« Now that some law schools are really closing, what will make other universities pull the trigger? | Main | More civ pro in SCOTUS »

Thursday, January 19, 2017

Sunstein on political polarization and "The Divided States of America"

The renowned Cass Sunstein is delivering the Rosenthal Lectures at NorthwesternLaw  this fall.  His topic is political polarization, its causes and effects.  In his first lecture yesterday, entitled "Whose Facts?," Sunstein presented the fascinating fruits of some extensive experimental work on public opinion and its dynamics.  

The central question is whether and to what extent individuals with views on a factual matter -- and Cass is here interested in important issues which bear on public policy, e.g., climate change, gun control, immigration, etc. -- will update their opinions in light of reported new facts, facts which reveal either good news or bad news.  So, take some who has a strong view on the matter of human agency and climate change.  When presented in an experimental setting with "evidence" that scientific consensus is that the climate is warming at  less than previously reported, would these "strong view" folks adjust their priors in a weaker or stronger direction?  Ditto those with a "weak" view (let's call them climate change deniers) or a "moderate" view.

A plausible hypothesis tested by these experiments, and one very much in line with the classic story of Bayesian updating, is that supportive info will help confirm their biases and info inconsistent with their prior beliefs will help folks update their views.  In short, facts matter.  And out of the marketplace of ideas should come a dynamic process in which folks refine and reshape their views as new information is generated and disseminated.  But Cass and his colleagues find a very different story at work.  Alas, the incorporation of new facts into their worldview is asymmetrical.  That is, folks will disregard to a great extent "bad news" for their prior views.  So, info that the climate is warming faster will not shape the convictions of climate change deniers; it may well in fact strengthen them in their beliefs.  (Cass quotes George Lucas commenting on a Star Wars dispute:  "I don't like this fact, and I don't believe it").  In short, non-confirming facts will push them further in their direction of their prior beliefs, and this will be true for folks on both sides of the ideological ledger.

This is a startling and rather dispiriting result.  It tells us that committed folks are not only inured to facts, but that the facts which reflect bad news for their beliefs (even if good news for the world) will augment their convictions.  And this will reinforce patterns which generate more polarization and all the bad impacts that such polarization represents.

In his second lecture, Sunstein promises to reflect on the topic "Whose Values?" and assures us that there is optimism to be found in the third and final lecture.   

Posted by Dan Rodriguez on January 19, 2017 at 11:44 AM | Permalink


I wonder if there will be information on the role of social media (e.g. Twitter) in creating and reinforcing such opinion. If so, could this result from application of Sunnstein's earlier work on Cascades? Did Sunnstein actively try to apply his cascades theory to support public support for government policy by weaponizing social media? If this subject arises I would love to read a report on it.


Posted by: ROBERT JOSEPH BLAISE MACLEAN | Jan 19, 2017 1:02:41 PM

Empirical researchers have long struggled with the effect Sunstein describes, and I wonder if he mentioned the old truism in his speech at some point: The two most common reactions to empirical data analysis are (1) we already knew that, and (2) that can't be true. This is among the many things that Jay Westbrook and the Consumer Bankruptcy Project researchers have told me that have really stuck in my head over the years.

Posted by: Jason Kilborn | Jan 19, 2017 3:09:32 PM

This is conflict narrative research. It's been going on for some time. It's hardly news.

Posted by: YesterdayIKilledAMammoth | Jan 19, 2017 3:24:48 PM

This data conflicts with what I believe about the influence of new facts, which is why I am glad that the data doesn't actually exist.

Posted by: anon | Jan 19, 2017 3:28:53 PM

"It's hardly news."

Well, damn, someone posted something on a blog for academics that isn't news! That's nearly as bad as someone posting something on 4chan that isn't art.

Posted by: Anon | Jan 19, 2017 6:10:45 PM

When something is called "experimental" that is just the baseline for research that has been ongoing (and makes no changes or observations about that baseline) it seems worthwhile to point out that it's not news and say where the research has been done.

Someone might say, "Oh, what does the research on conflict narratives say?"

But don't let that get in the way or your puerile mewling.

Posted by: YesterdayIKilledAMammoth | Jan 19, 2017 8:06:45 PM

When a stage magician pulls a rabbit out of a hat (or some other impossible feat), rational people don't adjust their notions of where rabbits come from based on the new evidence presented to them. They reject the evidence and remain committed to their previous worldview. This is normal and appropriate behavior, even if it can make things difficult for academics trying to solicit funds from the public coffers for climate modeling or whatever social science hypothesis is in fashion at the moment. The notion that the experimenter can remove himself, and the nature of his authority (or lack thereof), from the experimental setting is one of the great conceits of the social sciences. So none of this research proves anything beyond the fact many interesting notions in the social sciences are not the falsifiable hypotheses we would like them to be.

Posted by: M. Rad. | Jan 21, 2017 10:21:14 PM

M.Rad: Huh? I suppose it's the stuff like this, along with today's paean by K. Conway to so-called "alternative facts," that is our new normal. Sigh. Unless, hope against hope, this last comment was intended as a parody.

Posted by: dan rodriguez | Jan 22, 2017 11:48:28 AM

Post a comment