Bryan Caplan  

Social Desirability Bias vs. Intelligence Research

PRINT
Individualism: True and False... Thomas Szasz: A Life Well-Live...
When lies sound better than truth, people tend to lie.  That's Social Desirability Bias for you.  Take the truth, "Half the population is below the 50th percentile of intelligence."  It's unequivocally true - and sounds awful.  Nice people don't call others stupid - even privately.

The 2000 American National Election Study elegantly confirms this claim.  One of the interviewers' tasks was to rate respondents' "apparent intelligence."  Possible answers (reverse coded by me for clarity):

0= Very Low
1= Fairly Low
2= Average
3= Fairly High
4= Very High

Objectively measured intelligence famously fits a bell curve.  Subjectively assessed intelligence does not.  At all.  Check out the ANES distribution.

iqanes.jpg

The ANES is supposed to be a representative national sample.  Yet according to interviewers, only 6.1% of respondents are "below average"!  The median respondent is "fairly high."  Over 20% are "very high."  Social Desirability Bias - interviewers' reluctance to impugn anyone's intelligence - practically has to be the explanation.

You could just call this as an amusing curiosity and move on.  But wait.  Stare at the ANES results for a minute.  Savor the data.  Question: Are you starting to see the true face of widespread hostility to intelligence research?  I sure think I do.

Suppose intelligence research were impeccable.  How would psychologically normal humans react?  Probably just as they do in the ANES: With denial.  How can stupidity be a major cause of personal failure and social ills?  Only if the world is full of stupid people.  What kind of a person believes the world is full of stupid people?  "A realist"?  No!  A jerk.  A big meanie.

My point is not that intelligence research is impeccable.  My point, rather, is that hostility to intelligence research is all out of proportion to its flaws - and Social Desirability Bias is the best explanation.  Intelligence research tells the world what it doesn't want to hear.  It says what people aren't supposed to say.  On reflection, the amazing thing isn't that intelligence research has failed to vanquish its angry critics.  The amazing thing is that the angry critics have failed to vanquish intelligence research.  Everything we've learned about human intelligence is a triumph of mankind's rationality over mankind's Social Desirability Bias.



COMMENTS (29 to date)
MikeDC writes:

I find this research remarkably "stupid" because you're attributing absolute consequences to relative differences.

Intelligence should be treated the same as income.

As we (at least I) always told my econ students, relative inequality doesn't matter. If the poorest of us is a millionaire and the richest is Bill Gates, nobody should care about inequality. And if we're all living hand to mouth, we probably aren't preening over our highly equal distribution of income.

On intelligence, the operative question (and the one I think people implicitly answer in things like the ANES) isn't whether I'm smarter than anyone else, it's whether I'm smart enough to make good decisions (even though the question is framed as whether the person is smarter than "average").

In any case, we'd be better off figuring out how to structure society so that most even folks of below average intelligence can successfully navigate it, rather than introducing complication upon complication and then using them as rationales for taking choice away from the many and "stupid" and putting it into the hands of the few and "smart".

Hadur writes:

The prospect of public policy based on intelligence research (or social unrest caused by intelligence research) is sufficiently troubling to make me adopt the view that this is one truth worth supressing.

Matt Flipago writes:

They should have put the intelligence in terms of averages, not low and high. It's possible that people think, in general, human adults have a high intelligence, compared to a child, or a dog. This would also get rid of some of the unease when saying they have below average intelligence. What sounds worse, "they appear to have low intelligence." or "they appear to have moderately below average intelligence."?

MG writes:

I am generally sympathetic to the crux of Bryan's point, but I do think that (as MikeDC suggests in his fourth para.) the example cited here muddles the point because we do not really know what the interviewers are interpreting "average intelligence" to mean. I think they are more likely to interpret it to mean a baseline above which other responses are "meaningful and comparable" not as a variable to be more precisely distributed. This can lead to something other than a normal distribution. I think a more interesting experiment would require the interviewer to rank the responders (with ties not allowed). If they gave up -- such as if you would, were asked to rank your children -- then we are closer to establishing unwillingness to see differences, or an inability to see the impact/vilidity of small differences.

Kevin writes:

If everyone is rated above average (even those below average), then the interviewers must have an extremely low view of what the average is (perceived average IQ = 70?).

Or social desirability.

RPLong writes:

To Mike and Hadur's points, an idea I have been thinking a lot about these days is the concept that if a "democracy" isn't universally comprehensible, then it isn't really a democracy.

In other words, if we must become knowledgeable experts in a great many disciplines just to have an informed viewpoint on "the issues," whatever they are, then we're already excluding people of average intelligence, much less people of below-average intelligence.

And is this fair? Is it fair to make politics so specialized that only a minority of people can stay up to speed?

I say no. I say governments should be stripped down to the point that everything is easily understandable to everyone. That means we have to get government out of the sciences, out of engineering, out of specialized market regulation.

If the lower 50% of people can't understand what a favorable outcome actually looks like, then why on Earth should we expect the policy to produce favorable outcomes for anyone?

Brian H. writes:

MikeDC:

Who here is suggesting we take choice away from people below average intelligence?

One of the main reasons to engage in intelligence research at all is to tease out what consequences intelligence does have on everyday life -- i.e., your operative question. This is an unsettled question, and not only because of "Social Desirability Bias" -- it is simply a very difficult question to answer methodologically.

I'm not attributing this to you, but one of my general observations is that people project ideas about the consequences of intelligence research that aren't necessarily as prevalent in the literature as they think -- and object on that basis.

Of course there is the issue of non-academics (and less credible academics) who do take this unsettled research and use it as a rallying call to push their own ideologies. But my point remains.

Hadur:

I'm very thankful that the many scientists whose research has challenged religious orthodoxy (for example) have not followed your logic.

I do agree with several of the other posters that the ANES results may not be an optimal choice to illustrate this point, but I don't think Bryan meant them to be conclusive, either.

Steve Sailer writes:

The most virulent objections to IQ testing come from about the 97th percentile -- the folks smart enough to think that Stephen Jay Gould was a genius.

Les Cargill writes:

I am skeptical of IQ as a concept because it doesn't fit observation well, not because of any "social bias". Increasingly, the sorts of social markers that, say, Charles Murray writes about have very little to do with intelligence and more to do with conformity to relatively arbitrary standards.

There may well be a thing called "IQ" but I don't think that applying ... applied numerology to patterns in filled-in ovals on a card can be defended very easily. It's a reified measure - a high IQ person is someone who scores highly on the test, and the test is evaluated based on the shape of the distribution of answers. it's not a measurement like voltage or GDP, that has a material referent.

that Murray/Herrnstein concluded that race affects IQ *still* is a problem. There's no explaining that away.

Bostonian writes:

Yes, half of all people have a below-average IQ, but that does not justify calling them "stupid".

Hadur wrote, "The prospect of public policy based on intelligence research (or social unrest caused by intelligence research) is sufficiently troubling to make me adopt the view that this is one truth worth supressing."

We already have such policies. No Child Left Behind, which mandates that ALL students be "proficient" and that racial and socioeconomic gaps in achievement be eliminated, is a public policy based on beliefs about intelligence. I think the reason NCLB has failed in its objectives is that those beliefs are mistaken. Telling minorities that they are just as smart as other groups and that their lower academic achievement is due to social inequity is already a cause of social unrest.

Hana writes:

Interesting comments Bryan, but I think the choice of the 2000 ANES doesn't prove your point.

1) If you look at Table 6, page 38, of the survey you will notice that 'some college education' (or more) is overrepresented in all age groups. The lone exception is the segment identified as 'some college education' 20-29.
2) The 'less than High School graduate' is overrepresented in only one age bracket 50-59.
3) Setting aside the issue of 'signaling', one can probably safely assume that a population with more college attendees and college graduates has a higher IQ than one with less educational accomplishments.
4) The 'intelligence' was a subjective assessment made by the interviewer. Since the actual interviews took from 65 to 70 minutes, the interviewers should be able to make a reasonable assessment. However their assessment could have been biased by their own relative intelligence.

I believe in this particular case the population interviewed was likely to have been more intelligent than the US population as a whole. It is not surprising therefore for the bell curve not to fit. As a result I don't believe you can use this case to make a more general case regarding IQ, or intelligence research.

Dan writes:

I wish MikeDC would tell us what a society that most folks of below average intelligence could navigate would look like. Sounds like it would need lots of "guard rails" to prevent the less clever from running off the road, so to speak. Such as not letting people with low incomes take out mortgages with balloon payments at the end of five years, for instance. But who will design and erect these barriers to fatal mistakes? Why, the "smart", that's who, the very blood suckers MikeDC fears. There are so many ways for the dull-witted to go wrong that we will need a raft of regulations to keep them safe from their own folly and from their more clever and predatory neighbors. Hardly a simple society. Or does Mike have a scheme whereby the stupid will draw up clever rules to protect themselves from themselves?

Tom West writes:

Brian H.:

Who here is suggesting we take choice away from people below average intelligence?

Where we're talking about group differences, we already have numerous commenters indicating that some groups are inferior enough that we should be denying (or at least impeding) immigration from those groups.

Once it's made explicitly clear that we as a society openly classify certain groups as inferior or superior, you don't think limitation of choices inevitably follow?

Of course there is the issue of non-academics (and less credible academics) who do take this unsettled research and use it as a rallying call to push their own ideologies. But my point remains.

"there is the issue"? Crikey. It's like developing a portable power source that currently discharges a little quickly so its only use is to allow any random person to blow up a city. Sure, the long-term outcome might be interesting, but I'm damn sure this won't be a net positive...

It would also help if we weren't at the "phlogiston" level of brain/IQ research, were so little is actually certain that people can basically push any agenda they want and paint it with "science".

Bostonian:

Telling minorities that they are just as smart as other groups and that their lower academic achievement is due to social inequity is already a cause of social unrest.

Because official sanctioning groups as superior or inferior always results in social peace. Well, at least it would if only those other folks would know their place...

Brian H. writes:

Tom West,

  • When I began to write my comment only MikeDC and Hadur had posted.
  • You're "damn sure" it won't end up in a net positive? I'm curious how you can be so confident of that.
john hare writes:

Your capacity to consider some people as stupid increases remarkably when you spend a few self employed years trying to get quality construction work out of day laborers. Not all by any means, but a really unfortunate percentage.

Bostonian writes:

Tom West wrote, "Because official sanctioning groups as superior or inferior always results in social peace. Well, at least it would if only those other folks would know their place..."

I don't think the government should sanction any group as "inferior", not just reasons for tact and social peace but also because intelligence is only one human attribute. The possibility of racial differences in intelligence just means that the government should not MANDATE equal outcomes, as NCLB does and as "disparate impact" regulations do indirectly.

Tom West writes:

I'm curious how you can be so confident of that.

Well, in my admittedly over-the-top example, there wouldn't be any cities left. In group IQ research, it's because we are a *long* way from any practical aspect outside of trying to use the data (which I find a bit dodgy most of the time) to prove which group is superior/inferior.

I don't think the government should sanction any group as "inferior", not just reasons for tact and social peace but also because intelligence is only one human attribute.

Let's cut to the chase. Intelligence *is* pretty much used as the ruler for superiority/inferiority. Just read people's comments here. They're usually too polite to actually use the words superior / inferior, but they've got IQ walking and quacking like those words.

You'll find that same ruler used pretty much universally. If you're going to go to jail for killing someone, make certain they're mentally handicapped, as physically handicapped won't save you jail time, but mentally handicapped likely will.

As for NCLB, think of it as a tug of war. NCLB's unrealistically high expectation are simply the other end of the rope that has pathetically low expectations on the other end. Without NCLB's pull, the inertia and laziness that results in abysmal results would continue indefinitely. In a tug of war you don't achieve a decent equilibrium by standing right where you want that equilibrium to be, when there are powerful forces pulling on the other end of the rope. It's just reality.

(It's why, while I'm no Libertarian, I'm happy to see Libertarians staking a claim well beyond what I personally consider reasonable. It's the best hope of having an actual equilibrium where I want it...)

Loof writes:

@ Steve Sailor who says: “The most virulent objections to IQ testing come from about the 97th percentile -- the folks smart enough to think that Stephen Jay Gould was a genius.”

Steve likely has some smarts, some sort of intelligence, perhaps even measure above average on an IQ intelligence test. However, by saying what he does here, he's blind as to how he is stupid, which could be measured by the science involved with Multiple Intelligences (MI). Or to test for another sort of intelligence would come out as inferior as subjectively measured by the science involved with Emotional Intelligence (IE).

Peter writes:

The problem I am running into John Hare is trying to figure out if the reason the person I am dealing with is useless to the point they can't even stack two same size boards on top of each other and continues to fail so after training, retraining, chances, more chances, physical walkthroughs, etc is are they:

1: Willful ignorant
2: Low intelligence
3: Willful counterproductive
4: Lazy

Let's face it from a results perspective we write all four off as *stupid* but not all correspond to low IQ. This scenario also applies to more complex tasks. For much of my life I just figured people were most 2 or 4 but as I get much older I'm thinking it's mostly 1 and 3. For most tasks attitude beats ability in my opinion and IQ doesn’t measure this well (nor is it meant to). I agree plenty of people are stupid but not sure that matters at the end of the day for most things. Horse Water Drink etc problem.

john hare writes:

Peter,
My opinion is that the ranking would be 1,4,3,2 on your scale. I believe that many of the ones we are discussing never put forth the effort to learn to think. The practice of thinking things through is, IMO, at least as important as inherited ability that is called IQ.

Brian writes:

MikeDC, I don't follow your reasoning at all.

First of all, I didn't see anything to suggest that Mr. Caplan's point was to attribute absolute consequences to relative differences. Please, direct my attention to it if I've just missed it?

But even if that were his point, are you sure there no absolute consequences to being relatively "stupid?" Or do I need to be PC and restate that question as, "are there no absolute consequences to being considerably less smart?"

What proof or reason is there that intelligence be treated as income? Says who?

Likewise, who says what "the operative question" is? In a competitive situation, might not whether "I'm smarter (or faster) than anyone else" be the operative question as to whether I'm the one eaten by the predator, so to speak.

Lastly, I saw no suggestion that differences in intelligence was suggested to be a rationale for taking any choices away. Again, if I missed that point, please show it to me.

Thanks if you read this far.
Brian

MikeDC writes:

@ Brian H,
Since "here" is the internet, there are lots of folks saying that. More seriously, I have heard lots of suggestions here in the past about limiting voting and occupational choice, and decisions made about health, education, and a variety of other things based on the supposed (lack) of intelligence of the folks involved.

At a fundamental level, a market economy is based on the proposition that individuals are rational and will make value-adding trades. To the extent that people are "stupid", a market economy can't be expected to function very well.

@ Dan
Lots of choice-enabling simplification would be beneficial to both the smart and stupid. The fact that I'm smart enough to figure out how to correctly do my taxes doesn't mean I wouldn't also benefit from a simplified tax code. I'd save days of time. Health care regulations, immigration law and government sponsored loans (both for education and housing) would also be prime targets, along with the sort of regulatory gamesmanship RPLong mentioned, which is largely not the smart praying on the stupid, but the smart employing their rent-seeking talents against each other. Rolling that back and channeling those talents toward productive ventures would, however, be something you'd think all the smart folks like you should be capable of figuring out.

MikeDC writes:

@ Brian,

He says, "Take the truth, 'Half the population is below the 50th percentile of intelligence.' It's unequivocally true - and sounds awful. Nice people don't call others stupid - even privately."

The first sentence notes a relative difference between people. The third sentence equates stating that relative difference to stating an absolute- "stupidity". If we were locked in a room with Stephen Hawking and Issac Newton, we'd be below the 50th percentile of intelligence. However, it doesn't mean we're stupid.

There absolutely are consequences to being stupid. That's exactly why it's important to frame the question accurately. What sort of intelligence does a society require in order to fully participate?

Regarding the other stuff, Economic theory (at least as a science) is fundamentally constructed on a utilitarian framework. Shifting around the intelligence of actors by definition shifts the utility maximization outcome. I don't think that's per se wrong (and I'm not opposed to intelligence research), but I think we ought to be damn careful about it.

Loof writes:

MikeDC asks an excellent question: What sort of intelligence does a society require in order to fully participate?

Simply put: A general intelligence that can resolve practical problems like a jack-of-all-trades whose an intellectual master of none. Newton and Hawking are intellectual (seen as highly specialized intelligence) geniuses, amongst the greatest humanity has ever produced, but suspect they’d be stupid if they tried to resolve real problems in society. Evidently this is so with academics in general. Rarely do professors fully participate in society sensibly.

Brian H. writes:

MikeDC, when I said "here" I meant this post and thread, though as Tom West mentioned my statement became incorrect when comments to that effect were posted, including while I was writing my first post.

My point was that augmenting an argument with a response to a point that has not been made is not fair, though I do not think you were being disingenuous in any way.

Also, I don't know if it's legitimate to the make "here" the entire Internet given that you can trivially find virtually any ludicrous opinion stated by someone somewhere.

caltrek writes:

My, my, where does one begin.

Since we all seem to agree Einstien was a smart dude, maybe we should think about his concept of "relativity".

A test taker is smart and/or stupid relative to other test takers. So far so good. He (or she) is also smart and/or stupid relative to the person(s) who designed the test.

Here we come into problems. My father tested out at a higher IQ than myself. The secret he shared with me, understanding the psychology of the person(s) who designed the test. What answer(s) were they looking for?

Gee guys, could it be that the answers they are looking for are wrong?

Suppose I am given a test. I can either give the answer I suspect is expected of me, and be rated as "smart". Or I can give what I think is the correct answer, and risks being rated as "stupid".

So the "smart" guys decide to agree with BS, and the dumb guys have a more creative approach to the problem(s) being posed. Can you begin to see the problem with metrics that this poses?

Social Desirability Bias - interviewers' reluctance to impugn anyone's intelligence - practically has to be the explanation.

There's a simpler explanation: underestimation of the average level of inntelligence (due either to a self-enhancing bias or to propaganda from both liberals and conservatives about the masses' stupidity).

TerjeP writes:

Another possibility is that the interviewers are as dumb as doornails. However you probably didn't consider that option because it seemed a bit rude.

WMD writes:

There is no such thing as objectively measured intelligence. IQ is a deeply flawed concept based on subjective criteria, namely, the capacity for a certain type of rational or logical thinking.

Intelligence is not something that can be measured. It is essentially the creative capacity of an individual to express her talents and abilities. In that sense, there can be no classification of people of lesser or greater intelligence, only of those with more or less developed potential. To think otherwise is foolish at best and eugenic at worst.

Comments for this entry have been closed
Return to top