Kaj Sotala points us to a finding that people ignore info more when they feel powerful. On a related note, Michael Nielsen, quantum physicist extraordinaire, suggests a "bias toward power": A form of bias I’m interested in is the great deference we pay to power, often far more than is warranted by the facts. I’m particularly interested in the damage this does to powerful people, since it greatly reduces the incentive they have to perform well …
Is a Bias Toward Power different from obedience to power (as famously shown by Milgram)?
That is, I'm interested in the potential causality chain here. Is the bias to trusting/respecting/accepting things coming from a powerful figure different from the push to obey? Or do we obey only inasmuch as we are being deferential?
I've always felt a strong pull of the authority of Socrates on this score--when people are wise in one thing, they think they are wise in many. Thanks for confirming it. ;)
The book written when a writer is unknown takes as long as the writer thinks it needs; if the writer thinks it needs uninterrupted concentration, he or she can save up money and quit a job and do nothing but write. This can hardly be said to apply to the work of published authors, whose new books are often contractually tied to a deadline, and who generally work on the new book amid constant interruptions from those seeing earlier work into print. Faulkner is said to have written As I Lay Dying in six weeks while working as a security guard; it's easy to imagine writing with that kind of intensity under those circumstances, less easy to imagine someone writing such a book while simultaneously defending the MS of an earlier book from an Armada of Bridget Jones clones.
"People tend to assume that the professor is right 100 percent of the time, and the student 0 percent. A more accurate breakup in my experience is 60/40. ... "
The 60/40 I would buy because what I, as a professor observe, is that 60% of college professors are competent. I, of course, am among the 60%who ARE correct virtually 100% of the time.
Nice post, but I don't think Tom Clancy is an example of the phenomenon. A lot of people, including authors, have just one idea--once they've written their book, that's it, really, except for fumbling around. And, as for Clancy in particular, all of his characters are the same guy with different names and jobs--no reason for him to keep on, and I doubt that even a lot of effort on his part would make a difference. Maybe Harper Lee was smart to quit after "To Kill a Mockingbird."
Isn't "bias toward power" just another name for credentialism?
Nope. A credential is just an acknowledgment of competence: credentials aqnd other kinds of certification should rationally be taken into account whenever their issuer is a reasonably trusted party. AFAICT, power does not enter into the issue at all.
Eliezer_Yudkowsky: You may be right for certain authors, but the claim was for works in general. What about all the one-hit-wonders in music or movies?
(Btw, "number", as the comparative of "numb", is hard to parse because it looks at first like #.)
I wonder if this effect is any different for authors than for executives? There's a very definite Senior Author Syndrome with examples without number - Robert Heinlein, Orson Scott Card, Anne McCaffrey, and the list goes on. Perhaps this is related to the editors deciding that they can just sell any book by that author, or the author getting powerful enough to refuse editing assistance. Or maybe, once an author makes enough money to lead a comfortable life, or just get old enough to be number to negative emotions, they can't hurt their characters enough for a good story. Definitely, part of Senior Author Syndrome is that your characters seem to stop getting hurt, and nothing really bad happens in your worlds anymore.
I've got to say, the notion of fooled-by-randomness/regression-to-the-mean doesn't seem very plausible here. We're talking about a long series of good books, starting good, getting better, spread out over years, but then over decades gradually degenerating into plotless painless random event sequences.
Off topic: You know, I was thinking about how I object to the notion of predictive markets. I think there a semantics thing here, where I see 'predictions' as being somehow being a statement about expecting events in the future without regard to the consensus view of the likely outcome.
I think if you refered to, if you will, 'forecastive markets' you would be on more sound semantic ground (if less sound grammatical ground) and would have to spend less time defending predictive markets against the occasional failure as proof of their invalidity.
If you refer to it as a forecast in the first place, people will have an intuitive sense that it is like the weather forecast, an estimated probability.
Now, it won't be as sexy or as feakanomics-like, but maybe that's the price we pay for clarity.
You must work in a completely egalitarian department. In a more hierarchical culture, Nielson is right. (I either work in a hierarchical department or am astonishingly stupid. (Yes, I really am open to both possibilities.))
I thought those statements were both from Nielsen. one was talking about the inferences people make in that situation, and the other what inference they probably *should* make.
i.e., given the only evidence we have about A is that a prof claims A, and a grad student claims ~A, we should probably assign a roughly .6 probability of A, but most onlookers will assign a .99 probability of A largely due to the prof's power.
Er, sorry PK, for saying almost exactly what you did because I didn't read you post.
Wait -- this is a great opportunity to test out the theory! We should expect that whichever one of us has more power will be deemed the genius, regardless of the similarity or who posted first!
People's early works, when they are unknown, are often better than their later works, after they've become famous. See, e.g., Tom Clancy.
Isn't that more easily explained through the "fooled by randomness" effect? That is, whether the public at large loves a new artistic work is essentially random, and so the success of one work has little predictive power about the success of the rest, much like a mutual fund manager.
People's early works, when they are unknown, are often better than their later works, after they've become famous. See, e.g., Tom Clancy.
This could also be explained by regression toward the mean.
I'm not saying that there are no biases related to power but that's not the only explanation for subsequent works not being as good if the first one was really really good.
Bias And Power
Is a Bias Toward Power different from obedience to power (as famously shown by Milgram)?
That is, I'm interested in the potential causality chain here. Is the bias to trusting/respecting/accepting things coming from a powerful figure different from the push to obey? Or do we obey only inasmuch as we are being deferential?
I've always felt a strong pull of the authority of Socrates on this score--when people are wise in one thing, they think they are wise in many. Thanks for confirming it. ;)
The book written when a writer is unknown takes as long as the writer thinks it needs; if the writer thinks it needs uninterrupted concentration, he or she can save up money and quit a job and do nothing but write. This can hardly be said to apply to the work of published authors, whose new books are often contractually tied to a deadline, and who generally work on the new book amid constant interruptions from those seeing earlier work into print. Faulkner is said to have written As I Lay Dying in six weeks while working as a security guard; it's easy to imagine writing with that kind of intensity under those circumstances, less easy to imagine someone writing such a book while simultaneously defending the MS of an earlier book from an Armada of Bridget Jones clones.
"People tend to assume that the professor is right 100 percent of the time, and the student 0 percent. A more accurate breakup in my experience is 60/40. ... "
The 60/40 I would buy because what I, as a professor observe, is that 60% of college professors are competent. I, of course, am among the 60%who ARE correct virtually 100% of the time.
Venkat,Thank you for the link to your "The fifteen laws of meeting power." Definitely worth thinking through.John
Nice post, but I don't think Tom Clancy is an example of the phenomenon. A lot of people, including authors, have just one idea--once they've written their book, that's it, really, except for fumbling around. And, as for Clancy in particular, all of his characters are the same guy with different names and jobs--no reason for him to keep on, and I doubt that even a lot of effort on his part would make a difference. Maybe Harper Lee was smart to quit after "To Kill a Mockingbird."
Isn't "bias toward power" just another name for credentialism?
Nope. A credential is just an acknowledgment of competence: credentials aqnd other kinds of certification should rationally be taken into account whenever their issuer is a reasonably trusted party. AFAICT, power does not enter into the issue at all.
Isn't "bias toward power" just another name for credentialism?
Eliezer_Yudkowsky: You may be right for certain authors, but the claim was for works in general. What about all the one-hit-wonders in music or movies?
(Btw, "number", as the comparative of "numb", is hard to parse because it looks at first like #.)
I wonder if this effect is any different for authors than for executives? There's a very definite Senior Author Syndrome with examples without number - Robert Heinlein, Orson Scott Card, Anne McCaffrey, and the list goes on. Perhaps this is related to the editors deciding that they can just sell any book by that author, or the author getting powerful enough to refuse editing assistance. Or maybe, once an author makes enough money to lead a comfortable life, or just get old enough to be number to negative emotions, they can't hurt their characters enough for a good story. Definitely, part of Senior Author Syndrome is that your characters seem to stop getting hurt, and nothing really bad happens in your worlds anymore.
I've got to say, the notion of fooled-by-randomness/regression-to-the-mean doesn't seem very plausible here. We're talking about a long series of good books, starting good, getting better, spread out over years, but then over decades gradually degenerating into plotless painless random event sequences.
Off topic: You know, I was thinking about how I object to the notion of predictive markets. I think there a semantics thing here, where I see 'predictions' as being somehow being a statement about expecting events in the future without regard to the consensus view of the likely outcome.
I think if you refered to, if you will, 'forecastive markets' you would be on more sound semantic ground (if less sound grammatical ground) and would have to spend less time defending predictive markets against the occasional failure as proof of their invalidity.
If you refer to it as a forecast in the first place, people will have an intuitive sense that it is like the weather forecast, an estimated probability.
Now, it won't be as sexy or as feakanomics-like, but maybe that's the price we pay for clarity.
You must work in a completely egalitarian department. In a more hierarchical culture, Nielson is right. (I either work in a hierarchical department or am astonishingly stupid. (Yes, I really am open to both possibilities.))
I thought those statements were both from Nielsen. one was talking about the inferences people make in that situation, and the other what inference they probably *should* make.
i.e., given the only evidence we have about A is that a prof claims A, and a grad student claims ~A, we should probably assign a roughly .6 probability of A, but most onlookers will assign a .99 probability of A largely due to the prof's power.
Er, sorry PK, for saying almost exactly what you did because I didn't read you post.
Wait -- this is a great opportunity to test out the theory! We should expect that whichever one of us has more power will be deemed the genius, regardless of the similarity or who posted first!
People's early works, when they are unknown, are often better than their later works, after they've become famous. See, e.g., Tom Clancy.
Isn't that more easily explained through the "fooled by randomness" effect? That is, whether the public at large loves a new artistic work is essentially random, and so the success of one work has little predictive power about the success of the rest, much like a mutual fund manager.
Finally! A scientific mandate for sticking it to the man!
People's early works, when they are unknown, are often better than their later works, after they've become famous. See, e.g., Tom Clancy.
This could also be explained by regression toward the mean.
I'm not saying that there are no biases related to power but that's not the only explanation for subsequent works not being as good if the first one was really really good.