and through? Diversity would just get in the way, with all those non-libertarians interjecting their opinions into the mix. What if one wants a masculine perspective on things? Hard to do with women around. Diversity destroys other values, like self-selected conformity.
2.3.2006 4:35am
llamasex (mail) (www):
This isn't is a symptom of human nature. It is a symptom of the media and culture we have today. The soundbite arguement crossfire world has lead us here, rational discussion is unpopular. People see the Bill O'Rielly's as today's idea of a debator, but these people don't debate. They pick a side and run off talking points. Each team has them (even the third party Bad News Bears). The result being each side dug in.
2.3.2006 5:38am
Okie:
Gee, I'm shocked....
2.3.2006 5:52am
J..:
Perhaps you people are smarter than I (I suppose I shouldn't say "perhaps), but *I* found the results surprising.

The surprising part was *not* that people don't change their mind when presented with non-contradictory views. The surprising part was that people became *more* extreme when presented with non-contradictory views. It is that conclusion that I found interesting.

What is also interesting for a legal blog is that Sunstein came to similar conclusions as part of the Judges Project:

2. When Republican [judicial] appointees sit only with Republican appointees, and when Democratic appointees sit only with Democratic appointees, the gap grows -- by a lot. Republican appointees show far more conservative voting patterns when sitting only with other Republican appointees; the same is true for Democratic appointees on the liberal side.

Link.
2.3.2006 7:55am
Sailorcurt (mail) (www):
First of all, 15 minutes is not enough time to "deliberate" anything.

Secondly, as others have pointed out: like minded people don't deliberate, they reinforce.

Thirdly, this explains why it is so difficult to change a "corporate mentality". For example, most members of the mainstream media are leftist, therefore, they tend to reinforce each other's worldview until they believe themselves to be "the mainstream" regardless of the validity of that belief.

Finally, this is why current paradigms in "diversity" are totally ineffective. "Diversity" efforts concentrate on diversity of skin color and diversity of sexual orientation, not diversity of thought. You can have the most diverse group imaginable in regard to race, creed, color, sexual orientation, etc...but if they all have the same preconceptions and belief systems, the group is not "diverse" at all and will only serve to reinforce those belief systems and preconceptions within themselves rather than provide for an honest evaluation of those beliefs.
2.3.2006 7:55am
Medis:
J.,

I agree, there is a subtle point here that people may be missing. Some people are using the word "reinforce", but I think that is not quite right, at least insofar as it implies that people are simply settling into their original views, perhaps with more convinction or certainty. Rather, what is interesting about Sunstein's work is precisely that these "deliberations" are leading to a change in views--and the change is toward the extremes.

I think that is a nonobvious result. Moreover, I think it is a result worth "measuring", in the sense that it would be useful to know things like how quickly this effect can occur, its magnitude, and so on.

Splunge,

Speaking metaphorically, I think one way to state this result would indeed be to suggest that when like-minded people set out to "deliberate," they actually end up "cheerleading" instead, in that they work themselves toward even more extreme views. So, while I think it is obvious that "cheerleading" would have this effect, I think what is nonobvious is that people who set out to deliberate have ended up "cheerleading" instead.
2.3.2006 8:29am
Medis:
Additional comment to Splunge:

This thought occurred to me immediately after posting, so I apologize for the additional post.

Anyway, it seems to me the practical import of this result is that a group cannot simply avoid this "cheerleading" by resolving to conduct their discussions in a certain way. Rather, the answer to this problem must be structural: they have to foster the inclusion and participation of people with diverse views.

Again, I think this is an important idea. In short, I think a lot of us would like to believe that we are capable of choosing to have more self-critical discussions, but it may well be impossible to effectuate that choice without including actual critics in the discussion.
2.3.2006 8:34am
Cecilius:
Again, not surprising. Conservatives have been making the very same accusations against the media and academia for years. The one curious aspect of Sunstein's conclusions are that discussing an issue with like-minded folks "produce[s] extremism." Unless Sunstein defines the term differently (to mean something other than an outlying or unsupportable viewpoint), nothing about this case study suggests extremism. 50 years ago, if a group of anti-segregationists conferred for 15 minutes and agreed that "yes, racial segregation is in fact bad policy," how does their viewpoint suddenly go from a respectable and ultimately prevailing belief to "extremism"?
2.3.2006 8:43am
Paul Sturm (mail):
I can see the headline now -- "Like-minded people reach consensus". Scary stuff!

Or maybe, "wishy-washy middle-of-the-roaders helped to find consistency in their political views". :)
2.3.2006 9:29am
keatssycamore (mail) (www):
Sunstein says there were 60 people from Boulder and 60 people from Colorado Springs. Then he seemingly assumes that the 60 from Boulder are liberal and the 60 from Colorado Springs are conservative. Does he have anything with which to back up that key assumption?

Did the study pre-screen the people by political persuasion? This is left unclear. So as it stands, with just 60 people from a certain city, why could it not be that 'deliberation' was actually changing some minds?
2.3.2006 9:42am
Zubon (mail):
I would point to a different aspect in "reinforcement," in the way that it likely occurs. In each group, you have people will generally shared assumptions. Different degrees of conservatism and liberalism will often arise from how closely one maintains consistency: the more extreme people are those who are willing to follow their principles to their logical conclusions, while the moderates do not feel comfortable doing so. Moderate conservatives/liberals are perhaps also more likely to feel less passionately about those principles, the importance of holding to them strictly, etc. When placed together, the moderates feel intellectually compelled to remain consistent to the principles they claim to espouse, and are therefore led towards the extremes. Stylized examples follow.

Assume that you are a moderate liberal discussing global warming. You all agree that global warming is real, that ecological disaster is just a matter of time if nothing is done, etc. The moderates probably accept that idea in the abstract, but they feel hesitant to damage the economy. After all, your car and job are very concrete, immediately present things, while a few degrees a century away can seem pretty distant. You probably think that something ought to be doing something about that, but you do not think about it much. Confronted head-on with the issue, you are forced to concede that, yes, global warming will wipe out all life on the planet, etc. Are you willing to stand up and say, "But I really like my SUV," when you and everyone around you have conceded that you are contributing to the end of the world?

Assume that you are a moderate conservative discussing civil unions. You all agree that heterosexuality is the human norm, that marriage is a fundamental building block of society, etc. You probably think that civil unions are just a stepping stone to gay marriage, which you oppose, but you don't want to seem like one of those mean people who hate gays. Confronted head-on with the issue, you are forced to concede that, yes, the whole thing is a threat to the traditional family that will destroy our entire society. Now that no one around you will think you are mean, are you willing to be a squish on what could be the end of Western society?

Have I tossed in enough hyperbole? I don't want to argue either issue. I just think that, when pressed on their fundamentals, many people will give up their "yeah but"s that make them more moderate. I think many people approach political questions pragmatically rather than from first principles, and many of them would feel bad about it when pressed. We all like to think that our beliefs arise from important principles. When confronted with differences between our principles and the conlusions we claim to have drawn from them, we will probably cling to the principle. Once we have identified with the label, we all think we should be better conservatives, liberals, Christians, Objectivists, whatever, if only we could hold fast to those principles... In a 15-minute discussion, you do not have much time to formulate why you should be moderate in this specific case or why some other principle should dominate. And all this before groupthink sets in and we go along to get along.
2.3.2006 9:44am
Zubon (mail):
Or maybe Paul Sturm said it much more succinctly while I was reading and typing.
2.3.2006 9:46am
Taimyoboi:
I'm beginning to sense that posts on Volokh Conspiracy are becoming increasingly uniform...

As per Thomas Roland's suggestion above, I propose that contributing members pair up with new co-bloggers.

Some potential pairings:

Mr. Volokh with Pat Robertson
Mr. Kopel with the Brady Center
Mr. Kerr with Howard Dean...
2.3.2006 9:49am
Richard Bellamy (mail):
This just goes to show how people of my political persuasion have their beliefs strengthened and reinforced by rational arguments, while those with opposing beliefs get mired in groupthink.

But we all already knew that, right?
2.3.2006 9:53am
Tracy Coyle (mail) (www):
I call the situation presented, acting without a governor. I am a strong conservative, my partner is a strong liberal (lesbian relationship). We tend to balance each other despite neither of us changing our overall outlook much. We both work to have friends of different interests and persuasions. Take that away and the result is clear...fanaticism.
2.3.2006 9:55am
AlexM:

The surprising part was *not* that people don't change their mind when presented with non-contradictory views. The surprising part was that people became *more* extreme when presented with non-contradictory views. It is that conclusion that I found interesting.


The first explanation that comes to mind is that highly politicized people, activists and so on, are farther away from the political center and can also be quite persuasive when dealing with like minded individuals.

People who are generally inclined to support conservative or liberal POVs but are not politically active may not have necessarily thought everything through. Their opinions may have been moderated by somewhat contradictory bits and pieces that they get from the news media. However, once they face one or more well-spoken and seemingly well-informed individuals with a coherent worldview who say things that reinforce their core preferences, they can be easily swayed farther from the center.

Just a guess, of course, but I have seen it happen on a small scale.
2.3.2006 9:58am
Medis:
Cecelius,

Here is the relevant passage from Sunstein's post:

"Citizens expressed their views in three ways: anonymously, before deliberation began; in small groups, which deliberated and tried to reach verdicts; and anonymously, after deliberation concluded. Our key question was this: What would be the effect of deliberation on people's views?

Here are our three major findings. (1) Liberals, in Boulder, became distinctly more liberal on all three issues. Conservatives, in Colorado Springs, become distinctly more conservative on all three issues. The result of deliberation was to produce extremism -- even though deliberation consisted of a brief (15 minute) exchange of facts and opinions! (I am speaking here of shifts in anonymous statements, not of shifts between individual views and group views -- though groups were also more extreme than their individual members, predeliberation.) . . ."

Sunstein isn't using "extremism" in some overall normative sense. Rather, what he is saying is that on his scale, the individually-expressed views in the groups moved toward their respective ends of the scale as a result of these "deliberations".

So, your hypothetical might be a little off. You make it sound like the "anti-segregationists" are simply agreeing on something they already believed. But in an appropriate hypothetical, according to Sunstein the views of the "anti-segregationists" would actually move as a result of the deliberations. I'm not sure what that would mean in practice, but numerically it might mean something like this: before deliberations, the "anti-segregationists" on average might have said that on a scale of 1-10, where 10 is the worst, segregation scored an 8. But after deliberation, segregation on average scored a 9. I think that would be what Sunstein means by deliberations producing "extremism".

Sturm,

Again, I think it may be misleading to say this process merely results in "consensus", although it is apparently true that the views converge (that would be Sunstein's result (3). To continue the (perhaps unfortunate) hypothetical above, if the "anti-segregationists" originally averaged 8, before deliberation, then we might have expected them to converge on an answer of 8 after deliberation. But according to Sunstein, rather than converging on 8, they will converge on something like 9. Again, I think this is the most notable nonobvious part of his research.

Zubon,

I think the basic structure of your model might make sense: people might have a mix of both more and less "extreme" beliefs on any given issue, and even beliefs on the other "side" of the scale. The effect Sunstein is observing could be the result of the more "extreme" beliefs being reinforced and the less "extreme" or contrary beliefs being weakened.

But I'm not sure why you assume the more "extreme" beliefs are "principles" and the less "extreme" or contrary beliefs are something else. It seems to me that the beliefs falling in any of the respective parts of the scale could be either the "principles" or your something elses.
2.3.2006 10:11am
Houston Lawyer:
I would agree with AlexM. People in each group who are most likely to be well informed about the arguments will tend to lead the discussion. While there are extremely well informed individuals out there with heterodox views, they tend to be a distinct minority. Most of the rest of the world has opinions that fall into either the liberal or conservative category. These views have often been arived at based upon life experience instead of persuasion by argument. Once they hear a good argument that fits within their world view, most will quickly get in line.
2.3.2006 10:19am
Richard Bellamy (mail):
While not the point of the article, does it strike anyone else as odd that beliefs about "affirmative action" or "same sex unions" should in any way be associated with one's views on "an international treaty to control global warming"?

I understand that political parties are coalitions of people with varying beliefs, so don't need to be based on a single, unifying theme. But it strikes me as odd that an individual's view on the Kyoto Protocol should tell me ANYTHING about his view on the affirmative action, let alone that it be a strong predictor.
2.3.2006 10:28am
Medis:
AlexM and HoustonLawyer,

I also think your model is plausible--those with more "extreme" beliefs are moving the other people--although I wonder if even they end up with more "extreme" beliefs.

Again, though, I question the assumption that people with more "extreme" beliefs are moving the other people because they are better informed or better speakers. Alternative explanations would include that they are more emotional, less nuanced, less tolerant of contradiction, and so on. In other words, there are lots of ways to influence beliefs, and unfortunately rational discourse is not the only such way, and maybe not even the most prominent.
2.3.2006 10:36am
ficus:
This experiment has similarities to other experiments done by psychologists that show that people are comfortable when they agree with those around them, in non-political matters. I remember one in which the subject was in a group of people who were all plants, and they were all asked to give an answer to a simple factual question of some sort, the answer to which was obvious. The participants were polled in turn, and all the plants who answered before the subject gave the wrong answer. Very often the subject also then gave the wrong answer.
2.3.2006 10:59am
Blar (mail) (www):
This is just another example of the well-studied phenomenon known as "group polarization" (Wikipedia article here, Google for more). When groups of people on the same side of an issue discuss the issue, they tend to end up with even more extreme views in that direction. There are probably several mechanisms at play, including:

- The arguments, information, opinions, and values that get shared all point in the same direction, so people get persuaded to move even further in that direction

- In order to look good/impressive/smart/reasonable/moral/etc. to the group, people are inclined to make comments that match the group's views, or are even slightly firmer &stronger statements in favor of what the group has been saying

- The people who know the most, talk the most, make the best arguments, etc. are often more extreme than the typical group member ("activists"), and they pull the group towards them

The effect does not depend on political partisanship, since group polarization effects have been found in decisions unrelated to political ideology or values (the effect was originally called "risky shift", since it was found that groups were more likely than the individuals within them to recommend a riskier option over a safer option). But the red/blue divide probably exacerbates it.

In order for group deliberation to moderate people's views, it's important to have a diverse group of people (diverse with respect to the relevant attributes - opinions, beliefs, values, etc.), with at least one dissenter in the group (or even a devil's advocate playing the role of dissenter).
2.3.2006 11:02am
AlexM:

when pressed on their fundamentals, many people will give up their "yeah but"s that make them more moderate.


That's a reasonable hypothesis and perhaps that's how it works in some cases, but in my experience the mechanism tends to be somewhat different.

At this point I should note that I am a consultant by trade and I travel a lot, from NYC to the San Francisco Bay and from Corpus Cristi to rural Alaska. My customers range from MDs to computer critters to technicians, so I get a fairly representative sample across geographical and educational boundaries. I also minored in XXth century history in a prior life, so when we start talking about politics or history, I usually get to do some gentle and, since I am generally a mild-mannered man and wouldn't want to antagonize my customers, non-confrontational edumacating.

What I have noticed is that no matter how even-handedly I present the history of the subject under discussion -- be it a 10 minute lecture on the history of the HUAC/Smith Act/McCarthy controversy or "American involvement in the Russian Civil War in 1918-1920" -- my customers tend to emphasize the facts that reinforce their pre-existing notions. Overall, conservatives tend to become more conservative and liberals tend to become more liberal once they are presented with the same set of facts.

I should also note that on average, conservatives tend to be a little more excited about these discussions and move a little more to the Right vis a vis their liberal counterparts, which I am inclined to attribute to the "See, I knew that the liberal news media didn't tell us the whole story!" factor.

Depressing, isn't it?
2.3.2006 11:07am
Hunter McDaniel (mail):
Sunstein says there were 60 people from Boulder and 60 people from Colorado Springs. Then he seemingly assumes that the 60 from Boulder are liberal and the 60 from Colorado Springs are conservative. Does he have anything with which to back up that key assumption?
I'm a conservative who lives in Boulder and has spent a lot of time in Colorado Springs. Take it from me, Sunstein's assumption is very well founded. If that's not good enough for you, compare how they voted in the 2004 election - the Springs went overwhelmingly for Bush, Boulder even more overwhelmingly for Kerry.
2.3.2006 11:29am
Medis:
AlexM,

I actually don't find the problem itself too depressing, because as Blar and others have suggested, the solution is obvious--we should foster diversity and dissent.

What I do find a bit more depressing is that so many people seem adverse to doing that. Of course, that is particularly troubling when the people in question wield power, but even in their daily lives, some people seem uninterested in exposing themselves to contrary views.
2.3.2006 11:29am
t d e (mail):

Thirdly, this explains why it is so difficult to change a "corporate mentality". For example, most members of the mainstream media are leftist,

So corporate leftists?

The mind reels.
2.3.2006 12:19pm
AlexM:

I actually don't find the problem itself too depressing, because as Blar and others have suggested, the solution is obvious--we should foster diversity and dissent.


Well, what I found depressing was that my experience suggests that people, once exposed to the same new facts, will often use them to reinforce their pre-existing notions and preferences. To use one of the examples mentioned above, it would usually go along the following lines:

Me: "You have to keep the context of the American involvement in the Russian Far East as it began in 1918 in mind. WWI was still going strong, with the Germans making their last desperate push on the Westerm front. Bolsheviks had just signed a separate peace treaty with Germany, which made Russia a semi-ally of the Central Powers. There were hundreds of millions of dollars (a lot of money at the time) worth of equipment sitting in Vladivostok, a major supply depot during WWI, that could have been transferred to Germany. Japan, America's expansionist ally/rival, had occupied the city and begun establishing control over the area. The Czechoslovak Legions had just overthrown local Bolshevik governments throughout most of Siberia and a crazy quilt of competing embryonic governments emerged in their place. Some members of the previously elected Constituent Assembly, which had been suppressed by the Bolsheviks, tried to re-established a democratic form of government in the Volga region under the Legions' protection. It was a very complicated situation."

Conservative: "I always suspected that America was doing the right hting there, trying to help establish democracy and all that, and that the PBS documentary that I saw last year was misleading! Thanks for clearing it up!"

Liberal: "Isn't it amazing how we always manage to get into other people's business even though we have only the vaguest idea of the compexities involved? And then complain that we get stuck in a quagmire? Thanks for clearing it up!"

I am not sure how being exposed to diversity and dissent would change anything in the scenario described above.
2.3.2006 12:29pm
Joshua:
A similar dynamic may also explain why, in the Middle East and elsewhere, moderate Muslims are losing the internal, intellectual battle over the nature and future course of Islam to the fundamentalists.

To invoke one of Zubon's points from earlier in the thread, the short definition of the difference between moderate and radical Muslims is that while they both derive their worldview from the same set of teachings in the same Qur'an, the radicals are much more eager to carry those teachings to their logical conclusions. Furthermore, the radicals regard the moderates' relative reluctance to do likewise as a sign of wavering faith, and let the moderates know this in no uncertain terms. The upshot is that even those moderate Muslims who aren't intellectually brought around to the radicals' view through the deliberation process are often shamed into adopting it, or at least intimidated into publicly toeing the radical line. (TCS Daily's Stephen Schwartz recently posted a piece on this phenomenon as it has played out in America.)
2.3.2006 12:50pm
AlexM:

the radicals are much more eager to carry those teachings to their logical conclusions. Furthermore, the radicals regard the moderates' relative reluctance to do likewise as a sign of wavering faith, and let the moderates know this in no uncertain terms.


Interesting. This dynamic was also quite prominent on the Left between roughly 1915 (Zimmerwald Conference) and 1939 (the Stalin-Hitler pact), especially during the early years of the Comintern before the final split between socialists and Communists.
2.3.2006 12:57pm
Medis:
AlexM,

To be clear, I wasn't proposing a general solution to the problem of perceptual biases. But it does seem that when "conservatives" or "liberals" are "deliberating" together, those effects may be moderated.
2.3.2006 1:51pm
Cecilius:
Medis,

Thanks. I see what he meant now. Seems to be a measure of 'firmness of belief' or something like that.
2.3.2006 3:28pm
Blar (mail) (www):
AlexM, what you're talking about (with the Russian intervention example) certainly happens as well. People often interpret ambiguous information in a way that's favorable to them or their side. People also tend to minimize or discredit information or arguments that seem to count against their side, while accepting those that help their side relatively uncritically. There is a famous study in which partisans who read two research reports that had used different methodologies, one giving evidence that the death penalty had a deterrent effect and one giving evidence that it did not, ended up with more extreme beliefs about capital punishment, since they were far more critical of the methodology of the study that went against their side.

This kind of polarization based on ideologically-motivated reasoning does not depend on the presence of a group, and it might even be lessened through the deliberation of a diverse group, since people are forced to listen to alternative interpretations of the ambiguous information. Still, it obviously plays a big role in political discussions.
2.3.2006 9:22pm