Skip to content

Breaking the feedback loop: When people don’t correct their errors

OK, so here’s the pattern:

1. Someone makes a public statement with an error, an error that advances some political or personal agenda.

2. Some other people point out the error.

3. The original author refuses to apologize, or correct the error, or thank people for pointing out the error, and sometimes they don’t even acknowledge the correction in any way.

It’s happened in Perspectives on Psychological Science, it’s happened in the American Journal of Hypertension, it’s happened in the New York Times (of course), and, hey! Politicians do it too!

German Lopez reports:

Democratic presidential candidates Elizabeth Warren and Kamala Harris marked the five-year anniversary of the Ferguson, Missouri, police shooting of Michael Brown last week with tweets claiming that the cop who shot Brown “murdered” the 18-year-old black man.

But the evidence, including a report released by President Barack Obama’s Department of Justice, says otherwise. . . .

Between now and the time this post appears, Warren or Harris might issue corrections and apologies. But I doubt it. After all, Al Sharpton never apologized for attacking that prosecutor, Donald Trump never apologized for the things he said about 9/11, etc. My point here is not “false equivalence,” just that it’s considered standard operating procedure for people to not correct their errors, even when they’re making false statements about clearly-identified people, events, or statistics. The original statement could be an honest mistake, a reasonable misunderstanding, an opportunistic bit of exaggeration, a flat-out lie, or something in between . . . whatever. The point is that the statement is clearly wrong, the evidence is right there in front of these people, but they still won’t admit the error.

It’s a sad day when political figures on both sides of the aisle are behaving as badly as Ivy League professors and New York Times columnists. What’s the world coming to?

All jokes aside, I think this is a big deal. The self-correcting nature of science, or of politics, is a lot harder to occur when powerful figures in the system refuse to self-correct. It also provides a bad example for others and promotes an our-team-versus-their-team attitude.

I understand that this is going to happen—it’s human nature to lie—or, if not to lie, than to accept questionable claims that agree with our predisposed notions—and then to not admit the error. But at least we should call people out on it, also we should be suspicious of other claims made by people who are so comfortable with being associated with untruths.


  1. Anon says:

    Great post. Andrew. I do worry. Now, not even the ‘unbiased’ Nate Silver corrects his errors.. but blames it on the media or whomever he can find on twitter. Is there anything anyone can do at this point for institutional restoration?

    • Andrew says:


      I think/hope that Nate corrects his errors on his website, and that it’s just on twitter that he does the duck-and-blame thing. Overall, I think Nate’s excellent at admitting and correcting errors: that’s part of his brand identity, as it were, and it’s a good thing.

      • Anon says:

        You were just writing here a week ago that he hasn’t corrected himself..

        • Andrew says:


          I think of Nate-on-his-website and Nate-on-twitter as two different people. On his website, Nate is an analyst. On twitter, he’s a pundit.

          • Anon says:

            I’m sorry but no.. It doesn’t work like that.

            • Mark Webster says:

              The difference is that the people Andrew is referring to above aren’t apologising about mistakes made in their professional work, the place where they should feel most obliged to do so. While Silver might write nonsense on Twitter and not take it back, he does write articles about where he/538 went wrong on modelling something, i.e. his professional work.

              • Anon says:

                This idea that if you are a journalist and take a position on Twitter it doesn’t matter is insane. You don’t get to just dismiss his ~terrible~ opinions on Twitter as somehow a separate entity from him or not his professional work. It may even be more people read his Tweets than his website, given how news and information are consumed today. This is really sloppy thinking on the part of you and Andrew.

              • Andrew says:


                I don’t think it’s sloppy thinking; it’s just the way it is. Nate does excellent work on his website as an analyst, and he also has a sideline on twitter as a replacement-level pundit. I agree with you that this is a bad thing: my problem with Nate’s punditry is not that it’s mediocre but that he says lots of things that he’s not willing or able to defend.

              • Zhou Fang says:

                People use Twitter in various ways. Many people use it as a hole in the internet to vent into. For example many journalists will express opinions on political candidates on Twitter. It does not mean their professional work endorses that candidate from that moment forwards. Some separation exists.

              • Bob says:

                Your character is your character. If you are prepared to spout garbage for the benefit of the peanut gallery, or because it gives your side a short term tactical advantage, you are a liar and a charlatan.

                There’s no get out clause because it’s not your ‘professional work’. A liar is a liar.

  2. zbicyclist says:

    We have to be a bit careful here, because even official tribunals don’t establish absolute truth, and there can be room for opinion. I, for one, believe O.J. Simpson is guilty, although he was officially declared not guilty in his criminal trial.

    (yes, I know there were civil cases, so this isn’t a perfect example)

    One of the great things about the internet is that news stories CAN be updated in the archives, so the correction is included with the original article. This is a vast improvement over printing a correction in a later edition on the back pages.

    I wonder — do journals follow this practice? When an article is retracted, is the retraction noted prominently in the online version of the article?

    • Charles Stewart says:

      No, they do not. Cf. things have not improved much since then.

      Retracted articles continue to accrue citations.

    • Brent Hutto says:

      It wouldn’t matter anyway. People don’t actually read the articles they cite.

      I’m thinking of one paper I read early in my career, examining the validity of a certain survey instrument. The conclusion was quite clear, there no evidence of validity. The answers respondents gave to those survey items had little if any association with the criterion measure. The paper did not claim validity, it conclude the measure was not valid.

      That paper has been cited many times over the years by other studies using that measure. Clearly, none of those authors bothered to read the paper or even the abstract. The title was “Validation of the so-and-so instrument…” so they simply list the citation as evidence of validity and move on.

  3. Zhou Fang says:

    From an incentives point of view, people are almost never rewarded for admitting error. So why do it?

  4. Terry says:

    The depressing part of all this is that when facts no longer matter, and when people can’t be shamed for lying, we lose a common political ground, and we are left with raw tribalism and identity politics.

    That’s why Warren and Harris won’t admit they lied because they are sending a signal that “I am so completely on your side that I am willing to tell nakedly racist lies.” The lies are not inadvertent. They are deliberate signals.

    • Andrew says:


      I don’t know anyone in these campaigns so I can’t be sure, but I’m guessing that the problem is not that Warren and Harris are lying on purpose, as a signal, but rather that they don’t really care about the distinction between truth and lie. It’s similar to our experiences with the pizzagate guy, the why we sleep researcher, the ovulation and clothing researchers, etc.: They get things wrong (not necessarily “lying,” maybe they’re just making mistakes and not being careful) and they don’t really care.

      They have bigger fish to fry. Warren and Harris want to be president, or to influence national policy, etc.; that’s more important than the details of what some police officer did somewhere one day. Similarly, those researchers want to make and communicate big discoveries (as they see it), and that’s more important than details like how many slices of pizza did those people eat or what are the actual dates of peak fertility.

      From my perspective as a statistician, if you don’t get the details right, you’re dead on arrival and you can’t expect to have any hope of learning. From their perspective as policy makers, or would-be policy makers, or would-be scientific innovators, what matters is the big picture. For them, what’s important is telling a good story; the details don’t really matter, and people like me are at best nitpickers and at worst Stasi terrorists who want to ruin everyone’s party.

  5. Terry says:

    I’m guessing that the problem is not that Warren and Harris are lying on purpose, as a signal, but rather that they don’t really care about the distinction between truth and lie

    Could very well be, although it is a pretty fine distinction, and the two are nearly indistinguishable.

    …that’s more important than the details of what some police officer did somewhere one day

    No. The Ferguson lie is a proxy for a much larger dishonest narrative. When Warren and Harris were ballyhooing the Ferguson hoax five years ago, they certainly did not portray it as a trivial detail of what some police officer did somewhere. Many careers and votes depend on maintaining that narrative.

    Similarly, those researchers want to make and communicate big discoveries (as they see it), and that’s more important than details like how many slices of pizza did those people eat or what are the actual dates of peak fertility.

    You are being very kind to those researchers. But, I suppose it could be true of some of them, perhaps.

    • Dzhaughn says:

      It’s not so find a distiction. A phrase has many meanings. Saying “X” about Ferguson is (1) a statement of fact, and (2) an implicit claim about the attitude she would bring to governing. Sense (1) being obviously false commits her more to the claim (2.) She is saying she will govern as if it was a actually a cold-blooded murder. And I believe her. Sorta. I expect she would have done something showy along those lines.

      The current President of the USA understands this dynamic very well; you can see how much of the mocking might help him in another direction. He has the advantage (that Warren utterly lacks) that his supporters do not expect much of him as President.

    • Steve says:

      I think the distinction that Andrew is making is real and important. The philosopher John Austin made the distinction between locutionary, illocutionary and perlocutionary speech acts, roughly the act of conveying information, the act of conveying information with a particular force (e.g., asking a question, making a command), and the act of saying something to reach some particular end (e.g., making someone give you something or believe you). When politicians, salespeople and trial lawyers use language the perlocutionary act has nothing to do with the truth of the claims they are making. Sometimes a good strategy might be to convince people of a particular view or the facts, but that is often a terrible strategy to convince people to act as you want. So, we cannot expect politicians to have a primary focus on getting the facts right. But, when I see someone like Warren lie about the Brown killing, it strikes me as just lazy. There are ways to convey what she needs to convey, i.e., that she cares about police violence, without lying. We tend to belittle politicians for making carefully crafted evasive statements, but we should value it. They cannot change peoples minds on the facts in the time frame of an election, but if they can avoid outright lies that is about the best we should expect. Scientists on the other hand, should have as their primary goal advance scientific knowledge.

      • Andrew says:


        You write, “We tend to belittle politicians for making carefully crafted evasive statements, but we should value it.” That’s an interesting point that I hadn’t thought of. Alternatively, politicians could speak freely and then correct their mistakes, but I guess it’s hard for them to do that. When I post on this blog, I do make some stupid mistakes but usually I state things pretty carefully and avoid errors. That takes some effort, though: believe it or not, I read over these things before posting! If I were being interviewed by reporters every day, I’d make lots of mistakes, and I’d want to be correcting myself all the time.

        • Steve says:

          But, your purpose is to get to the truth, to expand our knowledge. But, if your purpose was something different, than you might evaluate statements differently. If you are coaching a basketball game and at half-time your team is down 50 points, your half-time speech is meant to encourage them not to give them a realistic assessment of the odds of victory. I am NOT advocating lying, but for many forms of communications, the goal is not truth and so like you said, the speaker may not really care about the truth of his/her statements. If your half-time speech got the team to come back and show themselves that they can compete, you would think you did a good job regardless of the factual content. In some instances of speech, the truth is just not the point. (Like I said, Warren was lazy to repeat the lie, but her point was to show solidarity with a group. She could have accomplished that without lying, but correcting herself now would undo her purpose. That’s very different than a journalist or scientist for whom the purpose is (should be) the truth.

      • Terry says:

        You have some good points.

        My point is that perlocution produces a string of false, dishonest, and sometimes true statements. Mostly the first two. So to the listener, there is little difference between a perlocutionist and a liar.

        But, Austin’s distinction is pretty handy. There are environments where we shouldn’t expect honesty — a listener shouldn’t expect to be told the truth by a perlocutionist. Courts assume this, and we should assume the same in political debate. So when we are in a perlocutionist environment, we should know that we have to let both sides speak. Each side should be invited to battle it out so the listener can decide whom to believe. When Warren tells brazen racist lies, she should pay a price and the false narrative the lie supports should be repudiated.

        I was trying to get at this earlier. Have we lost the common ground needed to do this, to shame brazen liars? Have we devolved into rank tribalism where signaling what team you are on is all that matters? Or, was it always this way?

        • Steve says:

          I agree. I hate lying. As a lawyer, I can’t mislead the Court. But, that means something very different than it should mean to a scientist. The Court knows that I am trying to persuade not find the truth. Deliberate misstatements should be punished. I find it appalling when they aren’t, but framing the facts in a way that makes one version of events seem more plausible than another is my job, but the same behavior should get a researcher in trouble. We have to acknowledge different standards in different context or I think that we risk false equivalencies that liars use to excuse their behavior.

  6. Ron Kenett says:

    cut and paste citations indicate how often cited papers are actually read, or not. the paper mentioned below estimates the extent of this phenomena. interesting read on corrections not done which provide an estimate of a prevalent effect:

  7. Anonymous says:

    Yeah and the officers didn’t beat the crap out of Rodney King for sport. They were acquitted!

    See the problem with stating that the court found, therefore is completely and obviously true. Especially regarding police officers. Especially regarding beating and killing unarmed black men. Shot 6 times. Unarmed. Feared for his life by a man running away. You can still swing that. And they did.

    Murder is murder even if “acquitted”. Not guilty is like 99% of the problem and why it locked off. Again. And will again.

    But i don’t expect you’ll correct yourself either.

Leave a Reply to Zhou Fang