The Bell Curve

[_Always Changing]

It is a book of exaggerations, written up to reinforce negative expectations based on social class.

[_Analytics]

It is a book of exaggerations, written up to reinforce negative expectations based on social class.

I'd prefer to refrain from judging the authors' intentions. I can definitely see how a lay reader could go from hearing "IQ is statistically significant and has the most explanatory power" to thinking "epsilon is close to zero" i.e. "IQ is "determinant" of success." Hell, I can clearly see how experts could get swept away in moments of carelessness. But the questions remain: what does the actual research imply, and what does the book actually say?

Maybe it is explicitly, emphatically, and repeatedly saying one thing to create plausible deniability for insidiously implying something else. I'm generally not that cynical though.

[_Gadianton]

Still no chapter 4? I don't think it's too much to ask for 4 chapters a day from a cognitive elite. I'm really enjoying the summaries to see how it unfolds compared to what I remember from the Geraldo episode. I have to admit so far, the Geraldo episode missed some interesting pieces.

I apparently missed the premise to to this thread which was a dispute between EA and A over a Sam Harris blog -- EA being slightly anti-new atheist and A being slightly prone to new atheism. That kind of sucks as when the thread started I thought A had discovered the book somewhat like vincenzo defrancesco discovered the Book of Mormon, without any context, and so here was going to be an fascinating review of a book that A doesn't know is dripping in controversy, and given his profession and equality leanings, what does he come up with?

So I won't get the contextless review, but it's still fascinating reading because I have to figure out a way to interpret the argument between EA and A totally based on prejudice either for or against new atheism. I don't have an answer yet.

[_EAllusion]

So I'm having a hard time interpreting what you mean when you say they imply an "exaggerated" relationship

The papers I've seen put IQ correlates with income distribution at the low end of their range.

The implication, which you are running with in the circumstantial case you've read, is that class stratification is heavily determined by differences in IQ. AFQT scores explain something in the neighborhood of 15% of the variance in income, and AFQT scores are already pregnant with dependence on non-fixed variables such as quality of education, stable parenting styles, etc. It's not nothing, but it's small enough to not be helpful in some of assertions you are making, such that it is hard to be a senior manager without having a high IQ, or pointed questions about how the top 10% are doing. Another way of saying it is higher IQ - having the qualities that lead to performing well on IQ tests - buys a modest income premium.

Humorously, when looking up studies on correlates between AFQT and wealth, the very first study that popped up was this one:

http://www.sciencedirect.com/science/ar ... 9607000219

It argues for a modest, but notable relationship between IQ derived from AFQT scores and income, but no relationship between net worth and IQ because people with higher IQ make decisions that cause more financial distress. In other words, while performing better on IQ measures is modestly predictive of a higher income, that gets canceled out because it's also modestly predictive of being having more negative cash flows in the form of credit card debt, alimony, student loan payments, etc. I laughed.

[_EAllusion]

Circling back to the Thomas Sowell review for a second, one of the things the authors of the Bell Curve are going to argue is that cultural biases on IQ measures are a red herring / not relevant, because the submeasures that you'd expect to be the least culturally dependent, and the ones that correlate best with g in factor analysis, are the ones that the poor performing ethnic groups do the worst on. In other words, while performance on vocab dependent tests that by necessity are going to have some bias related to one's language experiences are worse, more abstract measures like spatial reasoning are even worse than that. This isn't what you'd expect if the problem was just cultural biases built into the tests.

What Sowell points out in reply is that when you look at IQ testing in different culturally isolated populations, usually between between socially favored and disfavored groups, for whatever reason the disfavored group always seems to do worse in the most abstract components.

While Herrnstein and Murray do an excellent job of exposing the flaws in the argument that tests are culturally biased by showing that the greatest black-white differences are not on the questions which presuppose middle-class vocabulary or experiences, but on abstract questions such as spatial perceptual ability, their conclusion that this "phenomenon seems peculiarly concentrated in comparisons of ethnic groups" is simply wrong.

When European immigrant groups in the United States scored below the national average on mental tests, they scored lowest on the abstract parts of those tests. So did white mountaineer children in the United States tested back in the early 1930s. So did canal boat children in Britain, and so did rural British children compared to their urban counterparts, at a time before Britain had any significant non-white population. So did Gaelic-speaking children as compared to English-speaking children in the Hebrides Islands. This is neither a racial nor an ethnic peculiarity. It is a characteristic found among low-scoring groups of European as well as African ancestry.

In short, groups outside the cultural mainstream of contemporary Western society tend to do their worst on abstract questions, whatever their race might be. But to call this cultural "bias" is misleading, particularly if it suggests that these groups' "real" ability will produce better results than their test scores would indicate. That non sequitur was destroyed empirically long before Herrnstein and Murray sat down to write The Bell Curve. Whatever innate potential various groups may have, what they actually do will be done within some particular culture. That intractable reality cannot be circumvented by devising "culture-free" tests, for such tests would also be purpose-free in a world where there is no culture-free society.

His point seems to be that those subcultures have less experience with the sort of reasoning or testing of reasoning that happens in abstract portions.

In other venues, he's apt to point out that when IQ testers meet with black students to acclimate himself to them for a little bit prior to testing, their IQ scores tend to improve significantly. One potential implication of that may simply be that the situated social status of the testee may impact performance akin to what happened in in Jane Elliot's infamously unethical "blue eye - brown eye" semi-experiment. Is that a cultural bias of test construction? Not really. Is it a cultural bias confound of measurement? Yeah, I think that's a reasonable statement.

So when you say,

He also claims that there is an entire body of research that indicates that a "properly administered" IQ test isn't biased on race, etc.

I think you might need to take a wider view of the notion of "biased on race." Do these phenomenon explain the entirety of the gap? No, I don't think any research bears that out. Is it correct to say that modern IQ testing is bias free when it comes to racial experiences? No, that's dubious.

[_EAllusion]

Still no chapter 4? I don't think it's too much to ask for 4 chapters a day from a cognitive elite. I'm really enjoying the summaries to see how it unfolds compared to what I remember from the Geraldo episode. I have to admit so far, the Geraldo episode missed some interesting pieces.

I apparently missed the premise to to this thread which was a dispute between EA and A over a Sam Harris blog -- EA being slightly anti-new atheist and A being slightly prone to new atheism. That kind of sucks as when the thread started I thought A had discovered the book somewhat like vincenzo defrancesco discovered the Book of Mormon, without any context, and so here was going to be an fascinating review of a book that A doesn't know is dripping in controversy, and given his profession and equality leanings, what does he come up with?

So I won't get the contextless review, but it's still fascinating reading because I have to figure out a way to interpret the argument between EA and A totally based on prejudice either for or against new atheism. I don't have an answer yet.

I had an opinion on the Bell Curve from many years ago. Sam Harris out of the blue deciding to promote Charles Murray's IQ arguments a few weeks ago I thought was a dead-to-rights example of how he sucks. I didn't expect to even be challenged on that point by Analytics.

I like Daniel Dennett, often viewed as one of the four horseman of new atheism. My main complaint with him I think he's a bad public communicator when actively engaging people like Alvin Plantinga. But generally speaking, good dude.

Sam Harris is just the worst, though. I think he's bad at what he does and reinforces negative stereotypes of atheists as smug, philosophically shallow dilettantes who are secretly prone to religious thinking and are morally questionable. He's what someone like Tal Bachmann is to the ex-Mormon community.

[_Lemmie]

....In reply to my request that you be more specific in your criticisms, you said:

They imply a exaggerated relationship which you have been repeating here to the point that it is determinant. Determinant means, in this context at least, a factor that is important enough to be decisive in observed outcomes...

I need you to be more specific. The book says, specifically, "The correlations between IQ and various job-related measures are generally in the .2 to .6 range." They do an excellent job of explaining in lay terms how to interpret that.
So I'm having a hard time interpreting what you mean when you say they imply an "exaggerated" relationship. It seems like an earnest attempt at describing the actual research. Perhaps you are saying that almost all of the results are in the .2 to .3 range, and stuff in the .4 to .6 range are outliers? Therefore saying measures are generally in the .2 to .6 range is an exaggeration?

Analytics, I'm having a difficult time with your description of the statistics, and why you object to EAllusion's "exaggerated relationship" comment, specifically this:

And they repeatedly say that while IQ has the most predictive power of all of the variables in the dataset, individual factors that don't come through in the statistics are collectively the driving determinants of success.

So if individual factors, NOT captured by the statistics, are the "driving determinants" of success, then the only way that IQ could still have "the most predictive power" would be if those individual factors WERE correlated with IQ. But the main argument of these outliers, if I am reading you correctly, was that these individual factors were NOT correlated with IQ. So, which is it? IQ predicts? Or individual factors predict? Or is it that the authors are carefully asserting no individual is being talked about here, but their analysis still holds for groups and averages, hence EA's "exaggerated relationship" comment?

From your quote:

"The exceptions associated with modest correlations mean that a wide range of IQ scores can be observed in almost any job....

The exceptions do not invalidate the importance of a statistically significant correlation." (page 67)

So if it is only a "modest correlation" defining this relationship, why is the entire book, so far, predicated on fully accepting this relationship and its predictive power, including averages, for specific groups of people?

[_Always Changing]

My "blacks have a genetic predisposition towards lower IQ" professor always taught that there is considerable overlap in the black/white curves. Any given black could easily have a higher IQ than any given white. It is just a bit more likely for the white to have a higher IQ. So the message is softer than color-coding for IQ.

Again, "Bell Curve" was written to justify cutting off funds for leveling the playing fields between social classes.

[_Analytics]

Chapter 4: Steeper Ladders, Narrower Gates
This is the final chapter of Part 1, which attempts to describe the emerging world of the cognitive elite. I believe it's the longest chapter so far, and also has the most footnotes.

It starts out by saying, "We are not talking about what should have been but what has been." The basic gist is that having a high IQ has a very economic value that is increasing. While we aren't there yet, we are moving in the direction of a meritocracy where everybody who has a high IQ, regardless of their race, religion, sex, or background, will find there way to congregate in schools, jobs, and neighborhoods with other people with high IQs.

We are optimistic enough to believe that no administration, Left or Right, is going to impede the education of the brightest, or forbid the brightest from entering the most cognitively demanding occupations, or find a way to keep employers from rewarding productivity. But we are not so optimistic that we can overlook dark shadows accompanying the trends.

To this point, we have avoided saying what social consequences might be expected. This omission has been deliberate, for part of a candid answer must be, "We aren't sure." We can be sure only that the trends are important. Cognitive stratification as a central social process is something genuinely new under the sun. One of our purposes is to bring it to public attention, hopeful that wisdom will come from encouraging more people to think about it.

The basic argument of this chapter is an economic one. The world is getting very, very, complex, and the economic stakes are getting very, very, high. Because of this, there are huge competitive advantages and economic returns to having the best and brightest navigating companies through these complexities. As the world gets more complex, the economic value of being especially bright will continue to increase. I find the argument compelling. In my line of work, if you are a qualified actuary, nobody cares at all if you also have a Ph.D. Employers won't pay an extra nickel for those letters. What they care about is whether you can solve complex problems and then articulate the results--i.e., whether you have a high IQ.

There are a ton of footnotes in this chapter. I'm sure one could do an entire dissertation going through the finer points of this.

From what I've seen, people who do really well financially--i.e. top 2% of earners--are either really smart, really good at personal relationships and selling, or both. But even if you are good at selling, you probably aren't going to make $300k a year in our economy as a salesman unless you are smart enough to master complicated products in complicated markets and figure out solutions to complicated customer problems.

The basic idea is that the world is changing. The smartest kids are finding their ways to the best schools where they then get the best jobs, and then marry other people that did likewise. They then settle down in expensive neighborhoods with other very smart people, and then send all of their kids to school together. The smart kids born out of these neighborhoods will likely find their way to join their intellectual peers.

Is this book written for angry bigots who are looking for an excuse to look down upon and exclude blacks? Consider how the chapter ends, and recall that this was written two years before the launch of Fox News in 1996:

These phenomena are driven by forces that do not lend themselves to easy reconfiguration by politicians. As we leave Part I, here is a topic to keep in the back of your mind: What if the cognitive elite were to become not only richer than everyone else, increasingly segregated, and more genetically distinct as time goes on but were also to acquire common political interests? What might those interests be, and how congruent might they be with a free society? How decisively could the cognitive elite affect policy if it were to acquire such a common political interest?

I wouldn't expect to hear that line of questioning on FOX News. I'd expect to hear it in The New York Times about FOX News.

[_Analytics]

So if it is only a "modest correlation" defining this relationship, why is the entire book, so far, predicated on fully accepting this relationship and its predictive power, including averages, for specific groups of people?

Great question. The way these models work is that you have a big list of people in a population. You want to see if you can predict something for each person in the data (e.g. Y = income) by other elements of data (e.g. X1 = IQ, X2 = education level, X3 = race, X4 = income of parents, X5 = ZIP code, etc.). You do some fancy math and make an analysis about various pitfalls and see if there is a relationship. The relationship is going to be of the form:

Y = B1*X1 + B2*X2 + B3*X3 + B4*X4 + e

In this example, X1, X2, X3, and X4 are four specific factors that vary by individual, and Y is each individual's income. If the relationship is actually there and the formula is set up correctly, you can calculate what B1, B2, B3, and B4 are--those are the factors that make the prediction. By the math, you can calculate what they are and see if they actually have the ability to help predict Y if you know what B1, B2, B3, etc. actually are.

The final term is "e". This is the residual. The idea is that when you do the study, X1, X2, etc. can help predict Y, but they will rarely predict it exactly. "e" is the missing piece--the part that isn't explained by the model. If the model is good, then "e" will be normally distrusted around zero and independent from person to person.

What the book is saying that out of all of the factors we have in our data, IQ does the very best job of predicting success. It doesn't predict it perfectly--some people with low IQs make good, and others with very high IQs are colossal failures. Everybody does at least a little better or a little worse than what the model predicts. Despite that, there is a clear, mathematical relationship and that out of everything that is modeled, IQ is the most powerful indicator of success.

So how should this be applied? When this is applied to the population as a whole, it gives very powerful results. Have you seen the movie Money Ball? It is kind of like that. On any given baseball play anything could happen. But over the course of a season, patterns emerge. If you understand those patterns, powerful things can happen.

Does that help?