Reflections on Social Class and College Admissions, Part I

Socio-Economic Status, Standardized Test Scores & College Admissions

Many studies find a strong correlation between parents’ income and education level, and the child’s performance on standardized tests such as the SAT. Because of the role played by standardized tests in college admissions, especially for top tier schools, bias in test scores translates to bias in admission. And because education is held out to be the primary engine of social mobility in America, socio-economic bias in testing becomes a PROBLEM. Biased tests threaten the Republic.

Not so fast. A wee bit of counter-factual thinking will show that a teenager’s SAT scores must be correlated with their family’s income, education, and social position. Zero correlation—in America—would be the greater absurdity.

The argument is particular to America, so let’s start with a query about the people in America who make up the financially successful. Which of the following statements best describes financial success at maturity—attained affluence circa age 50—in the United States of America:

1. Affluence in America is random with respect to all personal qualities, being unrelated to intelligence, grit, hard work, and drive. It has no regular association with any personal quality or ability.

OR

2. Affluence in America has some non-zero association with personal qualities, chief among them intellectual capacity, grit, and a willingness to work hard.

To keep us all on the same page about the distribution of affluence: in 2016, above-average affluence is common among the professions—doctor, lawyer, dentist, engineer, professor; and also among managers, business people, and investors.  Below-average affluence is concentrated among those with little education, such as those who did not finish high school; those who work in occupations that do not require any degree; and those loosely attached to the labor force, who work part-time or only some of the time. Where neither of these descriptions applies, a middling level of affluence is commonly seen.

I think the first statement is absurd—in America.  But it is not per se absurd.  In some other country that followed a strict caste system, where birth status totally determined financial success, statement #1 would be the expected outcome. All healthy humans babies have equal prior probability of being smart or not, hard working or not.  Financial success always has some relationship with at least the personal qualities of energy and drive—unless a caste system interferes to constrain results by accident of birth.

The second statement better describes the USA (we can argue about how strong the association is, but have to accept that it is present). Given that, we can reason back toward the SAT results to see why the correlation with socio-economic status has to be non-zero and positive.

It is convenient to take professors and their children as the focal case (if only because I know this group well).  Take it from me: by the 21^st century, to be a tenured college professor, especially in favored subjects such as business, engineering, and law, had become a very good gig. Salaries have grown substantial; job and retirement security had always been great; and working conditions continue to be superb.  By age 50, the typical tenured college professor will earn 2X the median income (in unfavored subjects such as psychology) to 4X the median income (in more favored subjects such as business).  If he or she married an engineer, manager or other professional, family income will be 4X to 8X the median income, currently stuck at just over $50K in America.  That is affluent indeed!

Accordingly, if the child of a professor gets a high score on the SAT, that will contribute, in the aggregate, to a positive correlation between family socio-economic status and child performance on standardized tests and other measures of academic aptitude and intellectual ability.

But, and here is the point of this example, it is easy to imagine other reasons why the child of a college professor might do well on the SAT.  The professor parent had to pass through the eye of a needle to obtain his or her position.  This person had to: 1) do well enough in high school to be admitted to a decent college; 2) do well enough in that college to be admitted to a decent or better Ph.D. program; 3) be smart enough to handle graduate level course work; 4) have enough grit to complete the dissertation (historically, fewer than half of Ph.D. candidates complete that task); 5) distinguish themselves among other Ph.D.’s sufficiently to get hired onto the tenure track; and 6) distinguish themselves among other assistant professors, by their success at the challenging intellectual task of conducting original scholarship and getting it published.

We can state conclusively that the tenured professor is a man or woman with high intellectual ability, notable grit, and a considerable capacity for hard work. (Although, of course, he might later go to seed.)

I hasten to add that there’s nothing special about professors. The exact same passage through the eye of the needle applies to doctors, lawyers, engineers and other successful professionals.

* Important: the association between ability and affluence is asymmetrical.  We can know that the tenured professor has ability; but we cannot know that this other professor who was denied tenure, or this adjunct professor, or this Ph.D. student who didn’t finish the degree, or this college graduate who didn’t get admitted to a Ph.D. program, is relatively lacking in ability. Life is chancy. Success indicates ability; but ability does not always issue in success.

** Also important: the small business person may become very financially successful without a college degree or any evidence of above-average ability, and not even by grit or other person quality.  Showing up every day, plus good fortune, or maybe family connections, can be enough. Hence, my argument applies best to professions requiring education. Nonetheless, I’d argue that large numbers of the affluent in America are educated professionals, and that most successful business people at least score above average on personal qualities such as pluck, grit and drive.  In which case, my argument stands.

With the association between affluence and parental ability established, here is a second differential. Which of the following statements do you believe to be correct:

3. Intellectual abilities, and personal qualities such as grit, drive, and pluck, show zero inheritability. Parents with high levels of grit are equally likely to have children with low, medium or high levels of grit; very smart parents are equally likely to have slow, average, or smart kids.

OR

4. There is an imperfect but significantly positive association between parental grit and child grit, parental smarts and child smarts, etc.

As written, statement #3 verges on the absurd; but a softened version could be made plausible.  We would have to hypothesize an extreme form of regression to the mean, and, to nail it down, add a bit of a boomerang effect. If high levels of parental drive tended to produce only moderate levels of drive in the child (i.e., regression to the mean), and also a slight reaction formation (the child tends toward the lazy and shiftless, to differentiate himself from the father; or the lazy, shiftless parent produces a somewhat driven child, again as a statement of difference); then something like statement #3 would hold true.

But we do need a boomerang or reaction formation: if high grit parents produce children of average to high grit (a moderate regression to the mean), and easily discouraged parents produce children of low to average grit, then statement #4 will hold true.  That’s why I picked drive as my example quality—we all know the folklore in which hard-working, hard-driving parents produce lazy slacker kids.

It’s much harder to believe in a boomerang effect in the case of high (low) intellectual ability: really, do you think the children of lawyers tend toward halting speech and mediocre reading ability?  That the children of engineers tend toward sub-par performance in math class?

If not, then statement #4 must be true, just as statement #2 had to be true. And now we can pull together the threads of this post.

1. Among affluent parents, we will find over-represented individuals who were good test takers, and who showed grit and drive;
2. The children of these affluent parents will likewise inherit high levels of whatever those tests measure, and high levels of grit and drive;
3. Therefore, the children of parents who are above-average in affluence will tend to perform above the average on the SAT—even if the test is not the slightest bit biased in terms of content and style, but does what it is supposed to, which is measure intellectual aptitude, as reinforced or not by years of diligent school work.

Since the SAT and most other paper and pencil (today, mouse-click) tests are probably at least somewhat biased in terms of cognitive style, favoring individuals who are accustomed to sitting still in quiet environments and maintaining an uninterrupted focus for hours at a time, the end result should be a strong association between socio-economic status and SAT performance, even when that test’s bias is minimal, relative to what it purports to measure.

Preliminary Conclusion

If personal qualities have some heritability, and if success in America has some association with ability and grit, then the children of college professors are highly likely to do better than average on the SAT, and run the gauntlet of college admissions more successfully. The University of California at Berkeley, and every other flagship public university, will be full of the children of professors, and engineers, and lawyers, and doctors, and executives.

Not because their parents are affluent, or privileged, but because these kids had a favorable inheritance and grew up within a relatively rational system for allocating financial success, wherein their parents had succeeded by merit.

If this conclusion still sticks in your craw, how do you feel about these next two statements:

5. Americans who work hard, and who persevere against obstacles, should enjoy no greater financial success, relative to Americans who prefer to take it easy, and who seldom persevere when the going gets tough.
6. Americans who are fidgety indoors, and prefer to be in motion outdoors, and who hate to sit in front of a computer, and who would much prefer chatting with friends to reading or writing, and who would rather work up a sweat than do mathematics, should expect the same financial rewards and social position as their opposite: people who love to read, write and do math, and who can sit inside for hours at a desk doing reading, writing, and arithmetic.

The first statement is calculated to enrage large numbers of Americans.  That is radical egalitarianism taken a step too far. But if grit is a heritable characteristic, then #5, stated as a factual claim rather than an ideal, must be false.  People who have a favorable inheritance, in this case of the personal quality of grit, will be more financially successful than average. In a word, they will enjoy social position achieved partly through inheritance. That statement makes Americans uncomfortable too. Hence this essay: unspoken conflicts, such as hard work should be rewarded, capacity for hard work is heritable, and inherited social position is bad, provide an opportunity for insight.

Statement #6, of course, describes a large number of teenagers.  It’s the stereotype of the high school athlete who struggles in school to get B’s and C’s, not because he or she lacks ability, pluck or grit, but because of their preferred mode of living in the world, which is body-focused and active, rather than sedentary and cerebral.

Yet, in 21^st century America, what are the odds of professional success, if you don’t have a proclivity for sitting still in front of a computer, focused on reading, writing, and doing math? That proclivity is entirely cultural and temperament-based, of course, and has little to do with whatever neural substrate underlies what we refer to as “native intelligence.”  But, speaking in terms of population averages, who would be more likely to show that proclivity: the professor’s son and the engineer’s daughter? Or the carpenter’s son and the laundress’ daughter?

Cultural proclivities are not unrelated to socio-economic position; which is not unrelated to personal qualities; which merge imperceptibly into cultural proclivities.

To highlight the cultural element: in medieval France around the first millenium, that physically active individual in #6 was far more likely to enjoy financial success than his antithesis—he might succeed as a knight, and rule a seigneurie, while his antithesis could only hope to be a celibate monk under a vow of poverty.

Shall we condemn 21^st century America for being a place where the sedentary and cerebral have excellent career prospects, and the bodily-focused and physically active have dimmer prospects?

Loaded question: should colleges seek diversity by including the cerebral and the non-cerebral, inclusive excellence as it is sometimes called?

 

Next: culture and success on tests and in school

See also: the fantasy (and nightmare) of total social mobility