Rumors have it: Trying to correct political myths may only entrench them further

May 14, 2015

Illustration: Jose-Luis Olivares/MIT

By Science Daily

Bad news, fans of rational political discourse: A study by an MIT researcher shows that attempts to debunk political rumors may only reinforce their strength.

“Rumors are sticky,” says Adam Berinsky, a professor of political science at MIT, and author of a paper detailing the study. “Corrections are difficult, and in some cases can even make the problem worse.”

More specifically, Berinsky found in an experiment concerning the Affordable Care Act (ACA) that rebuttals of political rumors about the supposed existence of “death panels” sometimes increased belief in the myth among the public.

“Pure repetition, we know from psychology, makes information more powerful,” Berinsky says.


Read the full article by clicking the name of the source located below.

22 comments on “Rumors have it: Trying to correct political myths may only entrench them further

  • “It’s not that there are some people who believe a lot of crazy things,” Berinsky says. “There are a lot of people who believe some crazy things.”

    Well outside politics also.



    Report abuse

  • I read an article a few months back that education about the safety of vaccines just reinforced beliefs that vaccines were dangerous. This is just confirmation bias on steroids.



    Report abuse

  • I would suggest a minor amendment to Berinsky’s statement (which is probably implied):

    “There are some people who believe a lot of crazy things; there are also a lot of people who believe some crazy things.”



    Report abuse

  • A reason for the results, Berinsky surmises, is apparently minimal belief, among contemporary voters, in the existence of neutral sources of information.

    This strikes me as the key issue. Who do we trust when it comes to political issues: politicians, interest groups, the media, lobbyists, academics, experts…



    Report abuse

  • Who do we trust when it comes to political issues: politicians, interest groups, the media, lobbyists, academics, experts…

    Trust no one. And hope that you were lucky enough to have received an education that allows you to make rational evidence based assessments of the material put before you.

    If you adhere to an ideology, you’ve failed before you even wake up. Socialism. Capitalism. They’re all the same. Your ideology has already decided for you what your response will be. All your answers are predetermined. You are prevented from making a decision which follows evidence, because it conflicts with your ideology. You become just another follower. Just saying.



    Report abuse

  • Why? I’ve studied the main arguments for both sides and I am convinced that global climate change is being caused by man, as that seems to be far the most parsimonious reason. Occam’s razor and all that!



    Report abuse

  • Would the results change if you tested the participants knowledge of the relevant subjects and let them know how they did?

    What if you tested them on a completely unrelated subject?



    Report abuse

  • Well said David.

    Ideologists frequently claim to be non-dogmatic, yet they just as frequently fail to demonstrate free thinking. For evidence: see All Politicians.

    It is next to impossible to be permenantly skeptical (I’ve tried it) so, if you will allow, I propose a modified version of trust no-one. Rather; investigate, trust and verify.

    All politicians are suspicious even if only for one reason; that they want to be in charge of you, and me, and as many other people as they can get away with.

    Nevertheless, it would be going too far to claim that all politicians are sly, stupid, snakes-in-the-grass … all the time. Even political leaders struggle to achieve that level of scumbag.

    But politicians are only human (another reason to exercise our caution), and unless we are paragons of virtue we should remember that – particularly if we live with a measure of democracy – it’s our job to keep them in line. It’s a job of work. I requires that we actually use our time, energy and nous.

    There is one very special reason why we should thank Rupert Murdoch. His media stable, the standout example being Faux News, is a living, daily, testimony to the failure of the public to recognise that the Old Media (TV, Radio, Newspapers) are also frequently partisan political movements. Unelected, unapologetic and underhand, they use exactly the tactics that are outlined in the OP.

    And, if I may make so bold, I would like to slightly modify your ” … you were lucky enough to have received an education that allows you to make rational evidence based assessments … ”

    It’s never too late to learn.

    Peace.



    Report abuse

  • Berinsky’s experiment also produced new data about the attachment of the electorate to myths in general. He asked respondents whether they believed in any or all of seven different myths, six of which concerned politics — such as the myth that President Barack Obama is a Muslim, or the rumor that vote fraud in Ohio swung the 2004 presidential election to then-President George W. Bush.

    One of these things is not like the other…

    OK, so maybe it’s my own confirmation bias at work here, but this “rumor” is actually supported by a whole lot of actual evidence. I’m not saying it’s necessarily true, but there has certainly been enough evidence presented to elevate it above a mere “rumor.” There’s no evidence to support that the A.C.A provides for death panels (and clear evidence to the contrary), therefore it’s just a rumor. There’s no evidence to support that Obama is a closet Muslim (and clear evidence to the contrary), therefore it’s just a rumor. There is a mountain of evidence regarding voter fraud in Ohio during the 2004 presidential election, however. Whether that evidence is sufficient to prove that Bush “stole” the election is unknown (to me, at least), but it’s certainly not just a “rumor.”



    Report abuse

  • I need help from someone to clarify the methodology of the experiment. Anyone .

    Berinsky conducted the experiment with three separate waves of public-opinion surveys in 2010, encompassing nearly 2,000 voters. The surveys tested multiple methods of debunking rumors about the ACA on multiple categories of respondents.
    Among “attentive” voters who responded to the survey, 57 percent initially rejected the rumors about “death panels.” After seeing a nonpartisan correction of the myth, those figures changed slightly, to a 60 percent rejection rate. But a Republican-sourced correction raised that rate to 69 percent, while a Democratic-based correction only led to a 60 percent rejection rate.
    In each scenario, the acceptance rate of the rumor went up, while only the Republican-based debunking markedly changed the rejection rate. Other slices of the survey group yielded similar patterns in which both acceptance and rejection rates rose, with Democratic-sourced debunking efforts being the least effective.
    When Berinsky conducted follow-up surveys of voters weeks later, he found that about 43 percent of all respondents (not just “attentive” ones) rejected the rumor. A nonpartisan debunking of the rumor produced a 51 percent rejection rate, but a debunking stemming from Republican sources yielded a 58 percent rejection rate, while a Democratic correction led to 53 percent of people rejecting the rumor.

    I do not understand why both the rate of rejection and acceptance went up. The data shows that in the two different trials the rejection rate only went up. I understand that in the two different trials the first trial included only attentive voters, presumably an unspecified minority share of all voters in a total sample of 2,000. The second trial indeed shows a “drop” in in the rejection rate against a different base of all voters compared to the base of attentive voters (57% versus 43%) but ultimately two different groups are being tested even if attentive voters are subsumed within a category (“all respondents”) of all voters In terms of methodology I fail to understand how the hypothesis is proved since two different groups are being tested while the statistics only show a rise in rejection rates for both trials. I acknowledge I’m probably missing something that someone can clarify. Anyone?



    Report abuse

  • Hi Melvin, I don’t see where the article states that the first survey included only attentive voters.

    Unfortunately the author doesn’t give us the numbers to support the claim in the title. Instead it’s merely said that:

    “In each scenario, the acceptance rate of the rumor went up, while only the Republican-based debunking markedly changed the rejection rate. Other slices of the survey group yielded similar patterns in which both acceptance and rejection rates rose, with Democratic-sourced debunking efforts being the least effective.”



    Report abuse

  • I’m sorry, I see you quoted that exact section in your post.

    My guess is that participants had more than just (accept | reject) options and that in both scenarios some moved from one of the others to acceptance.



    Report abuse

  • Thanks Sean. Here’s how I interpreted the numbers in the probable absence of fluctuating data. For convenience, a base of 100 (100%) of respondents will be used.

    Attentive voters were identified and separated from the total. (The researchers are apparently breaking out two separate groups for measurement??) Presumably “attentive voters” were identified by likeliness to vote; by greater interest in the political process. In the U.S, such a group would be a minority let’s say plausibly a little more than a third or 36% of the total. The remainder fell into a presumed category of inattentive respondents or 64% (36% + 64% = 100%)
    Therefore: 36 people out of a hundred are attentive voters
    64 people out of a hundred are inattentive respondents

    First Polling: attentive voters (36%) only:
    57% of attentive voter group only rejected the rumor or .57 x 36 = 21 people out of 100.
    Depending on nonpartisan, Democrat, or Republican debunking sources the rejection rate went up between 3% (nonpartisan or Democrat sources) and 12% (Republican). Adjusting for error, let’s say the rejection rate increased and held stable at about 65% or .65 X 36 = 23 people out of 100.

    Second Polling: All Respondents (100%) ( Attentive Voters and Inattentive respondents)
    43% of the total rejected the rumor or .43 X 100 = 43 people out of 100. All we can infer from the first attentive voter scenario is that about 23 people (.65 X 36 = 23) “should have” held fast to rumor rejection. Using the presumptive majority of inattentive respondents at 100-36 =64, we can calculate the rumor rejection rate for this group. Adjusting for error, all respondents showed a combined net rejection rate increase from 43% to 55% or 55 people out of 100. Therefore: 55-23 = 32. 32/64= 50 or 50% of inattentive respondents. came to reject the rumor.

    Summary:

    Total Sample 100
    36% attentive voters or 36 people out of 100
    64% inattentive respondents or 64 people out of 100
    After reading debunking material, rate of rejection increases from 57% adjusted for error to 65%
    for attentive voters group or .65 X 36 = 23 people out of 100
    After reading debunking material in subsequent testing, rate of rejection increases from 43% adjusted for error to 55% for entire sample of 100 (both attentive voters and inattentive respondents)
    or .55 X 100 = 55.
    55- 23 (attentive voters) = 32 inattentive respondents for a rejection rate of 50% or 32/64 = 50% for inattentive respondents versus a rejection rate of 65% for attentive voters.

    The authors intend to show: “In each scenario, the acceptance rate of the rumor went up, while only the Republican-based debunking markedly changed the rejection rate.”…My reading finds evidence that shows only the rejection rate not the acceptance rate went up however more or less. I freely admit I’m probably the victim of a stumbling mental block in interpreting the data. I welcome clarification of what the researchers actually measured based on how their findings are stated in the article.



    Report abuse

  • I think the article just doesn’t give the numbers for the increase in acceptance. The numbers it does give are used to back another claim: the efficacy of the rebuttals was dependent on their source.

    What those numbers don’t show is that the acceptance rate did not also grow. I gave one reason why we can’t assume from the numbers given that the study doesn’t support what’s claimed in this piece.

    We just have to take the author’s word for it that acceptance of myths also increased, and I guess if you’re still interested read the paper when it’s published.



    Report abuse

  • Sean. Yep. I suspect you’re right… I suppose the Republican debunking source lends more credibility to the denial of a false rumor that works to Republican political advantage. The average person might be inclined to think, “If [some] Republicans have enough integrity to debunk a false claim that serves their interests, then they must be more likely than Democrat or “neutral” sources to be telling the truth.”



    Report abuse

  • “(I)nvestigate, trust and verify.”

    How? How will you know what data to trust? Which researchers and experts to believe? What conclusions to trust? Each side of a debate has its experts in support of their respective positions. Which expert will you trust? A recent column on this website (“The G.O.P.’s War on Science Gets Worse” https://www.richarddawkins.net/2015/05/the-g-o-p-s-war-on-science-gets-worse/“shows that you can’t argue a point if you can’t gather the data you need to support your point. It also shows that few people are really going to honestly and fairly weigh both sides of a debate (yes, this especially includes me).

    Consider this: if someone were to present convincing evidence that global warming isn’t taking place, that evolution never took place, that vaccines do, in fact, increase the chances that children will develop autism, that smoking has little effect on the rates of cancer and heart disease, do you honestly believe that you would change your mind on any of those subjects? Wouldn’t you more likely say, “Nope, fake evidence, shoddy methodology, and ignorant pseudoscientists”?

    Our ability to change our minds doesn’t depend on fairly considering all sides of an issue. It depends on how credible we believe the apologists for a position are. Especially if they support our own beliefs.



    Report abuse

  • Are you sure that the evidence that shows that Obama isn’t “a closet Muslim”, that the A.C.A. doesn’t “(provide) for death panels”, and that “there’s … a mountain of evidence of voter fraud in Ohio…” is really real?

    That’s the point behind confirmation bias. We place more value (even subconsciously) on evidence that supports our own beliefs than on the “wrong” beliefs.



    Report abuse

  • @OP – “Rumors are sticky,” says Adam Berinsky, a professor of political science at MIT, and author of a paper detailing the study. “Corrections are difficult, and in some cases can even make the problem worse.”

    The error is in allowing the propagandists/deniers/rumour-mongers, to set the agenda, and then trying to refute their outlandish claims.

    Start with the evidence and the sources, and you quickly discover the sponsored rumour-mongers, merchants of doubt, and charlatan liars, who have no substantiated sources! – Merely lies, quote-mining, and cherry-picked data presented to mislead, by people posing as honest expert “authorities”!

    The fan clubs of quacks are likely to be unimpressed with reason, but they cannot reason their way out of views they did not reason their way into. –
    Many have no concept of logical reasoning anyway.
    To them “logic”, is just another authority badge word, to stick on to their own unevidenced and irrational views.



    Report abuse

  • I go straight to the legislation… and then read it. Find the facts! The suggestion that a death squad amendment will be signed in if “they” win again is just laughable and should immediately be dismissed as rubbish, or investigated and then dismissed as rubbish.

    We’ve been conditioned to believe the system is out to get us….!



    Report abuse

Leave a Reply

View our comment policy.