What We Talk About When We Talk About “IQ:” Algebra and School Quality

The now-unemployed Jason Richwine is portraying himself as a numbers-driven policy wonk who has been unfairly pilloried for a nuanced, intellectually sophisticated Harvard dissertation, whose only crime was questioning liberal pieties on race and IQ. Byron York:

Richwine and others also pointed to the fact that his ideas were expressed most completely in a dissertation done at Harvard, of all places, under the supervision of a group of distinguished scholars, and that the dissertation was accepted and Richwine was awarded a Ph.D. It seems unlikely that a Harvard dissertation, finished in 2009, would qualify as hate speech, his defenders contend. But that is how it was portrayed in the controversy.

Over the past several days, I dove more deeply into Richwine's dissertation arguing that Hispanics are innately less intelligent than whites, and thus should not be granted citizenship. Let me acknowledge at the outset that I disagree profoundly with Richwine's conclusion. I find it inhumane to argue that political rights be conditioned on a test score. Richwine and the Heritage Foundation also downplay the reality that many important jobs in our economy–picking tomatoes, delivering food, cleaning buildings–require little formal education or demonstrated intellectual ability, and that native-born Americans will not do them

But what I want to address here, at greater length, is the attention this controversy has put on IQ testing as a means of judging innate intelligence. What do measurements of IQ actually consist of? Does Richwine's analytic work stand up to scrutiny? He argues that education can do very little to help Hispanic immigrants achieve. So does he demonstrate familiarity with the educational and economic research on poverty, schools, and human capacity? 

Richwine's dissertation may not be hate speech, but I emerged from it surprised that this document garnered a Ph.D from the nation's preeminent university. Richwine fails to grasp the difference between testing academic achievement and testing innate cognitive ability, claiming that an exam that includes algebra can be used to draw conclusions about inherited IQ. He explicitly ignores the well-documented, historically persistent reality of educational inequality across the United States, assuming that the only "environmental" factors that affect a child's test score are ones inside the home.

In Chapter 2 of the dissertation, Richwine acknowledges that the "language bias" in most American IQ tests makes it difficult to assess the intelligence of native Spanish speakers or those who grow up in Spanish-dominant homes. He therefore argues that gaps between Hispanic and white performance in math provide the strongest evidence of innate ability differences between the two groups. He draws many of his conclusions from the 1979 National Longitudinal Survey of Youth, which collected IQ scores from nearly 12,000 individuals who took the Armed Forces Qualification Test.

Is the AFQT an accurate measure of genetic cognitive ability, as Richwine claims — as distinct from academic achievement?

As Richwine writes, the AFQT "was designed for 17 and 18-year olds who speak English and have taken algebra." Some students in the data set may not yet have enrolled in algebra when they took the exam, so Richwine adjusts for school-entry cutoff birth dates. This ensures, he writes, that his entire sample has completed the same number of years of schooling, and thus their cognitive abilities can be accurately compared to one another using this test that includes algebra.

Here I was taken aback. I don't know about you, but I was not born knowing how to solve for x. It was taught to me at school, by teachers. Was Richwine truly claiming that 12 years of schooling in Scarsdale, for example, was equal in quality to 12 years of schooling in East New York? Was he claiming that there is no significant inequality in schooling across the the United States that could help explain differences in scores on a math test? As I continued to read, this did, in fact, emerge as Richwine's argument. From pages 65-66:

Screen Shot 2013-05-13 at 2.29.50 PM
Screen Shot 2013-05-13 at 2.29.59 PM

Sure, academic standards are more uniform today than they were 100 years ago, and we give schools more funding, even in the "inner city." But to claim that school quality no longer "varies enormously" is shockingly ignorant. Affirmative action in college and private school admissions does nothing to guarantee the typical young Hispanic child access to effective preK-12 schools or teachers; in fact, in recent decades, American schools have become more segregated by race and class, with the poorest children most likely to be stuck in low-quality schools. We know these children do much better when we get them into better schools and classrooms, because we've tried it. Poor kids score higher than their racially and socioeconomically identical peers when they are enrolled in schools with middle-class students. Teachers who are good at raising their students' test scores (like, in algebra) are also good at helping them graduate high school, avoid teen pregnancy, and get higher-paying jobs — all those achievements that can supposedly be attributed to genetic IQ. Economists have been demonstrating for 15 years that somewhere between five percent and a third of the achievement gap can be attributed to poor children's lack of access to effective K-12 teachers.

So while Richwine does acknowledge that both genetic and environmental factors contribute to IQ, he locates environmental variability almost solely in the home, as if there were no inequality in contact with good teachers, orderly classrooms, up-to-date textbooks, and engaging curricula. He praises only one school-based intervention, the Abcedarian pre-school project, which he admits demonstrated "modest, tentative" IQ score gains that merit "further research." He quickly moves on, however and his overall elision of school as a factor allows him to claim that "environment" itself is a function of low IQ, with poor parents too unintelligent to provide a stimulating environment for their kids, who inherit the genetic deficit. He does not seem to know or care that such families have, through no fault of their own, unequal access to good schools that can and do raise student achievement in algebra and many other areas.

I've written extensively about how difficult it is for schools to overcome the academic affects of poverty. Yet we know good schooling does make a significant, potentially life-changing difference, and that poor children, including Hispanic immigrants and their descendents, do not have equal access to good schools. When people obsess about IQ in the face of these obvious inequalities and the vast research literature dissecting them, one has to wonder: What is the motivation? Ta-Nehisi Coates takes some guesses here. So does Jamelle Bouie.

5 thoughts on “What We Talk About When We Talk About “IQ:” Algebra and School Quality

  1. Andy

    You seem to have missed the part of your post that shows differences in IQ have a small effect on educational outcomes and that quality of schooling dominates ability in importance. When people obsess over school quality, while ignoring the vast literature on intelligence, one has to wonder: What is the motivation?

    It’s rather disappointing that you would link to Coates. His article demonstrates the worst sort of anti-intellectual attempt to shut down discussion about difficult topics.

    Reply
  2. highly_adequate

    “Economists have been demonstrating for 15 years that somewhere between five percent and a third of the achievement gap can be attributed to poor children’s lack of access to effective K-12 teachers.”

    And that is a good sign that the gap is purely environmental? At most a third of the gap can be attributed to quality of teaching, which you emphasize as being a key feature in engendering the gap, and you conclude that that shows how only environment figures into outcomes? Any half way decent scientist would conclude in any case that the true number likely lies toward the middle of the two estimates you offer in any case — only roughly 20% of the gap. How is this evidence that the gap is strictly environmental, rather than evidence for the opposite conclusion?

    Do you ever think through even your own facts, or apply any kind of sanity check to your own arguments?

    Reply
  3. highly_adequate

    A further point.

    One way, of course, to test how much the environment figures into the gap is to consider transracial adoptees. The best way would be to study black children adopted into the supposed environmental acme of stimulation, the upper middle class white family, and see how they fare on IQ tests.

    Well, such an experiment was performed, and by researchers who very much hoped that it would show that this sort of environment would eliminate the IQ gap.

    While it seemed to show promising closing of the gap earlier in childhood, by age 17 the black adoptees had fully reverted in their scores to the mean of the larger black population of the area — fully 1 SD deviation below the average for the white population of the area.

    Link:
    link to en.wikipedia.org

    Reply
  4. Min

    “many important jobs in our economy–picking tomatoes, delivering food, cleaning buildings–require little formal education or demonstrated intellectual ability, and that native-born Americans will not do them.”

    Please. That is insulting to American workers. The reality is that American employers will not pay enough money for such jobs to attract native-born Americans.

    Reply
  5. Read the literature

    “So while Richwine does acknowledge that both genetic and environmental factors contribute to IQ, he locates environmental variability almost solely in the home, as if there were no inequality in contact with good teachers, orderly classrooms, up-to-date textbooks, and engaging curricula.”

    The school district a family lives in is strongly tied to the schools the kids are exposed to. That shows up as shared environment in twin studies, and yet there is very little role for shared environment in adult IQ: family effects fade when the kid leaves home.

    Adoption studies change both family and school district, and don’t seem to show lasting effects on adult IQ.

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>