The Cult of Smart by Fredrik deBoer, 2020, St. Martin’s Publishing
What is intelligence?
A book whose thesis seems to be ‘the unintelligent will always be with us, so it’s our moral duty to take care of them’ should spend some time carefully defining terms, but instead deBoer uses a slew of synonyms, not unlike Justice Stewart’s definition of obscenity: “I know it when I see it.” The problem with this kind of sloppy definition is that it leads to sloppy thinking, as we shall see.
Let us define intelligence, then. The Oxford English Dictionary has several definitions, but for our purposes, I’ll take two. First, towards the end of the fourteenth century: “The faculty of understanding; intellect.” Second, towards the end of the fifteenth century: “Knowledge concerning events communicated by or obtained from another; information, news; spec. information of military value.” Keep these in mind as we proceed.
deBoer puts great stock in the measurement of intelligence, going so far as to use IQ interchangeably throughout the book, beginning on page 23, in the Introduction. While deBoer says he has some expertise in psychometrics, and I have a few graduate classes in statistics myself, most people do not.
Forgive me, for we are about to get technical.
A century ago, eugenics was all the rage. Eugenics is a kind of meritocracy–those who scored well on tests were to be rewarded. Francis Galton, the major proponent of eugenics, advocated for individuals who scored well to be encouraged to reproduce–and those that did not score well to be forcibly prevented from reproducing. Like an early Mark Zuckerberg, he rated women as hot or not, but he didn’t stop there. In fact, he set up a lab where people paid for the novelty of being scientifically measured. Galton leaned hard on the “nature” part of the “nurture versus nature” debate (in fact, he came up with the entire debate, including the phrasing). He aimed to prove it with statistics, and he had true believers at his side. These ideas led directly to the Nazis.
From 1933 to 1945, Nazi Germany carried out a campaign to “cleanse” German society of individuals viewed as biological threats to the nation’s “health.” Enlisting the help of physicians and medically trained geneticists, psychiatrists, and anthropologists, the Nazis developed racial health policies that began with the mass sterilization of “genetically diseased” persons and ended with the near annihilation of European Jewry. With the patina of legitimacy provided by “racial” science experts, the Nazi regime carried out a program of approximately 400,000 forced sterilizations and over 275,000 euthanasia deaths that found its most radical manifestation in the death of millions of “racial” enemies in the Holocaust. (USHMM)
I point this out not because I want to embody Godwin’s Law, but to give context for intelligence tests. They were designed to prove Galton’s theory that well-educated, upper-class, straight, white British men were the pinnacle of society. Spearman, a eugenics advocate, did this by noting that these individuals scored well on certain test items. He also found that someone who scored well on one set of test items usually scored well on the other set of test items. Then, he juked the stats so hard to find the statistical connection between those test items that he invented a whole new branch of math, called “factor analysis.”
Spearman was undoubtedly brilliant, as well as evil. He went looking for a pattern in the data to justify why these upper-class, straight, white British men scored so well, and so consistently, across school subjects, and he found one, which he called g. I would note that the human mind is excellent at pattern recognition, to the point that we sometimes see a pattern where none exists.
It’s important to remember that Spearman found a statistical artifact in data. To this day, nobody can agree what g is, other than a pattern in the data. In fact, a handful of years after Spearman came up with g, Thomson pointed out that, basically, stuff happens, and that this kind of statistical artifact can happen even without g actually existing. Personally, I think Spearman went looking for the common factor among these upper-class, straight, white British men, and found it. Unfortunately, this eugenics goal has been used to measure humans ever since.
g is what intelligence tests purport to measure. From Spearman on, all intelligence tests are synched to this statistical artifact. If a test doesn’t find g, it’s not considered to be a valid intelligence test. Remember that there were multiple factors, and the degree to which a factor correlates, or goes with, is called “g loading.” We’ll come back to this.
Before we do, all introductory stats professors would like you to know that “correlation does not indicate causation.” Really. The classic example that my intro to stats professor used was that of ice cream and drowning deaths. They are very highly correlated, but one does not cause the other. Instead, people eat more ice cream and go swimming more often during the summer.
In other words, just because g exists, it does not follow that the statistical artifact of g causes people to do well on these intelligence tests. They’re just correlations. Also important is that we don’t know how g is distributed in humans, because of the nature of it as a mathematical construct. Yes, yes, your IQ score is distributed on a bell curve, but that’s because the test designers make it this way, not because we know that’s true.
- Eugenics is advocating for controlled selective breeding of human populations (with straight, white, well-educated British men of 100 years ago as the ideal)
- There is a straight line between eugenics and Nazis.
- Eugenicists came up with intelligence tests to justify their selective breeding ideas.
- g is a pattern in the eugenicist’s data.
- Intelligence tests are said to measure g.
- g has no generally agreed-upon definition.
- Correlation does not indicate causation.
- We don’t know how g is distributed in the general population.
The tragedy is that we’re still using these virtually unchanged IQ tests to rank and sort ourselves. For example, the SAT was designed as an intelligence test, and is still highly correlated to IQ, and in fact, your SAT/ACT scores can be converted to an IQ score with a basic equation. Somewhere between half and three quarters of all our high school graduates subject themselves to the SAT/ACT every year, carrying out the eugenicists’ mission of ranking students for meritocratic rewards. It’s worth examining how the researchers determined this. In the study, the researchers used the Raven’s Advanced Progressive Matrices test to correlate SAT scores and IQ.
What is the Raven’s Progressive Matrices (RPM) test? It’s a non-verbal IQ test, virtually unchanged since it was first published in 1938. It consists of a pattern recognition problem set, growing more difficult as you progress through the problems. It’s considered to be highly correlated with g and widely used because it’s non-verbal. The US Army used it in World War II, to sort and rank draftees, and Flynn used it to prove that IQ scores have been dramatically rising over time.
As a side note, autistics score an average of 30 percentile points higher on the RPM than they do on the WISC. This result has been widely duplicated, including in other languages. The WISC is the most widely used intelligence test in the world, and it is therefore the reason that autistics are widely considered to be mentally retarded. (I know, it’s a horrible word, but it is the official term of art when dealing with intelligence scores.) However, autism is in part a communication disability, and the WISC relies strongly on written and oral communication. In fact, many autistics are non-speaking all or some of the time, and the WISC penalizes them accordingly.
On the top is a sample problem from the RPM set, and on the bottom is my seven-year-old’s math work from last Thursday. I’ve highlighted the relevant bit from the homework in yellow. She does one of these pattern recognition problems every day, as part of a less commonly used but highly regarded curriculum.
You can see the similarity. Another argument against IQ as a valid measure of anything other than a statistical artifact is the correlation between IQ and reading comprehension. In fact, Jensen, not exactly a radical leftist, says, “It is common knowledge in psychometrics that a standardized test of reading comprehension is a good proxy for an IQ test.”
As students of the science of reading know reading is more than just accurate and automatic decoding. In their influential 1986 journal article, “Decoding, Reading, and Reading Disability,” Tunmer and Gough proposed a model for reading comprehension. Famously, this is the “Simple View of Reading.” In the article, Tunmer and Gough give a formula that explains how reading works: word recognition skills x language comprehension = reading comprehension
As our children progress beyond phonics, many summative reading comprehension tests are actually tests of background knowledge. Famously, Recht and Leslie did a reading comprehension study back in the late ‘80s that demonstrates this idea. This study is sometimes referred to as “the baseball study” because students’ reading comprehension was tested using a passage about baseball. Children who were good readers with lots of knowledge about baseball performed best of all, but surprisingly, children who struggled to read and had lots of knowledge about baseball came in a close second place, well above good readers with little knowledge about baseball. As Daniel Willingham notes in the Spring 2006 edition of American Educator, the close correspondence between background knowledge and test performance has been replicated many times.
|Good Readers||Struggling Readers|
|Lots of knowledge about baseball||1st place||2nd place|
|Little knowledge about baseball||3rd place||4th place|
Therefore, if background knowledge undergirds reading comprehension, and reading comprehension is a proxy IQ test, then background knowledge leads to IQ scores. Obviously, part of the reason that IQ scores have been rising is that we’re assimilating our children into the standard background knowledge that early 20th century eugenicists set. This idea that intelligence is knowledge is not a new perspective. Recall that the second definition of intelligence is “Knowledge concerning events communicated by or obtained from another; information, news; spec. information of military value.”
deBoer, and others who take IQ test results as an immutable facet of humanity, would argue that IQ is some kind of innate property, that you have it, or you don’t. He spends chapter five, Does School Quality Matter? Not Really, arguing that non-school factors matter much more to student success than the quality of the school, or test prep. More specifically, he votes for “student ability.” In fact, he seems to believe that “thanks to the heritability of academic ability, the range of the possible in the classroom is dramatically smaller than conventionally assumed.” (pg 121)
In fact, a meta-analysis from 2018 concludes exactly the opposite, that education positively affects student ability: “consistent evidence for beneficial effects of education on cognitive abilities of approximately 1 to 5 IQ points for an additional year of education.” This is, after all, the major, generally accepted reason for the Flynn effect, which deBoer fails to address in his explanation of the Flynn effect on page 45. deBoer’s insistence that education doesn’t matter is also misogynistic, given that the majority of teachers worldwide are women, and that proportion rises to three-fourths in the United States–apparently, he’s content to ignore the contribution of women to assimilating our children to this eugenicist standard.
Less obvious is that the whole argument for valuing g is misogynistic from top to bottom. Remember, the standard was set for those upper-class, straight, white, well-educated British men, so the original tests wouldn’t have included test questions that allowed women to score equally–and they didn’t. Spearman’s g was explicitly linked to academic subjects, and women didn’t achieve educational parity at the high school level until the 1960s, undergraduate level until the 1980s, and doctoral degrees in the 2000s. Coincidently, this is about the same time that women achieved parity on these tests.
The problem with “heritability” (statistics is not like arithmetic)
One more particularly thorny term needs to be defined, in order to properly counter deBoer’s argument. That is heritability. deBoer and others often confuse heritability with hereditary, but they are not even close to being in the same ballpark. Once more, I ask for your patience, as this is a bit technical.
I’ll start with Razib Khan’s excellent question:
When someone tells you that height is 80% heritable, does that mean:
- 80% of the reason you are the height you are is due to genes
- 80% of the variation within the population on the trait of height is due to variation of the genes
Many, including Charles Murray of The Bell Curve fame, would answer A. Specifically, Murray wrote, “ . . . when we say 60% heritability, it’s not 60% of the variation. It is 60% of the IQ in any given person.” This is exactly wrong. Let’s talk about why.
Heritability is a term of art from statistics, and much like scientists deeply regret using the word “theory”, geneticists are deeply regretting using the word heritability. No one is arguing that children don’t inherit something from their parents–up to and including social status and financial assets. Geneticists are particularly interested in genes. The general argument is that an organism’s traits are dependent on its genes, which intuitively makes sense. You may remember Punnett squares from high school biology:
But, and this is important, “Methods derived from population genetics to assess heritability provide no information about the causal mechanisms contributing to the development of an individual’s traits. Population geneticists study the patterns of transmission of traits in populations from one generation to the next.”
Again, we’re doing pattern analysis with statistics, and correlation does not indicate causation.
One of the first analysis problems is that humans are almost never binary systems. For example, my husband has brown eyes and I have blue eyes. In a binary set-up, my children would have either brown eyes or blue eyes. Since brown eyes are a dominant gene, they’d have a 75% chance of having brown eyes. But, since humans don’t work this way, my older daughter actually has hazel eyes, brown with a single ring of green. This sort of inheritance pattern is not unusual.
This creates a more complicated math problem, because most human traits are actually “continuous.” A little bit of this, a little bit of that, and we can stick the distribution of that trait on a bell curve. Very few people are going to be super tall, and very few people are going to be super short.
Slope and best fit lines
Originally, heritability was studied in plants, and occasionally domesticated animals. This is important, because geneticists could control variables in a way that they cannot with humans. The classic example, one that even middle school students do for science fair projects, is the height of plants. As a science teacher, I encourage my students to grow several sets of plants for this, because small sample sizes are prone to error.
Then, once my students have grown their plants, in carefully controlled environments, we can graph the results. Geneticists will graph the second generation on the y-axis, or vertical axis, and the first generation on the x-axis, or horizontal axis. You’ll get some sort of Jackson Pollack looking scatter plot, and then you create a “line of best fit”, which is to say a line that is drawn through your dots in a way that minimizes the distance from the line to every dot, over all. I’m so old we used to use rulers for this, but these days Excel or statistics software will happily calculate this for you.
Now, here’s where it gets interesting. If the line is perfectly horizontal, the first generation is irrelevant. If the slope is perfectly vertical, it’s entirely due to genetics. So the heritability is a measure of the slope of the line, and the bigger the slope, the higher the heritability.
I’m afraid we have to talk about variance. If our hypothetical middle school student found the average (mean) height for their plants, I would ask them to subtract the difference between each plant and the mean. Then they would square each result, to make it positive. Finally, I’d ask them to take the average (mean) of that set of differences. That’s the variance, and it reflects how far away the plants’ heights are from an average height of the plants. Naturally, some very tall plants in a sunny corner will mess with your variance and throw it off, because you didn’t adequately control all the variables.
Nevertheless, behavioral geneticists persist
So, population geneticists use a couple of general principles when doing their analyses. First, remember how Galton set up this nature versus nurture argument? Modern geneticists are over here copying him. At its most basic, their equation says that the variation between individual plants is either the genes or the environment, and never the twain shall meet. This is a problem, because in humans and other organisms, genes are influenced by the environment. One classic example is that of Himalayan rabbits, which develop black ears, noses, feet, and tails when reared in cold environments, and completely white coats when reared in warm environments.
Nevertheless, behavioral geneticists persist, and so too shall we. They go on to say that heritability is the fraction of a trait attributable to genetics:
genetic variance / trait variance = heritability
Other interesting variations include the idea that a trait is influenced by more than one gene, like eye color. Behavioral geneticists made up a fraction for that, too:
additive genetic variance / trait variance = heritability
What about dominant genes, like brown eyes? Got one for that, too!
additive genetic variance + dominance variance / trait variance = heritability
Behavioral geneticists, do in fact, admit that the environment affects genes, and they call that VGxE or variance from gene/environment interaction. If genes and environments vary together: for whatever reason, little plants select poor soil to grow in and ignore rich soil, there is a variable for that, too, called gene-environment covariation: COV(G,E)
All of this is to say, there are a lot of factors, and the math gets very hard, and you’re probably not doing it by hand. For some reason, evolutionary biologists tend to assume that the interaction between genes and environment is minimal and that genes and environment don’t often vary together, so they kind of ignore that. On the other hand, behavioral geneticists are really into
genetic variance / trait variance = heritability
and they call that h2 .
Twin studies (damn lies)
Back to Galton, again. Really, he’s the original axis of evil. Galton hypothesized that since identical twins have identical genes, and fraternal twins share only 50% of their genes (although that’s a separate, wrong, assumption), that you can identify how much more traits are affected by genes from those traits that are strongest in identical twins and weaker in fraternal twins. We’ll call that A. The common environment, the family home in which all these sets of twins are reared, and assumed to be treated identically, is called C. And finally, E is the stuff that varies that the twins don’t share–different schools, for example.
It’s a nice little model, and creates this pretty little equation, which then elegantly lets you calculate all the various parts that each contribute, and we can call that heritability. One introductory calculation is that for the heritability of height. The calculation gives you a fraction, 8/10. As elementary school children know, you can convert that into 0.8, and then a good student would want to convert it to 80% and say that 80% of the variance in height in the studied population of twins reared together is attributable to genetic differences.
Errors everywhere, as far as the eye can see
80% attributable to genetic differences would be wrong. Heritability is about correlation, not causation. Just because these A, C, and E all seem to go together doesn’t mean that one of them must cause the other. A classical example is phenylketonuria. Just having a gene for it doesn’t mean that you’re going to die of it, because we can change the environment. Furthermore, it’s about a group because that’s how you get that best fit line, right? You have a bunch of plants, or people. You are talking about how stuff varies within a population, not about how populations vary. “An individual does not have a variance.”
Furthermore, racists, whom deBoers disavows, extrapolate that differences between individuals must be the explanation for the differences between groups. Lewontin’s classic example is that of two crops of corn. One of them is grown in a greenhouse, perfectly controlled with all the perfect amount of light, water, etc. Therefore, the variation among the plants is entirely due to genetic differences, and heritability is 1. Another greenhouse got a bad supervisor, and the same type of corn was also grown in perfectly controlled conditions–but those conditions were bad. The heritability here is also 1. As you can see, the perfect heritability of the bad corn crop cannot be extrapolated to all the corn, in general, because environmental factors are to blame.
If this is confusing, here are a couple of elegant examples. First, I borrowed from Khan. If you use a twin study from a country like the USA, we might note that height is 80% heritable. In a poorer country, the uneven quality of nutrition might mean that height is 60% heritable. A richer country can afford to feed all of its students better, so that the “floor” of nutrition is higher–there is less variability in the environment, which gives us greater heritability. Here’s another way to put it: “the greater environmental inputs result in greater heritability!” This is not at all how most people think about 80% heritability.
Another example is that of the height of rural women from the Upper Midwest. They’re of good German stock out there, well fed, and the environment of the Great Plains is pretty similar, and the heritability of the height is .75. Because the environment is the same, the dissimilar genes have a greater effect. Now, we run that same analysis in New York City, with all its wide varieties of environments and peoples, and the heritability of the height is .40. Therefore, “there is more variation in the environmental factors that affect height than there is in the genes that affect height. In other words, the women of New York City have a bit more in common genetically than environmentally.”
statistics (deBoer’s assumptions about .8 heritability of IQ)
Now that I’ve dragged you through all of this, let’s talk about deBoer’s argument about IQ. He says that “the heritability of IQ increases over the length of a child’s life until it reaches its peak, as high as 0.80 perhaps, in the late teens or early twenties.” Most people, including me at one point in life, would read that and think that your IQ is fixed by your parents’ genes. But we know better, now. What we’re reading is that our environment changes within a group less and less as we get older, and so the variance among IQ gets smaller. This is a call to arms for raising the floor of how we treat our children, not an argument that we’re helpless to do anything about unintelligent people.
One argument to support that point is the well-known, uncontroversial fact that eldest children and singletons tend to have higher IQs than middle and later children. This is called the birth order effect, and it’s long been attributed to the resources that parents can bring to bear in child rearing. deBoer doesn’t address this at all.
deBoer spends a good bit of time quoting Turkheimer’s “Three Laws of Behavioral Genetics” to argue that family effects are less important than genetics, and “the portion of variation attributable to the shared environment is, for most traits, very close to zero.” You know what else is close to zero? The heritability of IQ in Genome-Wide Association Studies (GWAS).
As Turkheimer puts it:
Something, and presumably something that can be broadly characterized as environmental, makes siblings, even identical twins, different from each other. But whatever that something is, attempts to decompose it into an additive collection of systematic environmental causes that produce systematic differences in outcome almost always end in disappointment…. the molecular genetic project has foundered on the same shoals of developmental complexity that sank the non-shared environment.
For example, in a recent special issue of Nature Genetics, conceding the difficulty of finding genes ‘for’ depression or intelligence or schizophrenia, researchers turned to something that really ought to work: height. Height has a heritability upwards of 0.9; it can be measured almost without error; it has straightforward analogues in lower organisms; and it can be acquired from tens of thousands of participants without great effort. And when all the 65,000 participants were pooled together, half a million SNPs each, the specific associations between SNPs and height accounted for about 3% of the variance.
Complex human behaviour emerges out of a hyper-complex developmental network into which individual genes and individual environmental events are inputs. The systematic causal effects of any of those inputs are lost in the developmental complexity of the network.
(Variation and Causation in the Environment and Genome)
As one lab puts it:
Heritability is not fate. Heritability is not immutable. Heritability does not measure our ability to affect the trait. High heritability does not mean group differences are genetic.
deBoer’s paean to the immutability of IQ does not fly with me. If we absolutely must do the eugenicists’ work for them and use IQ as a synonym for intelligence, then we can at least admit that IQ is a statistical artifact based on the eugenicists’ admiration for the upper-class, straight, white British man and note that closing the “achievement gap” is really about enculturating our children to be more like that standard, and that the majority of that work is done by women.
IQ is not fate, because we don’t know what the distribution of IQ in the general population looks like. For a group of people whose IQ has been tested and found to be high, higher educational achievement is correlated—but correlation is not causation! Even to get that far, we have to test people’s IQ, and that is relatively uncommon. Just getting an IQ test in our society is positively correlated with socio-economic status. Why do I say that? Because when Broward County Schools decided to test all children, “the screening program led to a 180 percent increase in the gifted rate among all disadvantaged students, with a 130 percent increase for Hispanic students and an 80 percent increase for black students.” Other research is consistent with this, so much so that gifted programs are often rightfully thought to be SES sorting programs.
Again, deBoer knows this, when he says, “For much of our history, girls and women were casually assumed to be inherently less intelligent than boys and men” (pg 113), but he fails to follow through to the logical conclusion. Public opinion of women’s intelligence has changed since the 1940s, but the racist opinions about intelligence have not. deBoer eventually concludes that the academic success of women “is an example of structural societal changes producing changes in the classroom, not the other way around.” (pg 114) Given that, I am at a loss as to his conclusion that educational achievement is primarily genetically based.
IQ is not immutable. IQ changes depending on how you measure it–remember the autistic IQ with RPM vs the WISC? IQ changes over time, as per the Flynn effect. IQ changes as we age, particularly for teenagers. Multiple studies support the idea that background knowledge positively affects reading comprehension, and therefore, more background knowledge will positively affect IQ. “Comparisons between schooled and unschooled groups reveal a strong effect of education on intelligence test scores even on nonverbal tests. Only by systematic education can individuals’ intelligence emerge and approach an optimum.”
Heritability does not measure our ability to affect IQ. I am not unsympathetic to deBoer’s thesis that since intelligence is inherited, and since we cannot choose our parents, giving accolades and monetary rewards to the more intelligent is akin to a lottery. I differ in his casual assumption that intelligence leads directly to educational achievement.
It’s worth taking some time to discuss this. In Didau’s Making Kids Cleverer, he reminds us “When we compare the IQ of children of similar socio-economic status, most of the variance is explained by genes; but when the IQ scores of children of lower socio-economic status are compared, most of the variation is explained by environmental differences.” This is important, because deBoer repeatedly argues that most kids are doing fine, because “parenting and family environment matter very little.” (pg 128). He seems to base this primarily on Judith Harris’s The Nurture Assumption. However, Didau points out
“…when the IQ scores of children of lower socio-economic status are compared, most of the variation is explained by environmental differences. Heritability can be as high as 70% for middle-class children and as low as 20% for less advantaged children.”
The kids are not all right. 30% of US infants are beaten regularly. Almost half of states permit children to be beaten in school. Almost 15% of young children go hungry. Fully 1/3 of US infants don’t get their diapers changed regularly because, as Eminem so memorably put it, “food stamps don’t buy diapers.” “More than 1 in 6 children under 6 were poor ($494 a week for a family of four) and almost half of them lived in extreme poverty.” In my own state, my senator’s refusal to extend the Child Income Tax Credit returned enough children to poverty to fill the 50,000 seats at the WVU Coliseum. 2/3s of US ten year olds are not proficient readers. 3 out of 5 of US ten year olds are not proficient at math. Schools matter for education, not just for safe warehousing.
IQ does not mean group differences are genetic. One of deBoer’s first salvos is repeated throughout the book “…educational achievement is significantly heritable—that is, it passes from parent to child genetically, with biological parentage accounting for half or more of the variation in a given outcome.” (8)
As discussed earlier, this is simply not true, not in the sense that he means. This is akin to saying that medieval peasant educational achievement is significantly inherited, when we are all quite aware that access and opportunity matter. For example, the heritability of intergenerational wealth in the UK is 0.75, approximately the same as IQ. Does that mean that genes cause wealth? Of course not. That way lies aristocracy, and we fought several wars to avoid one of those. We shouldn’t make the same mistake with the eugenicist standards of intelligence tests.
To make this logic gap worse, deBoer knows this isn’t quite true, as he shows when he discusses “group differences” in IQ (as if IQ is linked to gender, a gender essentialist position if I’ve ever read one) when he says, “In 1970, men earned almost 60 percent of all degrees conferred by colleges and universities; by 2015, they earned less than 40 percent.” The Flynn effect is interesting, but I don’t think anyone is claiming that approximately half the population changed their genes and thus their IQ relative to the other half so significantly that women are now smarter than men are.
The importance of education
I have many problems with this book, but where deBoer is most strong, is in his argument that after K-12 education, we don’t really have any support for the lowest achieving (deBoer would say “untalented”) students. He makes this point most eloquently in the introduction. His basic argument is that the hierarchical nature of our educational system (a microcosm of our society) requires that there be losers, and those educational losers are needlessly suffering. Of course, the state of Pennsylvania would argue that’s fine, “What use would someone on the McDonald’s career track have for Algebra 1?”
That said, deBoer suffers from Ivory Tower privilege, and this narrows his understanding considerably. “We use academic performance as shorthand for a person’s overall human value.” (pg 5) “We” is doing a lot of work there. Perhaps in the rarefied air in which deBoer seems to have spent the bulk of his formative and adult years, folks care about where you went to college, but down here in the muck of the least educated, poorest, sickest state in the union, most folks absolutely do not care if, or where, you went to college. In my experience, they’re much more interested in where you go to church.
School performance is not ability. deBoer should know that, as he says “I was focused primarily on the measurement of student learning and I took classes in statistics, research methods, educational measurement, and psychometrics.” (pg 16). Kathleen Humble makes this point about the gap between intelligence and performance most eloquently in her brilliant article, “Gifted versus Gifted.” She says,
“When educators are talking gifted – particularly in journals, they generally define gifted as children who achieve in the top 10% of school or school assessments…. I will call this group E-gifted. (Short for Educationally gifted)…
P-gifted (Short for Psychologically gifted)… will have an IQ score in the top 2.1% of the population (2 standard deviations from the norm or average score. For the WISC, this would be a score of 130+). For children with a diagnosed disability affecting their communication or motor skills, they will instead have some sub-scores in the top ~2% and an average score (if it can be calculated) at least 1 standard deviation above average (for the WISC, this would be 115+). … They are not all high-achieving. They are present in all populations and at all socio-economic levels. If they are from a minority group or a poor background, they are unlikely to be identified at all.…For P-gifted children:
- a higher-than-average risk of dropping out of school,
- a higher than average risk of mental health problems,
- a risk of both misdiagnosis and missed diagnosis with possible co-occurring disabilities,
- and potentially a higher than average chance of going to prison.
…both – E-gifted and P-gifted…[may] have the high-IQ score; they are also high-achieving. This is the group every single longitudinal study of giftedness has focused on.
…And then there’s the 4th group – 2e. Twice-exceptional (or 2e) children are a sub-set of the P-gifted group….A small amount of these children may end up in the purple group – but most of them will be missed and not even identified as P-gifted. They will almost certainly NOT get the accommodations needed in school and are drastically more likely to drop out of school or be homeschooled.
In short, evidence of education is not evidence of IQ. There are plenty of highly intelligent folks out there who do not have college degrees.
We can and should support all individual students. Unlike most educators, deBoer repeatedly throws up his hands at attempting to assist students with the “great grinding race up the academic ladder” referencing the “creative class” and the “aspirational class” as possessing “cultural and intellectual capital” and replicating themselves through “assortative mating.” (pages 33-37). Apparently, we’re just supposed to sit there and take the “Dream Hoarders” as an immovable force. My mother is calling from the 1970s, and she would like a word.
This theme of biology as destiny is repeated from the beginning, “in general, the kids who had distinguished themselves in second grade were to the same applying to colleges in twelfth” (pg 22) to the end, “equality of opportunity is a shibboleth” (pg 163). Yet, deBoer excuses himself with this offhand comment: “we cannot now come even close to saying whether a specific individual will be naturally academically talented.” (pg 81) Well, which is it? Is biology destiny or are we to provide a free, appropriate public education to all? As a certified special educator, I know which side I’m on—we take students from where they are to as far as they can go, whether the student likes it or not.
Education is not only about the top students. I find myself absolutely boggled by his sheer willingness to leave our students behind:
“In fact, the notion that there is a strong connection between education and economic growth has recently been convincingly argued to be largely a statistical mirage. The data shows that what really matters is the academic performance of the top 5 percent of students. This makes intuitive sense; even in a very well-educated country, a relatively small number of people are doing the lion’s share of the intellectual work.”
This is, frankly, wrong on many levels, and he should be embarrassed to have written it. First, deBoer counts himself a friend of teacher unions (pg 39-40), and he decries those who criticize teachers—and here he is, doing it himself. Teaching is, at its heart, an intellectually demanding occupation, and according to the BLS, 6% of the US workforce is employed in education. But apparently, that doesn’t matter, because they’re not the top 5% of students? (It does matter—teacher academic aptitude matters a great deal in student reading and math achievement.)
This is also true in the obverse. There is some evidence that autistic people as a group have “high, but more or less imbalanced, components of intelligence.” Given the prodigious increase in IQ on the RPM versus the WISC, and the correlation between income and IQ, you’d think that autistic people would be wealthier. Instead, 85% of autistic people are unemployed. High IQ does not automatically lead to higher education and thus higher wealth. Correlation does not indicate causation.
In addition, I would borrow David Didau’s argument in Making Kids Cleverer, that “Intelligence is a social good. The greater the number of individuals with higher intelligence, the safer, happier and more productive the society in which we live.” This is why taking every student from where they are to as far as they can go is good for everyone, not just the student. Our educational system may be imperfect, and we may not “catch” and support all our brightest students and we may not serve all our students well, but we’re trying to do our best for everyone.
“Intellectual work” and “academic performance” are not the same. I invite deBoer to consider the concept of executive function, for one. For another, I’ll quote Noah Smith, a well-respected writer on economics, who paraphrased Sam Hammond as saying: “skills are not arranged on a linear scale from low to high, and the people we call “low-skilled immigrants” actually have plenty of skills — just not ones that happen to command a high wage premium in the labor market.” In a personal example, I made a terrible cashier because I couldn’t handle the intellectual demands—but I’m a fantastic student.
Perhaps most damningly, the author of the article he cited on cognitive capital, or the idea that the top 5% do all the work and therefore the rest don’t matter, didn’t actually come to those conclusions. In fact, Rindermann writes:
The two most important factors are invested time in learning from early childhood on and an achievement-oriented structure in education — tests, central exams, decisions/promotion based on objectively measured achievement, discipline for improved learning, the aim of school and instruction is learning/achievement, etc. — surrounded by further factors, such as higher teacher quality and more problem-solving in instruction. Discipline increases the time spent on tasks, and relevant content tested in central exams shows positive effects on cognitive achievement.
Schools can promote equality and opportunity. On page 59 deBoer says, “Our [education] system can promote equality or it can sort people into a hierarchy of ability. It can’t do both.” In fact, K-8 schools do this all the time. Despite the fact that low-SES students typically come in one full standard deviation behind, equivalent to a full year of development, we stick them all in the same classroom, because we don’t know what students already know. We’re providing a ground floor for all students. While some students are undoubtedly sharper than others, teachers are notoriously bad at figuring how bright those students are. Creating educational policy based on the heritability of IQ is not only a bad use of data, it’s impractical at the classroom level.
deBoer argues that tracking is bad because students can’t shift from one ability band to the other. I am confused about why he thinks that is bad: after all, he argues “in general, the kids who had distinguished themselves in second grade were the same applying to colleges in twelfth” (pg 22). I would argue that we don’t actually track kids in most of the USA. In fact, you can see the New York City arrogance showing here, assuming that other school districts run like New York’s.
In most school districts, K-8 is generally untracked, with a side detour for mathematics about 6th or 7th grade. “Students are not assigned to college preparatory or vocational tracks that then dictate coursework all through high school; that practice died out in the U.S. in the late 1960s and early 1970s.” (Brookings). We do use within class ability grouping in the earlier years, and eventually sort students into different math classes in high school. (Paul Tough’s The Inequality Machine does a good job explaining why this matters). In fact, this lack of tracking is why the National Association for Gifted Children wrote a policy brief advocating for acceleration in education for gifted children, precisely because it is so rare.
There is no limit on the number of high achieving students that we can produce. Perhaps most strangely, deBoer portrays the US education system as a zero-sum game (pg 59), which it most definitely is not. By law, we provide an education to everyone (although he’s simply wrong when he says that education is a right in the USA (pg 167).) Not only is it for everyone, it’s customized to their needs. Perhaps his view of education was accurate pre-IDEA, but a “free, appropriate, public education” is the law in the United States. While we don’t always succeed at this, we work hard at it, as the approximately 3 million public school teachers in the USA would attest. Even at the post-secondary level, there is no hard limit on the number of trade school seats or undergraduate seats.
Part of what makes the US system unique is that we give everyone the opportunity to succeed. deBoer argues that opportunity is insufficient, that positive liberty is valid (pg 74), and while I don’t disagree, I’d like to point out that on page 81, he says, “…we cannot now even come close to saying whether a specific individual will be naturally academically talented.” If we can’t tell, the best we can do is let them attempt.
College degrees have real benefits. Eliminating charter schools, loosening academic standards, and arguing that college is unnecessary for a better life is very much a “rules for thee and not for me” sensibility coming from a person with a PhD who actively uses their degree in their source of employment. Again, I will turn to a newsletter from Noah Smith:
A 2016 paper by Arteaga found that a reform in Colombia which reduced the amount of coursework necessary to earn a degree resulted in a substantial drop in future wages. Indeed, the college earnings premium rises over a worker’s career, which is the opposite of what you’d expect if it were just a signal.
The whole book is a hot mess of contradictory arguing. First, deBoer argues that the relative premium of a college degree has gone down because everyone is getting one (pg 48), and then he argues that the relative value of a college degree—has gone up? He quotes Goldin and Katz as saying the relative demand for college graduates has gone up to levels not seen in 100 years (pg 53), which he attributes to creating inequality. Well, yes. That’s how it works. You pay more for something you want more, and people are more likely to do something difficult if they get rewarded for doing so. Even the mafia rewarded their college degreed employees: “mobsters have significant returns to education of 7.5–8.5% , which is only slightly smaller than their neighbors and 2–5 percentage points smaller than for U.S.-born men or male citizens.”
PreK is a public good. Chapters 8 and 9 are more of what he thinks we ought to do, since we can’t escape the fact that there will always be smarter people. Here, he says that pre-K doesn’t have any lasting benefits, but we should do it anyway. I actually read the study that he quotes from, and the good folks from Duke University would differ:
“In conclusion, the scientific rationale, the uniformly positive evidence of impact on kindergarten readiness, and the nascent body of ongoing inquiry about long-term impacts lead us to conclude that continued implementation of scaled-up pre-k programs is in order as long as the implementation is accompanied by rigorous evaluation of impact.”
Segregation is not academic tracking. deBoer shows himself to be a poor student of the history of education by conflating two categorically different things: racial segregation and academic options (pg 55). As a brief reminder, SCOTUS held:
“To separate [black children] from others of similar age and qualifications solely because of their race generates a feeling of inferiority as to their status in the community that may affect their hearts and minds in a way unlikely to ever be undone.”(Brown, 347 U.S. at 494)
We can decry isolating Black and Hispanic students at the same time that we can hope for them to attend schools that would best support their academic interests and needs because these two things are not the same.
No, you can’t just Google it. deBoer shows himself as an unserious student of how people learn by repeating the tired tripe about Googling for facts, and arguing “it’s learning processes, skills, that make education valuable.” (pg 44) I have some books for deBoer to read, but also, this is that Ivory Tower worldview again—”the students I teach don’t need to memorize facts, so no students need to memorize facts.” The world is more than that particular oyster.
Unschooling is not the answer. I’ll admit, unschooling might appear to be a good idea at first glance. Children who are naturally curious will be those children whom deBoer credits with natural ability! And “We remember things better if we discovered them for ourselves” sounds like a reasonable proposition. Did you know that the ancient Sumerians, sometime around 8,000 BCE, invented the oldest writing system in the world? Humans took nearly 5,000 years, until 3,100 BCE, to discover how that system could be used for writing language—and the system still didn’t represent sounds. We took another 1700 years, until 1,400 BCE, for Mycenaean Greeks to discover how writing could represent the sounds of language. We don’t have thousands of years for our children to re-discover reading and writing for themselves. There are 940 Sundays from birth to age 18, and every one of those weeks counts.
Perhaps deBoer’s most annoying ‘reform’ is “lowering the dropout age to 12,” or basically allowing students to unschool themselves after 6th grade. He says this is good, actually, because “The simple fact of the matter is that not everyone is meant for school, for reasons of desire as much as ability.” (pg 170) He argues his point by saying, “we have misguided policy such as forcing unwilling students to participate in schooling when the unwilling will never benefit in the first place.” (pg 173)
At this point, I’m going to take a wild guess and hypothesize that deBoer doesn’t have children of his own. Because I do. And I homeschool my children. You know what children (like all other humans) don’t want to do? Hard stuff. You know what’s hard? Learning to read. Does it benefit them to do it when they are unwilling? Absolutely. His line of argument is utter nonsense.
If deBoer’s guiding idea is, “Do your own thing!” then what he’s saying is that what children learn doesn’t really matter. There is no one thing that is more important than another thing. And if that’s true, why teach children anything at all? After all, you just said it doesn’t really matter what they learn. As my friend Dr. Campbell says, that assumption is the underlying rationale for homeschooling: “there is no hierarchy of knowledge, no information that is objectively more important than any other information. The only arbiter of that is the individual learner.”
Our children can find all kinds of things on the Internet, but I would like to remind you about the octopus study. In a study at the University of Connecticut, the vast majority of students fell for a hoax website about tree octopuses. Researchers blamed the lack of generic reading skills, but as per the baseball study we know that in fact, students didn’t have the prior background knowledge about the natural world to understand that octopuses can’t live in trees. They couldn’t push back against false information without reliable background knowledge.
I believe that what children learn matters. That we are obligated to teach them about the world outside of their experiences. And deciding what to teach is best done as a carefully constructed, scaffolded curriculum, not as a semi-random assortment of whatever a 12-year-old finds interesting. As my friend Dr. Campbell says, “There’s no guarantee that a child is going to become interested in the unifying ideas of physics or world history or the Constitution, but those things are important if you want your child to understand the world around them and be able to function as an informed citizen and competent adult.”
Again, what the 12 year old wants is irrelevant. Karen Glass says, “children are no more fit to choose their own schoolbooks than their physical diet, and that what they like is no fit guide—‘they like lollipops but cannot live upon them.’’ Along the same lines, Willingham says, “Leisure reading is wonderful for building background knowledge, but students still need a strong curriculum.” And the home culture matters, he says,
“What sort of vocabulary do parents use? Do the parents ask the children questions and listen to the children’s answers? Do they take their child to the museum or aquarium? Do they make books available to their children? Do the children observe their parents reading? … There are no shortcuts and no alternatives to trying to increase the factual knowledge that the child has not picked up at home.”(pp. 53-54 in Why Don’t Students Like School? 2nd ed).
Powerful knowledge and cultural capital should help guide our instruction. I would encourage folks to read about Pierre Bourdieu’s idea of cultural capital and Michael FD Young’s idea of powerful knowledge, particularly how class and knowledge are related. For example, deBoer’s upper middle class attitude towards formal instruction might allow people who grow up in middle class homes, speaking General American English (GAE) at home, to bypass formal English language arts instruction and get away with it, but for children who grow up speaking another dialect, as I did and my native Appalachian children do, or children who grow up speaking African American English (AAE), or children who speak English as an additional language (EAL), or who have learning disabilities around language, then this isn’t going to cut it. Dr. Seidenberg has an excellent article about the need for formal instruction and AAE in the Summer 2021 issue of American Educator.
My children need clear, explicit instruction in the rules and standards of General American English, and they need to be able to code-switch on a dime, because people who speak with their accents or write in their native manner are discriminated against. Ignoring this is a luxury good that has real consequences that I can’t afford to ignore, for my children’s sake.
The Two Cultures of Educators
In the end, I suspect that the most fundamental difference between deBoer and I is that I believe in educating all students, from where they are to where they can go. As a public school teacher of the visually impaired, I worked with quadrapelegic students with traumatic brain injury so severe that the IEP goals revolved around working towards visual attention in ten second blocks. I have no qualms about working with the lowest, slowest students because we don’t know how far they can go. I’m proud to help them on their paths, however far they make it.
On the other side of the divide are the teachers who prefer working with the students who make them look good, the bright little sponges who can tolerate poor teaching practices and soak up information almost effortlessly. Those are the students that deBoer is thinking of when he advocates for letting students drop out at age 12, because they’ll continue to do well, to build their own knowledge base, if unevenly. However, autodidacts who will succeed will do so because they come from upper-middle class families, who can afford to support them into adult success.
I’m sure teachers on both sides of the divide exist at every school, even if every educator will say they care about all students. The Cult of Smart is more relevant to one side than the other, however.
On a personal note, I am a long-time conscientious objector to eugenics based standards. Both of my children have notes in their medical files that I do not consent to intelligence testing. I have stepped so far away from the “Cult of Smart” that I homeschool my children (“weighted effect size of .4 of a standard deviation on student achievement”, pg 120) without any expectation of getting into HYP. Intelligence doesn’t make someone better or more valuable than other people. Not every child will be gifted in some way, and that’s OK. It could be said that every child contains the divine spark, the portion of the creator that resides within each human being. For that alone, our children are precious, no matter their IQ.