In his book, Caplan goes on at great length about what level of academic achievement predicts probability of success in attaining a credential. But he complete ignores the fact that the probability of low academic achievement is skewed based on demographic attributes. While it’s fashionable these days to pretend that income is the great demographic divider, the mother of all disaggregations in America is race.
Let’s examine Caplan’s discussion of race in educational achievement. Go get your copy of Case Against Education and check the index. I’ll wait.
What about Hispanics? No one has the last name “Hispanic”, or “Mexican” or “Puerto Rico”, much less “Dominican” or “Salvadoran”, so the sum total of their mention is uno.
And mind you, I mean mentions. At no point does Caplan do anything so basic as discuss the academic performance of different demographic categories. Blacks and Hispanics make a brief appearance in name only during the Griggs discussion and never show up again.
How do you write a book that argues for draconian cuts in our education system—and not discuss race?
Education policy in America is obsessed with race. Name a single problem in education and it’s a mortal lock that it was either caused by the achievement gap or caused by a policy put in place to end the achievement gap. Any attempts to solve educational challenges will be sued out of existence, or fail, or simply ignored to death because of its impact on the racial achievement gap.
But Caplan never once explores whether the implications of his proposals might unduly affect certain demographics. He simply uses median scores and percentages for the overall population. I am not a huge fan of Tyler Cowen’s dystopic fantasies but in education, there’s no doubt that average is over and has been for years. Averages hide too much. In Caplan’s book, averages hide the implications of his “ability archetypes”:
Caplan advises people to use “ability archetypes” to ensure they are realistic about their goals:
Let’s consider the racial implications of his advice. Once again, we’ll use the NAAL report that Caplan discarded after culling a few shallow data points.
Here’s the results broken down by race in the four ability categories, from Below Basic to Proficient, for Prose and Document. For example, white comprise 70% of the population and 7% of the tested white population scored below basic in the Prose category. So 4.9% of the tested population was white and below basic in Prose. White scores are in gray, black scores in blue.
(I’ve been working on this forever, and just now noticed I didn’t put the percentage of each race’s contribution to each category. Sigh.)
Asian and Hispanic results are skewed by the conflation of immigrant and native results. But it’s instantly obvious that blacks, who were only 12% of the tested population, contribute far more to the lower categories and are almost non-existent in the skill categories Caplan considers suitable for college.
The columns in the graph below list the median score by race in each education category. The horizontal lines are the overall population percentiles. So 14% got Below Basic, while “Basic” scores went from the 14th to 44th, Intermediate from 44th to 85th, and Proficient above that. The “Excellent”, “Good”, “Fair”, and “Poor” classifications are those that Caplan defined and are at (very roughly) the corresponding percentile location. (“Good” is a bit low, I think.)
Notice that white high schoolers and high school graduates have roughly the same scores as blacks with 4 year degrees or more. This is a very consistent finding in most test score data.
Caplan argues that only students from the Excellent or Good categories should invest in college. The NAAL report finds that only two percent of blacks read at proficient levels, 31% score at the intermediate level. If blacks or colleges took Caplan’s directive and only went to college with that qualification (which is actually broader than Caplan would like) just 4% of the overall population would be black college graduates.
NAAL doesn’t disaggregate by race, education, and performance category. But another survey, done three years later, gives us some insight: The Literacy of America’s College Students. This literacy survey tested 25 randomly selected students from each of over 1800 universities.
This survey uses the same assessment as NAAL, and the same categories, to assess college students in their last semester of an AA or BA degree. Again, I’m restricting the comparison to blacks and whites.
First, I benchmarked the literacy data to the NAAL data for college graduates.
The literacy survey data is much higher for blacks than the NAAL data, particularly for black AA holders. But it’s pretty close for BA holders. Moreover, standards change over time so it’s at least possible that looking at brand new AA degrees would differ from the overall population.
Here’s the breakdown by score category. Black AA and BA candidates are on the left, whites on the right. Blue and green are intermediate and proficient categories.
And consistent with the first graph, these results seem quite high for African Americans. Only 5% of blacks in 4 year schools scored below basic? Blacks in 2 year colleges had no below-basic scorers? Really?
Still, this is fine for my purposes. 1 in 4 blacks about to get a BA had basic or lower reading scores, while less than 1 in 40 whites had the same low ability.
Caplan asserts “we” should be shocked that “under a third” of those with a BA or higher achieve Proficient levels in numeracy and literacy. But close to half of the white college BA holders achieved Proficient levels in the three categories ( 42%, 45%, and 40%). The same black proficiency scores are 16%, 17%, and 5%.
Whites are achieving considerably higher than the results Caplan sniffs at, while black scores are far worse than “under a third” but rather “under a fifth”. Moreover, Caplan argues that he’s giving this advice to prevent low-skilled people from failing in college–but clearly, these blacks are about to graduate and made it through with skills he deems too low to succeed.
The college graduate data above would almost certainly be replicated in all the other education categories. Whatever Americans Caplan decries as low-skilled and incapable of succeeding in education, rest assured that he’s skewering a group that’s considerably more African American than the overall population.
But NAAL results over time (below) suggest that our k-12 system has improved results for African Americans. Asterisked scores indicate significant improvement. Blacks saw significant improvement in all three areas. (note again Hispanic performance declined rather spectacularly, thanks to increased immigration)
What educational categories saw the most black improvement?
Well, hey now. Look at that. The blacks that graduated (or even dropped out!) of high school in the 10 years previous saw significant improvement in prose and quantitative skills.
Black proficiency scores on the NAAL survey are extremely low. But they have improved.
Caplan’s prescriptions run into all sorts of problems when evaluating black academic performance. If Caplan is correct about the skills needed for college, then why is the black college graduate average below the level that Caplan declares essential for college success? Certainly, as I’ve observed, colleges are lowering standards (for all admissions as well as blacks in particular). But while the average earnings of black college graduates are less than those of whites, black earnings increase with education nonetheless. So should they invest in more education even though they don’t meet Caplan’s criteria?
I pointed this out to Caplan on Twitter, and he observed that the ethnic group improvements were marginal and that the absolute level of basic skills were “terrible”. Which suggests he was aware of the ethnic group differences and just decided not to mention them.
Breaking down test scores by race can be incredibly depressing. No one likes to do it. But Caplan’s failure to include this information is simply irresponsible.
Caplan argues that people outside the top 30% of academic achievement should stop investing in school, the sooner the better. He sees this as both selfishly correct and also the correct government policy, so he thinks all funding for education past minimal skills should end. Those who are worth further investment can justify the expense to a bank or a parent. Meanwhile, we should end the child labor laws so that the very lowest academic achievers can get to work as soon as it becomes a waste of time to educate them.
Applying his policies to black Americans, around 25 percent would be in need of those changed labor laws, because Caplan wouldn’t spend a penny to educate them.
In his conversation with Hanushek, Caplan proposes giving low-skilled kids “more realistic” careers–the example being “plumber”, of course. Like most elites, Caplan uses “plumber” as a low-skilled proxy when in fact the occupation is one of the more cognitively complex of blue collar jobs. But I think his focus on the job is also a tactical choice. “Plumber” sounds good, like a meaningful career. You can be self-employed or build a business.
Imagine telling a kid his best option is “janitor”. Now imagine telling a poor black kid his best option is “janitor”. Then imagine telling about 1 in 4 black kids that yeah, “janitors” where it’s at for them.
If you can’t imagine doing that, then don’t write a book arguing that Americans get too much education.
When people talk about the “bad old days” of American education, they are referring to the era when people did exactly what Caplan advises. School counselors looked at the students’ test scores and gave them a list of possible careers. White kids had higher scores and were advised to go to college. Black kids had lower scores and were advised to go to factories or custodial work. For a guy who spent several pages on the likelihood of Griggs lawsuits, Caplan doesn’t seem to have spent a single second looking at the case history of school district consent decrees.
But then again the kicker: Caplan wants open borders. So in Caplan’s ideal future, all those teenagers of all races that have been kicked out of school because they aren’t worth educating will be competing for jobs and housing with millions or more adults from third world countries.
Earlier, I wrote:
I’ve been struggling with the best way to take on Bryan Caplan’s woefully simplistic argument about the uselessness of education. What do you do when someone with a much bigger megaphone takes up a position similar to one you hold–but does it with lousy data and specious reasoning, promoting the utterly wrong approach in seeming ignorance about the consequences?
Nowhere is this dilemma clearer than in Caplan’s utter refusal to engage with the racial implications of his proposals. I, too, want fewer people in college. The best way to keep unqualified people from investing in college is to make work worthwhile. But Caplan wants to devalue work to the point of worthlessness through open borders, all the while denying even the possibility of education to those who can’t afford it.
Caplan complains that no proponents of public education have seriously engaged with his book. That’s because no one has observed, in so many words, “Bryan Caplan thinks most blacks shouldn’t go to college because they’ll fail. He thinks state funded education is a waste of time. Kids whose parents can’t afford education should have to be smart enough to get a scholarship.”
That’ll get him some engagement. But then, he knows that.
Caplan is often rather smug about his media popularity. “Steve Sailer’s policy views are much closer to the typical American’s than mine. Compared to me, he’s virtually normal. But the mainstream media is very sweet to me, and treats Steve like a pariah. I have to admit, it’s bizarre.”
It’s not bizarre at all. Honesty usually goes unrewarded.