Second-Guessing Confirmation Bias

This got me thinking:

…Facebook users tended to choose and share stories containing messages they accept, and to neglect those they reject. If a story fits with what people already believe, they are far more likely to be interested in it and thus to spread it.

Hey! This sounds exactly like the kind of message that fits with what I and my friends are likely to believe, like, and share!

I already believe confirmation bias is a thing, so maybe I should go look for contrary evidence that confirmation bias isn’t a thing. After all, that’s exactly what confirmation bias researchers say we should do so as to counteract confirmation bias.

So, I thought about it.

Consistent with the idea of confirmation bias, I do think I probably read a lot of things I am likely to agree with and then believe those things. But when I read things I don’t agree with, I often end up changing my mind to believe those things too. In general, I end up believing a really high fraction of everything I read regardless of whether I start off believing it or not. It’s really easy to convince me.

I used to read a bunch of Supreme Court opinions in their entirety. If I read the majority opinion alone, I would believe it and wonder how the dissenters could have tricked themselves into believing otherwise. If I read the minority opinion alone, I would wonder what the majority was smoking. If I read the majority and then the minority, I would first believe the majority but then realize the minority had gotten the right answer. This would happen even if the minority was a lone dissenter: Alito and Stevens alike.

It has occurred to me that the explanation for this favored by Occam’s Razor is that I am just incredibly stupid.

But maybe that’s understandable. If I were to play each member of the Western Conference All Star Team in succession, they could all probably convince me that they were the best player ever even though they can’t all be. It would be understandable that I might not be able to differentiate between the best and “worst” players among such an elite set. They would all be unimaginably good.

The Supreme Court is even more selective. Each is a dedicated professional arguer at the top of their field. And they spend a good chunk of the year sharpening themselves against one another. It makes sense that I as a reader would be hopelessly outclassed.

But I don’t just have this problem with Supreme Court justices.

I finally saw Star Wars VII over the holiday break, and it didn’t quite have the impact on me that it seemed to on others here who were celebrating Star Wars-apalooza. The most I could muster out would be agreement with James K:

They did not screw it up.

It had been decades since I saw the original series. I remember the characters, but the associated feelings had long since attenuated. I can’t complain about the plot being recycled because the plot was new to me. I didn’t even remember that IV was the Death Star movie. Star Wars VII seemed to be well-executed, but ultimately not moving. I got the feeling that the emotional bits everyone else got were due to callbacks to other things these characters had done in prior movies rather than set up in the movie I was seeing.

But I wanted to see if there was a better explanation for this dissatisfaction. And I found it here and here. These criticisms made sense to me, and I was happy to have what I thought were my feelings better explained to me.

And then I saw this post by someone who liked the movie. This writer clearly enjoyed it more than I did, and he can’t make me go back and like it more than I did, but I at least now feel silly for swallowing the first two critical reviews I read so uncritically.

You don’t need to be a Supreme Court Justice to play tennis with my opinions.1

What does this all mean for confirmation bias?

If my mind is a typical one (ha), then confirmation bias may be more of a symptom of the real problem of believing everything you read. Maybe we seek out confirming evidence because we just want to avoid constant whiplash. Confirmation bias might be an adaptation to a world in which it’s easy for anyone to sound convincing whether they actually are or not..

  1. I’m now cringing at the fact that I review academic papers and the course of science in my field has in some small way been influenced by whether I found what someone had written convincing or not. I’ve written before about what I think helped make me a critical reader, but I’m obviously not critical enough. []

Senior Editor
Home Page Twitter 

Vikram Bath is the pseudonym of a former business school professor living in the United States with his wife, daughter, and dog. (Dog pictured.) His current interests include amateur philosophy of science, business, and economics. Tweet at him at @vikrambath1. ...more →

Please do be so kind as to share this post.
Share

9 thoughts on “Second-Guessing Confirmation Bias

  1. I can’t tell what agreeing with this piece means and if i should feel confident that i’m not sure or more uncertain. And if i’m uncertain about my possible agreement, is that good or bad.

      Quote  Link

    Report

  2. I’ve read things and failed to be convinced by them. There are feelings associated with them, generally mediated by the style of the piece and the distance from the conclusion and my prior opinion. The feelings range from a slight sense of wrongness in the piece to smug superiority about my opinion to a visceral anger at the author for writing whatever it is they wrote.

    These feelings never seem to relate to a conscious process. I suspect they are associated with my tendency to nitpick – my tendency to say “I agree, except…”. I sometimes wind up accidentally arguing myself out of that initial “I agree”. But again, from the inside, it never feels conscious.

      Quote  Link

    Report

  3. Well, just to be clear, nothing that goes before the Supreme Court is easy, or obvious. So yeah, it makes sense that both sides of the argument are going to sound pretty good.

    When we look at stuff that is more empirical, I think the bedrock is data. What does the data say? what are the limits to what the data can tell us? I keep a mental category called “Things I don’t understand” around. Sometimes data that doesn’t quite conform to what I think is true goes into this category, hopefully to be hauled out and reexamined later when even more data becomes available.

      Quote  Link

    Report

  4. I don’t think what’s going on is ‘confirmation bias’ anymore.

    Confirmation bias is when you hear ten things, and *remember* the two that confirm what you already thought, and not really the other eight, even if two of the other eight were an argument against that.

    If ten stories about a part of town you think are dangerous happened to pass through your knowledge, you’ll remember the two of them are about crime there. Meanwhile, if someone else is actively trying to fight the perception that that part of town is dangerous, they’ll only notice the two heartwarming stories.

    But that’s not what happens anymore. What happens *now* is selective pre-filtering, where people only *learn* things that confirm what they already thought.

    This always used to happen a little, of course. But it would happen in the form of rumors and whatnot, and at least part of our minds would put qualifiers on it. Maybe it’s true, maybe that guy is spinning it.

    But now…it’s actual news stories, on reputable news sites. And it’s mostly real facts we’re learning…we’re just not learning *opposing* facts. Nor do we receive any correction when it turns out the facts were wrong. (And if we do…traditional confirmation bias kicks in.)

    And our minds don’t seem to qualifies those things with ‘Wait a minute…if all my friends have the same biases as me, isn’t it possible they’re only *sharing* stories that agree with my biases? When was the last time *I* read something that argued against what I believe, and then *I* shared it?’

    This trap seems a lot harder to escape than ‘confirmation bias’.

      Quote  Link

    Report

Leave a Reply

Your email address will not be published. Required fields are marked *