Nature Is Not A Slate. It’s A Series Of Levers.

Last week I criticized pop social psychology while maintaining that social psychology itself was a pretty interesting window into human thought processes.

I was then handed a link to someone who apparently likes social psychology a lot less than I do.

You should read the whole thing, but here are the parts I’ll be talking about. As you can see, it’s kind of a conservative perspective saying social psychology is a liberal enterprise to deny human nature in favor of people being infinitely malleable based on their situation:

Personally, I find it very hard to fathom the idiocy of Mischel’s conclusion. It would mean that a person who others think of as for instance shy is really nothing of the kind. It just looks that way because we have only had chance to observe him or her in situations that elicits shyness. And if you think of yourself as shy you must be either plain wrong or stuck in a series of situations that by coincidence predisposes you to acting shy. This idea may sound like a joke, but the zeitgeist of the 1960s was left of sanity and lots of “intellectuals” believed Mischel the way they believed in Marx, Lenin and Mao.

For that reason, social psychology became a major branch of psychology. After all, if it was all in the situation then this was the important field of research. Personality barely survived and its proponents, like Hans Eysenck and Arthur Jensen, were often dismissed as racists and right-wing lobbyists.

Unsurprisingly, it soon became evident that Mischel was wrong – there really was such a thing as a shy person. As traits became real again, personality psychology grew but at the same time social psychologists kept a grip on their dominant position by introducing interactionism, the study of both situations and personality. This way they blurred the line between the fields and managed to claim a lot of the newly available positions in personality research. But at heart they were never interactionists; they started out as situationists because of their political views and they have stayed that way ever since. To this day they rarely perform experiments in which personality measures are used. Their focus is very much on the situation. Look at the collective social psychology blog in the links to the right of this post – it’s even called The Situationist, not The Interactionist. For most of these psychologists, interactionism was just a word with which to neutralize the enemy.

And today? Well, it’s like the French say, the more something changes, the more it stays the same. This can be seen in a recent post in the above mentioned The Situationist (which is still interesting to follow because not all social psychology is crap) about Harvard Professor Francesca Gino’s book Sidetracked, in which she describes how small things or situations derail our plans, intentions and even our morals. As an example she mentions an experiment in which she and psychologist Dan Ariely equipped participants with high-end sunglasses. Half of the participants were told that they were actually counterfeits, while the other half were told they were the real deal. The participants were then instructed to perform a mathematical task which left room for cheating. It turned out that 70 percent of those who thought they wore knock-off sunglasses cheated compared to 30 percent in the other group.

This may sound like compelling evidence for the power of the situation, but is it really? The participants were all young women rather than a representative sample. But more importantly, they were informed that they were participating in a psychological experiment and then told to wear counterfeit sunglasses. That’s pretty far from any kind of real life situation. It’s more like saying, “let’s play a game – you will be the bad guy.” It supports the idea that social psychology is, as someone put it, a list of how people behave in weird situations. Needless to say, Gino and Ariely didn’t use any personality measure since that would only distract attention from the power of the almighty Situation.

So if wearing fake sunglasses can make a person dishonest, how about the situation of being brought up by criminal parents? Now that should be a way more powerful situation. Psychologist Sarnoff Mednick and colleagues investigated this in the mid 1980s using data from over 14 thousand nonfamilial adoptions (in which the adoptive parents are unrelated to the child). They found that when both biological and adoptive parents had no criminal convictions the adopted child was eventually convicted in 13.5 percent of the cases, so that’s our baseline. When adoptive parents had convictions but biological parents had not, the number of convicted adoptees only rose very slightly to 14.7 percent. So fake sunglasses will have a profound effect on your honesty, but being brought up by criminals will only marginally elevate your risk of being convicted of a crime. That must be some sunglasses.


[Social psychologists] just need to construct some even more artificial situations in order to deliver those results that will prove that Marx was right all along. And no outsiders need to concern themselves with exactly how they go about doing that.

Where to start, where to start?

First of all, it is too bad that Staffan finds the importance of situation hard to believe, but at least he is in good company. They do not call this problem the Fundamental Attribution Error because it is rare (or for that matter because it is a correct and tenable position). In experiments, people consistently overestimate the effect of personality and underestimate the effect of situation. Insofar as social psychologists are the people trying to correct that, they are doing God’s work.

On the other hand, there are also personality psychologists, who study personality. They, too, are doing good work. Contrary to Staffan’s assertions, they are an integral and well-beloved part of the field of psychology. Of the two personality psychologists Staffan claims were villainized and dismissed, both made the list of the fifty most eminent psychologists of the 20th century. Eysenck is the third most-cited psychologist of all time.

Jensen was indeed “dismissed as a racist and right-wing lobbyist”, but this was less because he dared to study personality and more because he spent much of his time trying to prove black people were genetically less intelligent and because he took large amounts of money from right-wing organizations. This seems like the sort of case where “racist and right-wing lobbyist” might be a perfectly acceptable value-neutral description – although I agree Jensen’s work, much of which was completely unrelated and brilliant, hasn’t gotten the recognition it deserves.

Because of the success of personality psychology, some of Staffan’s claims about social psych are a No True Scotsman argument. Personality psychology is considered a different field than social psychology. When Staffan notes that social psychologists accept the importance of personality but don’t study it, this is exactly symmetrical to personality psychologists accepting the importance of situation but not studying it. So getting upset at social psychologists for not studying personality is a lot like getting upset at cardiologists for not studying the liver.

The sunglasses experiment? Staffan correctly points out (as I did last week) that several similar experiments are under challenge, but those were on unconscious priming whereas the sunglasses experiment (from the little I read on it) seems to be about conscious priming, which is on firmer ground. To me it seems like exactly the sort of thing that might be true – in the extremely artificial conditions of the laboratory. This meshes with the point I made last week – that social psychology does a great job illuminating certain processes the brain goes through, but that we should be wary about assuming they have what doctors call “clinical significance”.

The criminal adoption experiment? Probably 100% correct. As anyone who’s read The Nurture Assumption (or my review thereof) knows, psychologists are constantly unable to find any effect of parents on their childrens’ personalities or actions. This is sufficiently distressing that most people refuse to believe it, but it keeps being confirmed again and again. Personality is 50% genetic and 50% some other factor which people have yet to illuminate but which definitely doesn’t involve upbringing (I hear Judith Rich Harris’ book No Two Alike purports to explain what this factor is, but I’m only halfway through it and can’t comment).

(seriously, why do I have to spend so much time insisting to racists and eugenicists that genes are seriously really important? This should not be as big a niche in the blogosphere as it is!)

But basically, social psychology has discovered the correct fact that situation is more important than people think it is and personality is less important than people think it is in determining behavior, while not denying that both are pretty important. It correctly claims that priming can have very large short-term effects on unimportant decisions, and correctly notes that being raised by criminals has no effect on anyone’s personality. So far I think it’s doing pretty well, as long as, once again, you are skeptical about trying to do social engineering with its results outside the laboratory.

Now let’s get to the part about Marx. For this we go to those experts on all things Marxist, More Right. In their latest post, Drew Summitt draws a distinction between what he considers a conservative view – that there is such a thing as human nature and that political systems need to take that into account – and a progressive view – that there is no such thing as human nature, people are infinitely malleable, and once we create some kind of utopia we can perfect mankind. The quote is his, the emphasis is mine:

“He defends this proposition with the assertion that “the conservative realist view of human imperfectability and their commitment to ordered liberty as rooted in nature, custom, and prudence“ can see great support in modern evolutionary theory because modern evolutionary theory contradicts what Thomas Sowell calls the “unconstrained vision” that liberal intellectuals and theorists are tempted to hold. In contrast to this Sowell sets up a “constrained vision” of human anthropology that is limited in its capabilities by an intellect being the servant of the passions, the reality of sin, or boring genes telling our memes what they can and can’t do. In order to support the idea that evolutionary theory supports a conservative political vision Arnhart traces the foundations of human capability to nature, custom and prudence. The conservative hierarchy of nature, custom and prudence is what constrains the idealist impulse for reason to govern and judge, and indeed seek to overthrow, custom and nature


But the Liberal Egalitarian Free-trade Technocrats also recognize human nature as being essential to political order. If you agree with Steven Pinker on the Humean Is in regard to human nature, that is, if you think he’s got his facts right, you must have an independent account of the Humean Ought, because Pinker is a Liberal, though one of the deflated, Clintonian Liberals. This is a form of Liberalism not touched enough reactionary circles. Why is it the Daniel Dennett, Steven Pinker, Peter Singer and many other experts in contemporary studies in human nature self described liberals? They may reject modern queer theory, as Dawkins does, or they may think that the brain is not a Blank Slate, as Pinker does, but they don’t consider these positions to be dangerous to Liberalism writ large or if they do they are terribly good at hiding it.

So he is wondering why, if liberalism is founded on the idea that human nature is an infinitely malleable blank slate, are the world’s greatest scientific experts in evolved human nature liberals?

To steal a delightful turn of phrase from Terry Eagleton, this is like wondering why, if Tony Blair is an octopus, he has only two arms.

Let me make an analogy to medicine. Unlike the brain, there is no debate on the “nature” of the heart – the literal blood-pumping heart, not the fuzzy emotional version. We know the heart is fully one hundred percent genetically programmed (minus a little morphogenetic variation), that it’s not malleable by schooling or brainwashing or being raised in a commune. It is a social engineer’s nightmare, a system founded entirely on human nature without the slightest wiggle room.

And yet doctors routinely make the heart do what they want. If they want to raise heart rate, they give a dose of epinephrine. If they want to lower heart rate, they give a dose of propranolol. If social planners could control the brain as easily as doctors control the heart, we’d already be living in a communist utopia.

The heart has an immutable nature, and that immutable nature is to respond to different situations in highly predictable ways.

The heart is neither a blank slate nor a fully inscribed slate. It’s not a slate at all. The heart is a series of levers. If you pull one lever, it will do one thing. If you pull another lever, it will do another thing. It is, paradoxically, hard-coded for malleability. It’s not infinitely malleable – there’s no drug you can inject to make the patient’s heart beat out the drum parts to Beatles songs – but you can shift it a little bit this way or that.

We have reason to believe the brain works the same way. Not everything is a lever – if you send a kid off to be raised by criminals, it won’t activate any of the hard-coded IF-THEN statements, and nothing will happen. But if you surround someone by stimuli that prime the idea of criminality – whether sunglasses or a broken window, that will pull on a lever that will make criminal behavior a little bit more likely.

(except for “levers”, read “extremely complicated things that run through chaos theory at some point and so are inherently unpredictable except in the broadest and most statistical sense”)

All of this reminds me of a video I saw this afternoon on the second day of The Hospital Orientation. Please excuse me if I change it around just a little to turn it from a quality improvement case study to a morality tale.

There were two hospitals, Hospital A and Hospital B. Both, like all hospitals, were fighting a constant battle against medical errors – surgeons removing the wrong leg, doctors giving the wrong dose of medication, sleepy interns reading x-rays backwards, that kind of thing. These are deadly – they kill up to a hundred thousand people a year – and terrifyingly common.

Hospital A took a very right-wing approach to the issue. They got all their doctors together and told them that any doctor who made a minor medical error would get written up and any doctor who made a major medical error would be fired. Rah personal responsibility!

Unfortunately, when they evaluated the results of their policy they found they had exactly as many medical errors as before, except now people were trying to cover them up and they weren’t being discovered until way too late.

Hospital B took a very progressive approach. They too got all their doctors together, but this time the hospital administrators announced: “You are not to blame for any medical errors. If medical errors occur, it means we, the administrators, have failed you by not creating a sufficiently good system. Please tell us if you commit any medical errors, and you won’t be punished, but we will scrutinize what we’re doing to see if we can make improvements.”

Then they made sweeping changes to what you might call the “society” of the hospital. They decreased doctor workload so physicians weren’t as harried. They shortened shifts to make sure everyone got at least eight hours of sleep a night. They switched from paper charts (where doctors write orders in notoriously hard-to-read handwriting) to electronic charts (where everything is typed up). They required everyone to draw up and use checklists. They even put propaganda posters over every sink reading “DID YOU WASH YOUR HANDS LONG ENOUGH??!” with a picture of a big eye on them. You can’t get more Orwellian than that.

And yet, mirabile dictu, this was the hospital that saw their medical error rates plummet.

The administrators of this second hospital didn’t ignore human nature. Instead, they exploited their knowledge of human nature to the fullest. They know it’s in human nature to do a bad job when you’re working on no sleep. They know it’s human nature to try to cut corners, but that people will run through checklists honestly and effectively. They even know that studies show that pictures of eyes make people behave more prosocially because they feel like they’re being watched.

You don’t have to tell me all the reasons this doesn’t directly apply to an entire country. I can think of most of them. But my point is that if I’m progressive – a label I am not entirely comfortable with but which people keep trying to pin on me – this is my progressivism. The idea of using knowledge of human nature to create a structure with a few clever little lever taps that encourage people to perform in effective and prosocial ways. It’s a lot less ambitious than “LET’S TOTALLY REMAKE EVERY ASPECT OF SOCIETY AS A UTOPIA”, but it’s a lot more practical.

(Although I’m also kinda okay with making every aspect of society a utopia, as long as we do it right.)

This entry was posted in Uncategorized. Bookmark the permalink.

55 Responses to Nature Is Not A Slate. It’s A Series Of Levers.

  1. Jared Harris says:

    So what’s the (probably incompatible) anti-liberal argument against people having these levers and society using them? There has to be one…

    • Oligopsony says:

      Nature is not meant to be a passive object to be endlessly manipulated by technology, Aristotelian teloi, trying to change complex systems always results in disaster, hubris, &c. &c.

  2. Sniffnoy says:

    See, to me, both hospital’s responses seem to assume that this series of levers exists, the difference is that one bothers to actually look for them and the other just assumes they’re the obvious thing.

    • Sniffnoy says:

      Oh, wait, I think I see. I was interpreting the purpose of writing-up/firing as a way of preventing errors in the first place by deterrence, but it could also be interpreted as a way of rapidly cycling through bad doctors to find the good ones. The latter fits the immutability idea.

    • Nick T says:

      This is a good point, “mutability of human nature” seems like a pretty clumsy way to compress the difference here.

      Somewhat nitpickingly, the first hospital’s approach is “the obvious thing” to me as well, but I doubt it would be to everyone; it takes a certain starting point for “there should be rules and people should submit to them” to be more obvious than “people’s environment should be changed to make the right thing easier”. (I don’t know how to characterize that starting point, or if there is a single one.)

      I notice that I have the strong feeling that the first hospital’s approach should work, that people should be able to just follow rules (especially ones with this good a justification), and that the second hospital’s approach is… giving in, maybe? Not holding people to the correct standard? I largely don’t endorse this on reflection, but it appears that part of me strongly terminally values authoritarianism in a way that’s hard for me to work around.

      • Scott Alexander says:

        Possibly an issue of whether the doctors are slacking off versus being basically good people who have reached the limits of their capability? But I don’t know how to operationalize that.

    • houseboatonstyx says:

      @Nick I notice that I have the strong feeling that the first hospital’s approach should work, that people should be able to just follow rules [….]

      Well, that would depend on what sort of rule it is. A rule like “Don’t sneak in under the side of the tent” is pretty much up to voluntary control. But “Don’t be so tired that you read an xray backwards” is not.

  3. Randy M says:

    Tangent–Is there a label you are comfortable with? I called you a progressive rationalist after linking a piece on facebook and getting a bit of a rebuke for it.

    • Scott Alexander says:

      Not really. Tell whoever rebuked you I’m kind of okay-ish with that.

      I did describe myself as “up-wing” after a discussion in which I decided I was located at a point floating several meters above the Two-Axis Political Spectrum.

  4. Douglas Knight says:

    I see a disconnect between the first part of this post, about social psychology, and the second about practical use of levers. None of the levers in the successful hospital (except maybe the eye) are social levers. Nor did the bad hospital follow personality psychology. I guess that the connection between the two parts is that they both claim to compare politics to psychology, but the politics of the morality play seem pretty arbitrary, in contrast to the genealogy of the politics of branches of psychology.

    I think you are unfair to Staffan. He doesn’t find the importance of situation hard to believe; he finds hard to believe that personality has no importance, because it does. Well, actually, what he finds hard to believe is that some people held that position. When you say that social psychology does not deny that personality is “pretty important” you are simply contradicting him. Maybe he’s wrong and you should contradict him, but you don’t acknowledge that contradiction, and instead put stupid statements in his mouth. I don’t know about the personality-situation debate, historically or today, but I certainly hear psychologists today making sweeping statements in favor of nurture over nature.

    • gwern says:

      > None of the levers in the successful hospital (except maybe the eye) are social levers.

      First off, ‘maybe’? The eye one is like the paradigmatic example.

      Second: you think that results about checklists and sleep and blame do not fall under social psychology? Could you explain what fields you do think they fall under and provide citations to papers demonstrating those results by people identified as working in those fields?

    • Scott Alexander says:

      Hm. What would you say is the proper procedure for even debating “Does social psychology admit the importance of personality” when Staffan admits that most social psychologists claim that they agree personality is important?

      • Douglas Knight says:

        Yes, it’s hard to debate whether people lie about their beliefs.

        You are right to point out that some of what he says is symmetric between fields, but I didn’t criticize that paragraph. But his criticism is predicated on a lack of symmetry. You deny the lack of symmetry (which I find utterly unconvincing). That’s the end of the argument. Maybe you should spell out how that disagreement propagates through the rest, but you should make it clear that it is the point of departure and not criticize his process of elaboration of the previous belief.

        But why do you care about this meta stuff? Why does Steffan? It’s a disjointed essay and he doesn’t really say why. The best I can string together is that the field started out corrupt in one way and while that way has changed, it inherited other forms of corruption. In any event, why don’t we focus on the object level accusations of corruption?

        The only interesting thing Steffan says is to criticize the sunglasses experiment not for being too artificial or narrow (which are serious problems, but endemic in psychology), but for being role-play. That is, maybe the subject knows what’s being tested and wants to confirm the experimenter’s theory. If so, the existence of a political echo chamber is relevant, though Steffan fails to make that connection. This is a standard criticism of the Zimbardo prison experiment and of stereotype threat. It’s also why deception is pretty common in psychology experiments, as in the faithfulness experiment you mentioned.

  5. ShardPhoenix says:

    I don’t agree that those two hospital scenarios map to “right wing” vs. “progressive”.

    • Deiseach says:

      Hospital A’s approach reminds me of what “Private Eye” magazine’s M.D. describes as the “culture of blame” in the National Health Service; hospitals are so bound by having to meet government-set targets about waiting lists, mortality rates, etc. that they stifle any honest assessment of how they’re doing because if they fall behind or fail to meet what’s been set, they get budget cuts and held up as failures.

      So if something is really going wrong, whistleblowers are discouraged and if they don’t get fired themselves, they’re bound by confidentiality agreements; there is no incitement to try and improve conditions through working on the problems because it’s a ‘blame and fire’ culture, and nobody wants to be labelled as a troublemaker.

    • Charlie says:

      The same cultures demonstrably play out in conservative vs. liberal school improvement policies, so I think it’s at least plausible.

    • Scott Alexander says:

      Imagine that instead of medical error, we were talking about crime. Police Force A says “Let’s increase sentences!” Police Force B says “Crime is the fault of society, we need the administration/government to change society to remove the factors that make crime tempting.”

      Does that make the right-wing/progressive mapping seem more plausible?

  6. Deiseach says:

    “(T)rying to do social engineering with its results outside the laboratory.”

    That is my main problem with social psychology, because it’s all too common a temptation (and one not often resisted) for government policy to try and socially engineer for ‘the good of society’ (and I mean all governments by this; I’m not saying left vs.right or conservative vs. liberal). It’s very tempting to say “Let’s reduce teenage pregnancy rates or increase literacy levels or find out what businesses want in their employees and educate school leavers towards that standard” and impose large-scale plans that will then get scrapped by the next lot who get into office; meanwhile the schools or the unemployed or the ‘underclass’ get treated like lab rats and nobody really knows what works or how it does work if it does work at all.

  7. Sarah says:

    What you’re calling “conservative” and “liberal” I see as “people” and “process” approaches.

    Do we try to accomplish our goals by hiring better people? (or motivating/teaching the existing people to be better?) Is it basically about virtue and merit?

    Or do we try to accomplish our goals by setting up better management processes and incentive systems? Is it basically about context?

    To give an example of how this cuts across political lines, “Free markets create wealth” is a process-focused libertarian idea, while “innovative individuals create wealth” is a people-focused libertarian idea.

    Teach For America is a people-focused program (improve schools by hiring ambitious Ivy Leaguers) within the government institution of public schooling. Lean Startup methodology is a process-focused program (improve business by a/b testing everything) that’s become popular among private companies.

    Atul Gawande’s checklist manifesto? The ultimate in process focus. Peter Thiel’s theories about startup success? 100% people focus.

    I don’t subscribe wholeheartedly to one at the expense of another. But I think we are living in a period when process focus is the most popular frame, and people focus goes underrated.

    Past institutions (such as the administration of the British Empire) got things done with a vastly smaller bureaucracy than we do now. My personal opinion is that modern institutions are trying to use management process as a substitute for individual excellence, and it mostly doesn’t work.

    Bad people plus good process gets you CYA. People optimize for the targets of the process, or the incentives, which aren’t exactly the goals of the institution.

    Good people plus bad process gets you hacky solutions. What the Indians call “jugaad” — something clever, makeshift, jury-rigged, maybe a little crooked. Smart, good people find ways to get their own small piece to work in corrupt and dysfunctional systems. Of course in a sufficiently bad system, jugaad turns into a form of defection — crime, or some kind of opting out.

    Naturally, you’d want good people AND good process.

    • Scott Alexander says:

      I am annoyed that you always comment with what I was trying to say, only better. One of these days you need a non flocked blog.

  8. James says:

    The idea of using knowledge of human nature to create a structure with a few clever little lever taps that encourage people to perform in effective and prosocial ways.

    This sounds like Thaler & Sunstein’s
    Nudge theory. They call themselves libertarian paternalists; the clever little lever taps are “nudges” and the structure that taps the levers is called “choice architecture”.

    • Berry says:

      Yup. Trying to make the optimal solution into the default solution for most people but without overt coercion.

  9. Randy M says:

    So, the hospital example/analogy: Given that it doesn’t have to be either or, and no hospital should do things like allow doctors to operate (or be on call) efffectively intoxicated due to lack of sleep (isn’t that obvious?), let’s take a look at a possible outcome ten years later. Hospital A has had some more patients have died due to the lack of good processes that Hospital B implemented. But let’s assume that they were diligent in uncovering the sources of error and fired the doctors, well enough that some mistakes were made but all the doctors were eventually caught (maybe not after the first error, but subsequent) or else adopted the best practices of those who were more careful if they were capable of doing so. So both Hospitals have now reduced their error rate, A with attrition and scaring off some portion of those less capable, B with the kind of ‘paternalism’ you describe. Which has reduced it further? And are the doctors/nurses at A now intrinsically better surgeons, diagnosticians, etc.? Is A the more far-view system? (I don’t know the answers, and I’m not assuming that either is correct, nor even that the analogy holds elsewhere)
    As I said, in a hospital they should certainly look at both (instituting best practices, firing for gross negligence if those are ignored repeatibly, etc.). But it occurs to me that perhaps this is why early moral systems that survived tended to be draconian–that was the only form of eugenics available.

    • Randy M says:

      (sorry for the block of text attack)

    • Intrism says:

      My assumption would be the opposite. Doctors at Hospital A face draconian management, a high error rate, high churn, and frequent and nasty office politics. In short: it’s gonna be a terrible place to work. The ones who can get out of Hospital A almost certainly will. These, of course, will be the good doctors. So, Hospital A is going to end up with staff that are good enough (at medicine, lying, or luck) not to get fired but bad enough that they can’t work somewhere nicer.

      • Randy M says:

        I don’t know. That seems like a good point, but what are the draconian rules? Don’t kill a patient? It sounds like they actually face wide latitude, but harsh punishment for error. Many might prefer that to the micromanagement present at B.
        Perhaps there are good and bad ways to go about implementing each error reduction method.

        • Earnest Peer says:

          I think the really draconian rule isn’t the one where you get fired for big fuck-ups but the one where you get written up for small ones. At least, that’s the bad one, because that one makes covering up much more of a habit than the other one.

    • Scott Alexander says:

      Your comment makes me think of two things.

      First, this is kind of like the nature-vs-nuture debate in that the construct we’re looking for depends on the inherent variability on both sides. Like, if there’s a huge amount of variability between competent and incompetent doctors, then firing the incompetent ones could work. On the other hand, if all doctors are basically equally good and they just make errors for reasons like tiredness, firing the ones who commit errors will just be random.

      Second, the strategy could be difficult for one hospital versus all hospitals. If there is a limited pool of doctors, then Hospital A’s firing strategy just leads to the bad doctors getting employed at different hospitals – the distribution of medical errors changes, but the number stays constant. On the other hand, Hospital B’s policy actively reduces the overall number of errors without shunting them over to someone else.

  10. roystgnr says:

    Sorry for the derail, but: “the heart is fully one hundred percent genetically programmed (minus a little morphogenetic variation), that it’s not malleable by schooling or brainwashing or being raised in a commune”, really? There aren’t any effects of exercise on heart size or capability?

    • Randy M says:

      The heart is an organ that changes itself in mostly predictable ways in response to stimuli.
      The brain is an organ that changes itself (and the rest of the body) in mostly* unpredictable ways in response to stimuli.

      *I may not be up to date on the correct strength of this adjective.

  11. Nick T says:

    Aaron Swartz has another good story on the anti-authoritarian/people-focused theme (that also namedrops the fundamental attribution error).

    Are there similarly good anecdotes in defense of the authoritarian/process-focused approach?

    • Douglas Knight says:

      There are two parts to the GM/Toyota story. One is social, about having a team ethos. The other is about having good processes. They is some connection, because the ability of the workers to stop the line or contribute ideas builds morale. Gawande tells a story of trying to get hospital workers to wash their hands more and asking for suggestions about how to do it. He said that he didn’t get any new ideas, but he did get buy-in.

      • Deiseach says:

        Here’s a stupid question, but did anyone ever try the effect of putting hand lotion dispensers up as well as hand sanitisers/wash hand basins? Because if you’re washing your hands multiple times a day, and you’re doing it with the kinds of harsh cleansers necessary for killing bacteria, and then drying them under hot air rather than on towelling, it is going to have bad effects on your skin.

        I wonder if part of the reason people don’t wash their hands as frequently as they should is down to having their skin dry out and peel and all the rest of it, so if there were emollients available as well to counteract that effect, would people wash their hands more?

  12. Leonard says:

    why do I have to spend so much time insisting to racists and eugenicists that genes are seriously really important?

    Because (a), the politically dominant religion of our time, progressivism, is radically egalitarian, and recognizes the ideological threat to egalitarianism from scientific study of race and genes. This is not your fault, of course, but it is a background against which you are being judged. And (b) you are squirrelly about admitting that genes are seriously really important. Yes, you’ve propounded some of the more politically acceptable implications of human evolution above the neck. Even reading Judith Harris verges on thoughtcrime. Good on you. But I’ve yet to see you admit any of the more politically incorrect implications of evolution, for example, that it is almost certain that various human races vary in average intelligence. (And yet you know many of the basic hatefacts, i.e. as you demonstrate in this recent post and its comments.)

    This should not be as big a niche in the blogosphere as it is!

    This is our belief. And yet: it is. Name me a higher-profile liberal or progressive blogger who is a scientific racist.

    It is not our side that makes it so: it is the progressive side, with its witchhunts and anathemas, which you are so carefully keeping one foot in. Not that I blame you — you are not using a pseudonym, and your team makes it dangerous to hold politically incorrect beliefs. Can you be allowed to be a doctor if you admit that Jenson was right? Doubt it. Racists should not be doctors! Therefore, you don’t talk about it and probably should not. I dunno. Truth is hard to value.

    • Sly says:

      “Because (a), the politically dominant religion of our time, progressivism”

      I love the smell of equivocation in the morning.

      • Berry says:

        I love the smell of equivocation in the morning.

        It smells like: “argument from my opponent believes something.”


    • Alex says:

      I’m not particularly interested in how brave you are, but rather whether your beliefs are accurate or not. Would you mind defending the main assertion of your post? (that blacks are significantly less intelligent). Even assuming that is true, why should we care if the races differ by a few IQ points? Are the costs of discrimination worth the benefit?

      • Mary says:

        Are the costs of the current regulatory regime work the benefit? One notes that even they were exactly equal, statistics say you will get disparate impact in locales, because variation is like that. If there is a significant if small difference, you will see even more, without any racial discrimination at all.

        Should government bureaucrats be able to put you through the wringer because of results that you ought to expect even in the total absence of racial discrimination?

        • Alex says:

          I don’t know what you mean by the current regulatory regime, or what you mean by the ringer.

        • Randy M says:

          EEOC, and lawsuits & hearings, etc., respectively.

        • Alex says:

          In that case I am not qualified to answer. I don’t know if the current regulatory regime is optimal or even a net positive, however, I happen to think that discrimination without regulation is a net negative but I’d rather discuss this another time.

        • Mary says:

          In the absence of laws positively requiring them to discriminate, businesses do not discriminate on basis of race. Do you know how the seating requirements that the Atlanta boycott was directed at were put into place? The city took to randomly stopping the buses and arresting the driver if the law was not followed. Prior to that, the company had ignored it.

          To be sure, if blacks are worse employees for whatever reason, they will find it harder to find jobs. The one technique that South Africa found effective in keeping blacks from “whites only” jobs was to impose minimum wage standards.

      • Leonard says:


        I said nothing at all about the intelligence of blacks in my post. I did assert that human races vary in intelligence. This is true, indeed must necessarily be true, since natural selection is happening (it must happen so long as we continue to have births and/or deaths) and different races have different intellectual environments. Of course, the degree to which races face different intellectual environments is unknown; thus my simple statement does not give any reason to believe in significant differences in intelligence.

        Now, as it happens, your assertion that blacks are significantly less IQ-full than whites is true. Blacks average one standard deviation below the white average (85 IQ, versus 100). This is a scientific fact; shall I google it for you? But I didn’t say it. You did. Which suggests that you, like Scott, know some hatefacts. Oh my!

        Of course, you may feel free to weasel upon the normal things, that “race” is ill-defined, that “IQ” just a stat-mining construct, that in any case “intelligence” is not “IQ”, etc. Sun got in your eyes; I get that.

        But you ask, why should we care? Should we not ban knowledge that violates our religious belief in equality? What good can come from the truth?

        If you really want to know, let me quote the great John Derbyshire here:

        Why do I say that the DZGD [the idea that all groups are equal] is a dangerous and evil doctrine? It is false, to be sure; but a false doctrine need not be dangerous. If the generality of Americans came to believe that Jupiter is further away from the Sun than Neptune, they would have come to believe a falsehood; but in all probability, society would go on much as before, and only pedants would feel any distress.

        Suppose you are a black American… Looking around, you notice all the familiar statistics of black America: the high rates of incarceration, single parenthood, and other dysfunctions. You also note that black Americans do not do very well in school (statistically speaking), do not have a fair proportion of good jobs, and so on. What is your logical deduction from all this?

        If you cleave to the DZGD, as everyone from the President on down insists that you must in order to be accepted into polite society, there is only one possible conclusion you can come to: Some force is keeping black people down. Since, on the DZGD, the statistical profile of your group on all measurable abilities is just like the statistical profile of any other group, there must be some force keeping black people away from society’s goods. What other force can that be, but the malice of nonblack people? Oppression! Racism!

        The DZGD thus generates discord and hatred. It is touted as a sine qua non of the modern civilized outlook. In fact it is a poisonous, anti-social doctrine, as well as a false one.

        I prefer the truth, because I would prefer to live in a society with racial harmony.

        • Alex says:

          “I said nothing at all about the intelligence of blacks in my post.”

          “Jensen was indeed “dismissed as a racist and right-wing lobbyist”, but this was less because he dared to study personality and more because he spent much of his time trying to prove black people were genetically less intelligent”

          “Can you be allowed to be a doctor if you admit that Jenson was right?”

          If I misinterpreted you as asserting that Jensen was right then I apologize but its very easy to read your statement that way.

          Before we misunderstand each other further let me pose a thought experiment.

          Suppose that I were to get a group of upper class white mothers to become surrogates for embryos that were obtained from poor black parents and vice versa (poor black mothers become surrogates for embryos from upper class white parents). Suppose also that the surrogate mothers and their spouse raise the babies that they give birth to ie. babies of the other race.

          I would predict that the black children raised by upper class white mothers would have an average IQ that is much closer to their adoptive/ surrogate parents than to their biological poor black parents and vice versa. What would you predict?

          “Blacks average one standard deviation below the white average (85 IQ, versus 100). This is a scientific fact”

          Alright so due to historical contingency the races are unequal. But that is not what I am disputing, I am disputing that they are significantly _genetically_ unequal.

        • Douglas Knight says:

          Alex, why do you care about whether the difference is genetic? Why do you think it has consequences for discrimination? What is the difference between something inherited through genes and something inherited through some unidentified mechanism? If the mechanism were breast feeding (sometimes proposed as 10% of the gap, but I don’t buy it), then we should promote breast feeding, but this only affects people born today.

        • Alex says:

          Whether or not I care is irrelevant – it was asserted and I am challenging said assertion. I might be willing to discuss why its important some other time though if you really want to.

        • Alex says:

          The main reason to care though, is that genes cannot be changed but environmental factors can be.

        • Douglas Knight says:

          Your original comment, which you claim is about genetics, not just the gap itself, claimed that it would have cost in terms of discrimination. What difference does genetics make to this?

          Maybe environmental factors can be changed, but I already gave two reasons why not: (1) we don’t know what they are; (2) even if we did know, we probably can’t change adults. Either way, it’s irrelevant to who should be hired as firefighters today.

        • Mary says:

          Actually, IIRC, someone did an experiment once tracking the children adopted by upper-middle-class whites in some state. Some were white, some were mixed race, some were black.

          The three groups were needed because such adoptions are an IQ raising factor in themselves.

          They found that the whites did better on IQ test than the mixed race, who did better than the blacks. Some children were mislabeled as black when they were mixed race. They performed like the other mixed race children.

        • Alex says:

          Thanks would you happen to have a link? It’s clearly different than my thought experiment and does not affect my prediction.

    • Nick T says:

      Reading comprehension fail? Scott is saying that the “racists and eugenicists” are the ones not taking sufficient account of genetics, not that he (Scott) has to defend himself by showing that he takes account of genetics.

    • Scott Alexander says:

      I wasn’t asking why people need to point out genes matter, I was asking why I kept having to point it out to the racists and eugenicists, ie to the people who claim to be leading the push to talk about how genes matter.

  13. Douglas Knight says:

    What is the psychological theory of the bad hospital? Other commenters have hinted at two theories.

    One is homo economicus, who responds to incentives. Economists are right to point out that this often applies when people don’t expect it to, but it often fails to apply. In particular psychologists know well that large rare punishments don’t work well. Also, even by this theory, the hospital has effectively given people a pay cut, by increasing the risk without increasing the reward. (This is, I think, a very different complaint than Intrism’s, which would apply even with a compensating reward.)

    The other theory, which you might lump under personality psychology, is that there are some bad apples and you have to weed them out. Of course, the personality psychologists suggest identifying them before hiring, perhaps by a conscientiousness screening. Of course, the whole of the doctors’ lives has been conscientiousness screen, so this probably wouldn’t do any good.

  14. Mary says:

    Eh, I remember the essay test that purported to show the fundamental attribution error.

    I notice that there is no mention of whether the essay writer did, in fact, support or oppose Castro. It is frequently not screamingly difficult to tell whether a writer actually believes what he’s arguing. It can be, I grant you, but it frequently isn’t. So it would be the way to bet, since the odds of opposing your own view may be fifty-fifty, but the odds of your both getting it and being able to aptly describe it are lower.

  15. Multiheaded says:

    A pretty excellent post!