Jenny Toomey: We’re bring­ing up Brendan Nyhan, who is com­ing to us all the way from Dartmouth College. Brendan in his past ran with some col­leagues of his a ter­rif­ic watch­dog polit­i­cal spin called Spinsanity between 2001 and 2004, which was syn­di­cat­ed in Salon. And he has a best­seller, All the President’s Spin. Back in 2004 one of the ten best polit­i­cal books of the year. Recently he put out a report called Countering Misinformation: Tips for Journalists” which I sus­pect may have some­thing to do with what he’s about to talk about. Thank you.

Brendan Nyhan: That’s right. Thank you very much. So in my past life I was a fact-checker. So if you remem­ber the com­mer­cial from the Hair Club for Men, I’m not only the pres­i­dent, I’m also a client,” well, I’m not only an aca­d­e­m­ic, I used to do this. So I’ve expe­ri­enced first-hand the chal­lenges of try­ing to cor­rect mis­in­for­ma­tion, and in part my aca­d­e­m­ic research builds on that expe­ri­ence and tries under­stand why it was that so much of what we did at Spinsanity antag­o­nized even those peo­ple who were inter­est­ed enough to go to a fact-checking web site. So it was a very select group of people.

And at first it’s a very puz­zling phe­nom­e­non that we antag­o­nized half of our read­ers every day. And I know the oth­er fact-checkers in the room will under­stand that sit­u­a­tion. But if you think about moti­vat­ed rea­son­ing from the per­spec­tive Chris has described, it’s pre­cise­ly those peo­ple who are best able to defend their pre-existing views and who have strong views who are most like­ly to go to a site like that in the first place. And that’s what made it so difficult.

So, I’m very proud of the work that we did at Spinsanity. This was a non-partisan fact-checker that we ran from 2001 to 2004, sort of a pre­cur­sor to factcheck​.org and PolitiFact—more sort of insti­tu­tion­al­ized fact-checkers. This was two friends and me doing this in our spare time. And we even­tu­al­ly wrote a book and then decid­ed that the mod­el was unsus­tain­able and we shut it down.

A man standing before several people at a meeting table, staying, "On second thought, DON'T correct me if I'm wrong."

But it was a fan­tas­tic expe­ri­ence, and what it made me think about was why it’s so dif­fi­cult to get peo­ple to change their minds. And I think Chris has done a great job lay­ing out all the rea­sons that peo­ple don’t want to be told that they’re wrong. And let me just add to that that it’s very threat­en­ing to be told that you’re wrong. There’s a cog­ni­tive ele­ment to this and a polit­i­cal ele­ment to this, but there’s also a self-image or self-concept aspect to this that my coau­thor Jason Reifler and I have explored in our research. But I just want to under­score how threat­en­ing it can be to be told that you’re wrong about some­thing. And that the defen­sive reac­tions that threat can gen­er­ate are part of what make cor­rect­ing mis­per­cep­tion so difficult. 

So what I want to do is think about what are dif­fer­ent approach­es to try­ing to cor­rect mis­per­cep­tion. And the obvi­ous place to start and a place that I have myself called for in my writ­ing is for the media to be more aggres­sive in try­ing to fact-check mis­per­cep­tions. This is a con­fer­ence about truthi­ness in online media, but the main­stream media is still a very potent source of infor­ma­tion, very impor­tant in the polit­i­cal cul­ture of this coun­try. So what if the media were more aggres­sive in try­ing to counter misinformation? 

False Death Panel’ Rumor Has Some Familiar Roots, Jim Rutenberg and Jackie Calmes, The New York Times, August 132009

This is an exam­ple, the death pan­el sto­ry in 2009. The press was more aggres­sive in fact-checking that claim than say, they were in the run-up to the Iraq War. So, is that like­ly to be effec­tive? And what my coau­thor and I did was we actu­al­ly looked at this exper­i­men­tal­ly. We took under­grad­u­ates and we gave them mock news arti­cles where we could exper­i­men­tal­ly manip­u­late whether or not they saw the cor­rec­tive infor­ma­tion. So we have pre­cise con­trol of what they’re see­ing and we can mea­sure exact­ly what their reac­tion is to it. And the ques­tion is what hap­pens when we give them this cor­rec­tive infor­ma­tion. Is this effec­tive at get­ting them to change their minds about the giv­en fac­tu­al belief.

And the answer is gen­er­al­ly no. So, the pat­tern across the stud­ies we’ve con­duct­ed is that there’s fre­quent­ly a resis­tance to cor­rec­tions on the part of the group that’s most like­ly to want to dis­be­lieve that cor­rec­tion. This is some­thing we call dis­con­fir­ma­tion bias, and it’s very con­sis­tent with the sto­ry that that Chris described to you a moment ago.

This is a claim was made by John Kerry and John Edwards in 2004. They made state­ments sug­gest­ing that President Bush had banned all stem cell research in this coun­try. That’s not true. He lim­it­ed fed­er­al fund­ing to pre­ex­ist­ing stem cell lines. But the lan­guage that was used implied that there was a com­plete ban on stem cell research. 

So we exposed sub­jects to a mock news arti­cle about this claim, gave them a cor­rec­tion. Well, who’s like­ly to want to believe this claim? Liberals who don’t like President Bush. And what you’ll see is that our cor­rec­tion was not effec­tive at reduc­ing their report­ed lev­els of mis­per­cep­tions. Conservatives were quite hap­py to hear that President Bush had­n’t done this and to go along with it. Liberals on the oth­er hand did­n’t move. So the cor­rec­tion isn’t working.

So while that might be trou­bling enough, it gets worse. In some cas­es, cor­rec­tions don’t just fail to change peo­ple’s minds, they make the mis­per­cep­tions worse. And this is what we call the back­fire effect. We doc­u­ment two of these in our arti­cle, which I’d encour­age you to read. Here we’re talk­ing about the claim that President Bush’s tax cuts increase gov­ern­ment rev­enue, a claim that even his own econ­o­mists dis­agree with. Virtually no expert sup­port for this claim.

Again, lib­er­als, per­fect­ly hap­py to go along with a cor­rec­tion of that state­ment. Conservatives dou­ble down, in exact­ly the way that Chris describes, becom­ing vast­ly more like­ly to say that President Bush’s tax cuts increase rev­enue rather than less. So in our efforts to com­bat mis­in­for­ma­tion, if we’re not care­ful we can actu­al­ly make the prob­lem worse. And this is some­thing that every­one should think about in this room when they’re think­ing about how to address mis­in­for­ma­tion. The Hippocratic Oath of mis­in­for­ma­tion. Try not to do harm. Because you can. 

Let me just add anoth­er point here. There are also peo­ple out there who mean well and are not moti­vat­ed rea­son­ers in the way way that Chris and I have dis­cussed. And cor­rect­ing mis­per­cep­tions can still make them more mis­in­formed, too. And one mech­a­nism for this is what’s called an illu­sion of truth effect. So this is from a CDC fli­er of facts and myths about the flu. This is not some­thing that peo­ple have strong pre­ex­ist­ing beliefs about in the same sense as their pol­i­tics, right. Most peo­ple are not epi­demi­ol­o­gists and experts in the char­ac­ter­is­tics of the flu. So we tell them some things and we say, These ones are true and these ones are false and here’s why.”

But when some psy­chol­o­gists showed peo­ple this and then divid­ed the ones who saw this— Now, some of them only got the true state­ments, some got the true and the false. And then they looked at how well they retained this infor­ma­tion. What they did was they divid­ed those folks and gave a thirty-minute delay for some peo­ple. And after just thir­ty min­utes, a sig­nif­i­cant per­cent­age of peo­ple start to mis­re­mem­ber the false state­ments as true. This is a well-documented phe­nom­e­non in psy­chol­o­gy where famil­iar claims start to seem true over time. So again, in try­ing to cor­rect the mis­per­cep­tion, you’ve made the claim more famil­iar, and that famil­iar­i­ty makes peo­ple more like­ly to mis­re­port these state­ments as true. 

So again, these are folks who have no par­tic­u­lar dog in this fight. So again we have to be very cau­tious about the steps we take. And again let me just under­score that the rea­son we can tell that these effects are hap­pen­ing is because we’re test­ing them exper­i­men­tal­ly. That gives us full con­trol and abil­i­ty to dis­en­tan­gle exact­ly what’s going on, which is very very dif­fi­cult otherwise.

A final note about cor­rect­ing the prob­lem. The oth­er thing I would just say is while we can talk about fan­tas­tic tools we could devel­op to help cor­rect mis­in­for­ma­tion, the prob­lem we have is that the folks who seek those tools out may be those whom we’re least inter­est­ed in reach­ing, because they may already believe what we want them to believe in any par­tic­u­lar case. And even for them when we do chal­lenge them, coun­ter­at­ti­tu­di­nal infor­ma­tion is only a click away. 

This is a snap­shot of the results I got when I typed Obama birth cer­tifi­cate” into Google this morn­ing. Let’s say I want to believe Obama’s birth cer­tifi­cate is real. Well, I can click on the Snopes debunk­ing, but it’s sur­round­ed by a head­line that says it could be a forgery. And news about Sheriff Joe Arpaio’s press con­fer­ence. So, choose your own adven­ture. Which one do you want? It’s very easy to seek out sup­port­ing infor­ma­tion for what­ev­er point of view you want to defend. So when we do chal­lenge peo­ple, the tech­nol­o­gy that we have makes it eas­i­er and eas­i­er to reach out and but­tress that view that’s been chal­lenged. So again this is a very dif­fi­cult chal­lenge. It’s much eas­i­er to but­tress those view now than ever before. 

Now that I’ve depressed you, let me talk about things we can do that are per­haps bet­ter approach­es. And let me just say that these are part of a New America Foundation report [sum­ma­ry] that my coau­thor Jason Reifler and I wrote with help of sev­er­al peo­ple in this room—it’s avail­able on the table out there. And there’s a cou­ple of oth­er reports that are part of that pack­age about the fact-checking move­ment. But these are series of rec­om­men­da­tions that we’ve devel­oped based on avail­able research in psy­chol­o­gy and polit­i­cal sci­ence and com­mu­ni­ca­tion, on how you can com­mu­ni­cate in a more effec­tive man­ner that’s less like­ly to rein­force misperception.

The first thing. This is what not to do, okay. Remember I told you about that illu­sion of truth effect where see­ing the false claim and it becom­ing famil­iar to you makes you more like­ly to mis­re­mem­ber it as true. This Politico arti­cle in its sixth or sev­enth para­graph even­tu­al­ly gets around to say­ing actu­al­ly, there’s extreme­ly strong evi­dence that Obama’s a cit­i­zen and this is all non­sense.” But if you look at the top of the arti­cle, which is what’s excerpt­ed here, what it’s done is it’s restat­ed that claim both in the head­line and the lead state­ment. And by restat­ing these claims again and mak­ing them more and more famil­iar, we’re actu­al­ly like­ly to cre­ate this flu­en­cy that caus­es peo­ple to mis­re­mem­ber these state­ments as true. So when I say best prac­tices here, this is what not to do. And I have an arti­cle on the Columbia Journalism Review blog about this if you’re inter­est­ed in read­ing more about this problem.

Another prob­lem, nega­tions. Again, there are well-intentioned peo­ple who are con­fused some­times. We’ve often been talk­ing about moti­vat­ed rea­son­ing and peo­ple who don’t want to be con­vinced. But even those peo­ple who are open-minded can have a tough time with nega­tions. When we try to say some­thing is not true, we may end up rein­forc­ing that claim we’re try­ing to invalidate. 

So this is an exam­ple of some well-intentioned folks try­ing to debunk, to dis­cred­it a claim that MMR caus­es autism. The prob­lem is they have MMR…cause autism” in the head­line. You stare at that long enough and peo­ple will start to get ner­vous. And my coau­thors and I have have done an exper­i­men­tal study find­ing sim­i­lar effects, that try­ing to cor­rect the MMR/autism asso­ci­a­tion can actu­al­ly make peo­ple less like­ly to vac­ci­nate rather more. 

A third rec­om­men­da­tion. And this is real­ly for the jour­nal­ists in the room. But the notion of of arti­fi­cial bal­ance, in which each side has to be equal­ly rep­re­sent­ed in fac­tu­al debates, has been shown to mis­lead peo­ple quite a bit. So Jon Krosnick in Stanford and his col­leagues have a study show­ing that pro­vid­ing a bal­anced report in the sense of one sci­en­tist who says glob­al warm­ing will have destruc­tive con­se­quences and one who says it’s great—the plan­et’s nice and warm, (which is this guy’s mes­sage)… Providing that bal­anced point of view is dra­mat­i­cal­ly changes how peo­ple inter­pret the sci­en­tif­ic evi­dence. And I don’t want to pick on this par­tic­u­lar debate, but just to say that report­ing needs to reflect the bal­ance of the evi­dence. And the he said–she said per­spec­tive that Kathleen men­tioned ear­li­er that leads to this sort of quo­ta­tion of fringe sources to pro­vide bal­ance can real­ly mis­lead peo­ple. And that’s some­thing to be avoid­ed when­ev­er possible.

Another rec­om­men­da­tion. Graphics. People are real­ly good at argu­ing against tex­tu­al infor­ma­tion. At least in the exper­i­ments we’ve con­duct­ed, graph­ics seem to be more effec­tive for those quan­ti­ties that are… Let me say that a dif­fer­ent way. For mis­per­cep­tions that can be graphed, right. There’s lots of mis­per­cep­tions we can’t graph. I don’t have a graph of Iraq’s weapons of mass destruc­tion. I can’t put one up for you. But what I can do is show you that four dif­fer­ent inde­pen­dent sets of tem­per­a­ture data show dra­mat­i­cal­ly increas­ing glob­al tem­per­a­tures and are extreme­ly highly-correlated. This is from a NASA press release. We found this was quite effec­tive, much more effec­tive than equiv­a­lent tex­tu­al cor­rec­tion, at chang­ing peo­ple’s beliefs about glob­al warming. 

Experts Debunk Health Care Reform Bill’s Death Panel’ Rule, Kate Snow, John Gever, Dan Childs, ABC News, August 112009

Another approach, and this is some­thing I don’t think we have talked much about so far. But it’s impor­tant think about cred­i­ble sources to peo­ple who don’t want to be con­vinced. This is an exam­ple of what I thought was an exem­plary report on the death pan­el con­tro­ver­sy from ABC news that said— It’s framed as experts, right, doc­tors agree that death pan­els aren’t true. So it’s going to med­ical exper­tise, it’s get­ting out of the polit­i­cal domain, and it’s say­ing that even experts who do not sup­port the ver­sion of the health­care reform bill being pro­posed by President Obama agree that death pan­els aren’t in the bill. And it goes on to quote a Republican econ­o­mist with impec­ca­ble cre­den­tials on that side of the aisle say­ing there’s lots to oppose about this bill but death pan­els isn’t one of them. And to the extent that we can reach out and find those more cred­i­ble sources to peo­ple who aren’t will­ing to lis­ten to the nor­mal peo­ple who are typ­i­cal­ly offered to them, that may be an espe­cial­ly effec­tive approach. 

So for more I’d com­mend to you the report that I men­tioned ear­li­er as well as those by my col­leagues about the fact-checking move­ment. Thanks very much.