While this blog is called “Overcoming Bias”, I don’t recall explicitly addressing biases in a while. So let me revisit a bias which, though it is pretty clearly visible, most people don’t even bother to hide:
Well, it's part SDB, would say Bryan Caplan. One seldom sits alone. Plus confirmation bias: "As I sit there, I have reason to believe X - and good luck changing those views." - plus selection: If you really sit where you do not stand, it is much more likely you will change places. (Best critics are often those who've been there, done that, thought that - and then learned and left in disgust.)
One important clarification is that it's important that group membership is causally upstream of someone's belief. Otherwise, this fallacy doesn't apply.
To be a bit more concrete, I'll use myself as an example. I weakly self-identify as a rationalist. I've been reading Slatestarcodex, Lesswrong, and the associated ratsphere for a couple years now.
I also believe that artificial intelligence represents a real and credible threat to humanity. Your reaction to that confession might be: "Ah ha! Of course, the rationalist worries about artificial intellegence! Not so rational now are we!"
But I think this would be an incorrect inference. I don't believe in AI risk because I'm a rationalist. I'm a rationalist because I believe in AI risk. And this is plausible because, unlike the generation you're born into, rationalism is something that must be actively chosen.
In general, I find it important to distinguish between inherited and chosen identities.
Robin did say "Yes, most of us can quickly point out exceptions, where our views go against that predicted by where we sit. But this is less about particular views and more about an overall pattern". So n=1 isn't falsification here.
Anyway, I agree with you that it would be a good idea to have punitive taxes on trade unions. :-)
In statistics, bias is one of the ways to break down predictive error. Predictive error can be attributed to a combination of bias and variance. If there is no error, there is no bias.
If I am a member of group X, and I have good and sound reasons for saying that a policy that would benefit group X is better for society overall than the alternative, then this is not bias, because my reasoning is sound. For example, if a scientist claims that evidence-based science should have more influence over public policy decisions than it does, he is not guilty of bias. Because he's right and can base it in sound reasoning.
Taking it to a logical extreme, if I am in a legal dispute and some logical fact such as 2+2=4 happens to be essential to my case, then it is not bias for me to take the position that 2+2=4, even though it is to my benefit for others to accept this position.
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it."
There was a great excerpt from the book "Wild Problems: A Guide to the Decisions That Define Us" shared in this newsletter: https://threetimeswiser.sub...
I recall that your colleague Bryan Caplan has written extensively on how attitudes toward redistribution are NOT driven by whether one expects to personally lose or gain. And your link for "generation" discusses how an expected political development based on aging is NOT happening with Millennials.
I remember when I was at school I hit on the idea of asking the teachers, "If there was one subject through which all the other school subjects should taught, what subject should that be?"
I thought teachers might exhibit "own subject" bias, but I was quite surprised to find how strong that effect was.
Yes, I deliberately didn't specify. Another way to say the sentence would be "Is ridding oneself of this bias likely to help people get what they want?" What people want can include all the values/attributes you mentioned, and will differ between individuals.
"optimal" is meaningless unless it's associated with the value or attribute that one is seeking to optimize (preferably linked to some kind of metric function). Optimal in terms of happiness, wealth, relationships, health, karma....?
Is it optimal to rid oneself of this bias? As you say, it is rewarded. Maybe the more interesting question is whether society would be better off if everyone got rid of it? What function is the bias performing, is it an important function, and could it be done in other ways? A bit off the topic perhaps, but it's what I thought of when I read the article.
Hmmm. I'm a 72 white male, strongly in favour of affirmative action in racial and gender matters, and supportive of redistribution, punitive taxes on the oligarchy, trades unions, and proportional representation in all elections. Oh, and banning capital punishment and private gun ownership. Predictable?
Well, it's part SDB, would say Bryan Caplan. One seldom sits alone. Plus confirmation bias: "As I sit there, I have reason to believe X - and good luck changing those views." - plus selection: If you really sit where you do not stand, it is much more likely you will change places. (Best critics are often those who've been there, done that, thought that - and then learned and left in disgust.)
What happens after you overcome bias?
One important clarification is that it's important that group membership is causally upstream of someone's belief. Otherwise, this fallacy doesn't apply.
To be a bit more concrete, I'll use myself as an example. I weakly self-identify as a rationalist. I've been reading Slatestarcodex, Lesswrong, and the associated ratsphere for a couple years now.
I also believe that artificial intelligence represents a real and credible threat to humanity. Your reaction to that confession might be: "Ah ha! Of course, the rationalist worries about artificial intellegence! Not so rational now are we!"
But I think this would be an incorrect inference. I don't believe in AI risk because I'm a rationalist. I'm a rationalist because I believe in AI risk. And this is plausible because, unlike the generation you're born into, rationalism is something that must be actively chosen.
In general, I find it important to distinguish between inherited and chosen identities.
Robin did say "Yes, most of us can quickly point out exceptions, where our views go against that predicted by where we sit. But this is less about particular views and more about an overall pattern". So n=1 isn't falsification here.
Anyway, I agree with you that it would be a good idea to have punitive taxes on trade unions. :-)
In statistics, bias is one of the ways to break down predictive error. Predictive error can be attributed to a combination of bias and variance. If there is no error, there is no bias.
If I am a member of group X, and I have good and sound reasons for saying that a policy that would benefit group X is better for society overall than the alternative, then this is not bias, because my reasoning is sound. For example, if a scientist claims that evidence-based science should have more influence over public policy decisions than it does, he is not guilty of bias. Because he's right and can base it in sound reasoning.
Taking it to a logical extreme, if I am in a legal dispute and some logical fact such as 2+2=4 happens to be essential to my case, then it is not bias for me to take the position that 2+2=4, even though it is to my benefit for others to accept this position.
That's not really what "bias" is. There's no simple litmus test for whether bias is right or wrong....
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it."
There was a great excerpt from the book "Wild Problems: A Guide to the Decisions That Define Us" shared in this newsletter: https://threetimeswiser.sub...
I highly recommend this newsletter by the way, it's a gold mine: https://threetimeswiser.sub...
It's not always bias, to adopt positions that favor your own group! It's only bias if you're wrong.
I recall that your colleague Bryan Caplan has written extensively on how attitudes toward redistribution are NOT driven by whether one expects to personally lose or gain. And your link for "generation" discusses how an expected political development based on aging is NOT happening with Millennials.
I remember when I was at school I hit on the idea of asking the teachers, "If there was one subject through which all the other school subjects should taught, what subject should that be?"
I thought teachers might exhibit "own subject" bias, but I was quite surprised to find how strong that effect was.
Yes, I deliberately didn't specify. Another way to say the sentence would be "Is ridding oneself of this bias likely to help people get what they want?" What people want can include all the values/attributes you mentioned, and will differ between individuals.
"optimal" is meaningless unless it's associated with the value or attribute that one is seeking to optimize (preferably linked to some kind of metric function). Optimal in terms of happiness, wealth, relationships, health, karma....?
Is it optimal to rid oneself of this bias? As you say, it is rewarded. Maybe the more interesting question is whether society would be better off if everyone got rid of it? What function is the bias performing, is it an important function, and could it be done in other ways? A bit off the topic perhaps, but it's what I thought of when I read the article.
Hmmm. I'm a 72 white male, strongly in favour of affirmative action in racial and gender matters, and supportive of redistribution, punitive taxes on the oligarchy, trades unions, and proportional representation in all elections. Oh, and banning capital punishment and private gun ownership. Predictable?