The Available Pool

(This is by far the most Voldemortean topic I’ve taken on in a while. Brace up.)

Some readers might have noted a potential flaw in my observation that ed schools can’t commit affirmative action. If the average elementary school SAT score is 500 per section, and the average content SAT score is 580 in the relevant subject, then there shouldn’t be a shortage. Plenty of African Americans have those scores, right?

Well, it depends on what you mean by “plenty”.

Just ask Malcolm Gladwell.

Four words I’d never thought I’d say. I liked Gladwell’s article about ketchup. I also find him useful as a predictive sorter: when I meet someone who admires his work, I run like hell.

But recently I came across a page I’d either missed or forgotten about since the last time I flipped through his book.

Gladwell even cites Jensen.

Conceding what he sees as a minor aspect of IQ to make a larger point, Gladwell acknowledges that regions, or thresholds, of IQ exist. But beyond these broad ability differentiators, IQ differences are irrelevant compared to factors like luck, birth, language, rice paddy history. Given certain thresholds, IQ is relatively unimportant in outcomes.

And given certain thresholds, Gladwell’s not terribly wrong, as Jensen confirms.

There’s just one pesky little problem still left to plague modern society: the thresholds. The regions, as Jensen describes them, that differentiate between broad ability levels. The ones that even an IQ pishtosher like Gladwell accepts as given. They’re kind of an issue, if by “issue” you mean the fatal flaw lurking in most of our social and education policies.

Jensen’s regions correspond to the IQ standard deviation markers. The average IQ is 100, with a standard deviation of 15. An IQ of 70 is 2 SD below the average of 50 (2nd percentile), 85 is 1 SD below average (16th percentile), 115–the marker for graduate level work, according to Gladwell and Jensen—is 1 SD above the mean.

Translating Gladwell and Jensen into standard deviations: in order for an American student to be ready for a college graduate program, he needs to have an IQ at the 84th percentile, with “average” (this is Gladwell’s word) as the 50th percentile. Give or take. IQ tests are finicky, no need to be purist. These are broad strokes.

Using those broad strokes, we know that average African American IQ is a little less than one standard deviation below that “average IQ” (again, Gladwell’s term), which means that the 84th percentile for all IQs is attained by just 2% of blacks. Test scores consistently prove out this harsh reality. While the mean African American IQ has risen five points since 1970, test performance has often remained stubbornly 1SD below that of whites. As Chistopher Jencks observes, “typical American black still scores below 75 percent of American whites on most standardized tests”, and often as much as 85% (or 1SD). Much has been written about the 1 SD difference; you can see it in the SAT, the GMAT, and the LSAT. (The SAT is much easier these days; before the recentering, just 70 blacks got over 700 on the verbal, whereas today it’s 2100, or 2%. In 1995, 90% of African Americans scored below 430 on the verbal section whereas the unrecentered LSAT has a score distribution chart registering no black scores over 170.)

(You’re thinking oh, my god, this is Bell Curve stuff. No, no. This is Gladwell, remember? Secure position in the pantheon of liberal intellectual gods. It’s all good.)

We are oversupplied with whites with IQs over the 115 threshold, all of whom have the requisite tested ability to be lawyers and doctors and professors. Since these fields are highly desirable, the educational culling process weeds out or rejects all but the most cognitive elite candidates. Thus all the cognitively demanding fields have a sorting process for whites: medicine, law, academia, science, technologists, executives, politicians, venture capitalists, mathematicians, yada yada yada all the way down to high school teachers, the peasants of the cognitive elite.

The available pool of blacks with the requisite Gladwellian-approved IQs to test into graduate education is barely toe deep.

To build cohorts with blacks exceeding single digits, graduate schools in law, medicine, and business, to name just a few, commit deep discount affirmative action, regardless of legal bans. Ed schools can’t, for reasons I described in the last post. Given the wide range of choices blacks with anything approaching the requisite cognitive ability have, it’s hard to say if any sorting occurs at all.

Much has been written of the supposedly low standards for teacher licensure exams but what do we know about the standards for becoming a lawyer in Alabama or a doctor in Missisippi?

I often ask questions for which data is unobligingly unavailable. Sometimes I just haven’t found the data, or it’s too broad to be much good. Sometimes it’s like man, I have a day job and this will have to do.

Med school: Not much data. See Razib Khan’s efforts.

Law school: For all the talk about mismatch or the concern over dismal bar exam passing rates for blacks, the reality is that low LSAT scores, law school, and persistence can still result in a licensed black lawyer. State bar exam difficulties aren’t uniform (which is also true for teaching). This bar exam predictor says that a law school graduate with an LSAT of 139, three points below the African American mean, attending an Alabama lawschool not in the top 150, graduating in the bottom tenth of his class, has a 26% chance of passing the bar. In Iowa, the same person has a 17% chance–in California, just 4%.

If that predictive application has any validity, the cognitive abilities needed to pass the average high school math or science licensure test in most states are higher than those demanded to pass a bar exam in states filling out the bottom half of the difficulty scale. Passing the math or science licensure exams with an SAT score below the African American mean would be next to impossible in most states. English and history probably compete pretty well on that front as well. It wouldn’t surprise me if the cognitive demands needed to pass elementary school licensure tests in tough states (California) are greater than those needed to pass the bar exam in easy states (Alabama). (sez me, who has passed the tests in three subjects, and sez all available information on average SAT scores for passing candidates).

Here we are back at the cognitive dissonance I mentioned in the last post. Received wisdom says teachers are stupid. Reality says teacher credential tests have significant cognitive barriers, barriers that appear to exceed those for law and may do so as well for medicine—and the other professional tests are presumably easier still.

Before I looked into this, I would have assumed that licensure tests for law and medicine weeded out a “smarter” class of blacks than those weeded out of teaching. Now I’m not as sure. It seems law schools and med schools keep out the “not-as-smart”whites and Asians while admitting blacks and Hispanics who would only be “not-as-smart” if they were white or Asian. The med and law school licensure exams, in knowledge of this weeding, are gauged to let in the “not-as-smart”, secure in the knowledge that these candidates will be mostly black and Hispanic. (A number of “not-as-smart” whites and Asians will make it through, assuming they paid a small fortune for a low-tier law school, but jobs will be much harder to find.) Understand that I’m using “smart” in the colloquial sense, which means “high test scores”. And most evidence says these are the same thing. I’ve said before now I’m not as certain of this, particularly with regards to African Americans.

This isn’t enough to prove anything, of course, and I wanted more. What else could I could use to—well, if not prove, at least not disprove, what seems to me an obvious reason for a dearth of black teachers?

Employed persons by detailed occupation, sex, race, and ethnicity

I made some predictions going in:

1. Blacks would be a higher percentage of elementary/middle school teachers than of high school teachers. I couldn’t sort out academic teachers from special ed and PE teachers, and I wasn’t sure whether sped teachers would be included in the count. But given the easier licensure test, I was betting the percentage would be higher.
2. There would be more black school administrators than black high school teachers.
3. The ratio of black lawyers and doctors to black high school teachers would be higher than the ratio of white lawyers and doctors to white high school teachers (in absolute numbers).
4. The ratio of black social workers to black teachers would be much higher than the same ratio for white teachers.

So this table shows the total employed in each category, the percentage black and white, the absolute number black and white:

This table calculates the ratio of each non-teaching occupation to K-8 and high school teachers by race. So the number of black high school teachers is 25% of the number of black K-8 teachers, and the population of black high school teachers is 65% the number of black education administrators, and so on.

I didn’t want to over-interpret the data, so this is just simple Excel, pulling the numbers right off the table (calculating white percentage by subtracting the other races). And I was right about a lot, except I underestimated the number of black professionals in the highly cognitive fields of doctors, dentists, and lawyers and I didn’t know this basic fact:

There are more white lawyers than white high school teachers!

Still, this data mostly bears out my predictions. I threw in some other categories: entertainment/media, and nursing, just for compare/contrast.

Many blacks become social workers, far more than become high school teachers or even K-8 teachers. Now, I know teachers complain about low pay, but social work has really low pay, less attractive vacations, and a client base even less cooperative than the average high school student.

I was wrong about lawyers, obviously, but less wrong on doctors. White and black doctors/dentists have roughly equal parity compared to white and black high school teachers–black doctors and dentists are about 85% of black high school teachers, whites about 87%. In med school, Whites have to compete with Asians, who are 20% of doctors (and just 5% of lawyers), but if the professions were cognitively sorting on anything approaching an equal basis, there should be a lot more black high school teachers, shouldn’t there? And if you go the less cognitively demanding but still intellectual field of nursing, black nurses outnumber black high school teachers by nearly twice the ratio that white high school teachers are outnumbered by white nurses.

So blacks are choosing skilled health care work over teaching at considerably higher rates than whites are making that same choice, and the number of black doctors/dentists have near parity with black high school teachers compared to whites in the same professions.

Then there’s my amazing perspicacity in predicting the overrepresentation of black education administrators. Pretty obvious, really. Districts can only practice affirmative action in teacher hiring to the extent they have black candidates. But administrative positions are wide open for affirmative action. While I’m sure there’s a test, it’s got to be a piece of cake compared to the high school subject credential test. I can’t really take all the credit, though.

CJ Cregg first alerted me to affirmative action in principal selection. But before you shed all sorts of tears for Tal Cregg, remember that the Brown decision resulted in thousands of black teachers and administrators losing their jobs, all in the name of racial equity and equal access.

I only had one surprise. When I started this effort, I figured that I’d include a snarky remark like “Want more black teachers? Raise the cut scores for the bar exam.” But no, lawyers, it turns out, are whiter even than high school teachers. That might explain why the cut scores are set so low on the bar exam, and it suggests that the predictive application knows its stuff. The legal profession in many states is doing its best to bring in more black and Hispanic lawyers by lowering the cut score—in others, not so much.

Steve Sailer noticed something I’d missed in my original post on teacher SAT scores—namely, teachers had strong verbal scores regardless of the subject taught. Law, too, is a field heavy on the reading and talking. So maybe whites are drawn to fields that reward this aptitude. It’s arguable, in fact, that America’s entire educational policy through the century was informed, unknowingly, by its unusually large population of unambitious smart white people who like to talk. We might want to consider that possibility before we start demanding diversity.

Anyway.

Step one in investigating the lack of black teachers should start with the oversupply of black social workers and see why, given their strong interest in community work, they aren’t going into teaching. The uninformed yutzes who presume to opine on education policy think ed schools are either prejudiced against or just uninterested in recruiting black teachers. Those actually interested in creating black teachers think it’s the licensure tests. I’m with them.

So go find out. If I’m right, we can start talking about lowering the cut scores for k-3 licensure tests. Once we realize that the Common Core goals are a chimera, we might create high school teaching tiers, with easier tests for basic math and English classes. (In exchange, maybe, for loosening up the affirmative action grip on administrative positions, if such a grip exists.)

Given the tremendous overrepresentation of blacks in our prisons, I’d argue we need to spend our time and policy creating more black lawyers, not black teachers. Better pay, better status and who knows, maybe better justice.

The available pool of black cognitive talent is small. Tradeoffs must be made. If we want more black teachers, we’ll have to lower the cognitive ability standards required for teaching or reduce the number of black professionals in better-paying, higher-status jobs. To a certain extent, the first of those options make sense. The second one’s just stupid.

I got into this because of that damn TFA announcement saying that 1 in 5 of their teaching corps was black, and the congratulatory nonsense that spewed forth in the announcement’s wake. And you still should be wondering how TFA is getting so many blacks that can pass the licensure tests. Next up, I promise.

Teacher Quality Report: Lacking a Certain Quality

October flew by. I actually did a lot of writing, but not where anyone can see it. Plus, I’ve had my butt kicked, bronchially speaking. Apparently I’m asthmatic, but the only thing that means to me is that I hang on to coughs forever. So here it is October and just two posts? 20 days between? That might be a record. Even weirder, this has been my biggest month–25K+ views. Go figure. But let’s see if I can get two done in a day.

First up is a bit of a greatest hits montage occasioned by this report on Gains in Teacher Quality by Dan Goldhaber and Joe Walch.

I noticed some, er, broad claims.

Smart Teachers are Better?

From the report: And the evidence on the importance of teacher academic proficiency generally suggests that effectiveness in raising student test scores is associated with strong cognitive skills as measured by SAT or licensure test scores, or the competitiveness of the college from which teachers graduate.

Nice to see a mention of licensure test scores, although not in the right context. But as Dan Goldhaber himself observed,

..we see that Black and other minority students appear to benefit from being matched with a Black teacher regardless of how well or poorly that teacher performed on the Praxis tests, and these positive effects due to matching with Black teachers are comparable in magnitude to having the highest-performing White teachers in the classroom. Removing the lowest of performers on the exam would necessarily remove some of the teachers that appear to be most effective for this segment of the student population.

He also found that there’s no relationship between licensure scores and reading effectiveness, but that licensure test scores operate as a reasonable screening device for white teachers and math. But not black teachers, I guess.

And RAND found less than that:

The results show large differences in teacher quality across the school district, but measured teacher characteristics explain little of the difference. Teacher licensure test scores are unrelated to teacher success in the classroom. Similarly, student achievement is unaffected by whether classroom teachers have advanced degrees. Student achievement increases with teacher experience, but the linkage is weak and largely reflects poor outcomes for teachers during their first year or two in the classroom.

Is there anything relating teacher quality to SAT scores? I couldn’t find it, but (and I mean this seriously) I may have missed it. Can You Recognize an Effective Teacher When You Recruit One? says there isn’t much of a link:

some researchers have found that teachers with stronger academic backgrounds produce larger performance gains for their children (see, for example, Clotfelter et al. (2006, 2007), in addition to the reviews cited above). However, there are also a number of studies which do not find this relationship (e.g., Harris and Sass (2006) on graduate course work and Kane et al. (2006) on college selectivity). …While several early studies failed to find a significant relationship between college admissions scores and principals’ evaluations of new teachers (e.g., Maguire (1966), Ducharme (1970)), a well cited study by Ladd and Ferguson (1996) did find a link between scores on the ACT exam and student achievement growth.

I often notice a study mentioning the “consistent” or “pervasive” link between teacher cognitive ability and student achievement, but all the research I find says that no conclusive link has been found, that some studies find a relationship, some don’t. Research is no more supportive of the smarter teachers campaign than it is for stricter gun control or tougher drug laws. Sorry.

Teachers used to be smart women, but no more

Another quote:

Over the course of the next 35 years, women still made up the vast majority of the teacher workforce, but their academic credentials began to decline. Research by Sean Corcoran, William Evans, and Robert Schwab indicates that the likelihood of a female teacher having been among the highest-scoring 10 percent of high school students on standardized achievement tests fell sharply between 1971 and 2000, from 24 to 11 percent.

Corcoran, Evans, and Schwab said that, but they said a lot more:

In the results presented here we find some evidence of a slight but detectable decline in the relative ability of the average new female teacher, when ability is measured as one’s centile rank in the distribution of high school graduates on a standardized test of verbal and mathematical aptitude. The magnitude of this decline is even greater when measuring ability using standardized scores. We also find that examination of the entire distribution of new teachers is more informative than trends in central tendency alone. Over the 1964–2000 period, women near the top of the test score distribution became much less likely to enter the teaching profession than their peers near the middle of the distribution. The apparent consequence has been a much lower representation of women of very high academic ability in the pool of elementary and secondary teachers. While the sample sizes of male teachers are much smaller, we detect the opposite trend among men.

Huh. So on average, women teachers are mostly as bright. (Consider that we have a lot more special ed teachers, who are predominantly female and who have scores as low or lower than elementary teachers.) But fewer really really smart women. Count that as a big “so what”, particularly since, as the report observes, we’re getting some more bright men.

But again, we don’t even know if we need really smart teachers.

We have no data on teacher quality

Absent persuasive evidence on the impact of efforts to raise the bar, some people have speculated that the rise of test-based accountability associated with NCLB and the ongoing push to establish more-rigorous teacher evaluation systems have made teaching less attractive and thereby contributed to further decline in the quality of the teaching corps. (emphasis mine)

Oh, come on. That’s crap. ETS has been telling the world that teacher metrics, particularly in elementary and middle school, have increased dramatically. I’ve written about this extensively, but I’ll just link in #5 on the list of heavily trafficked posts, Teacher Quality Pseudofacts, Part II, which has the most linked image on this site, and a link to this ETS report on teacher quality:

In summary, the following can be said about overall licensure patterns and academic quality during the last decade, at least for the states included in this study:

• Passing rates have decreased substantially.
• The academic profile of the entire candidate pool has improved.
• The academic profile of those passing the Praxis tests has improved.
• These improvements are consistent across gender, race/ethnicity, and licensure area.
• Profiles are markedly different for secondary subject teachers in contrast to elementary, special education, and physical education teachers.
• The decrease in passing rates is likely attributable to increasingly demanding testing requirements put in place during these intervening years.

Yeah, the licensure tests! How about those?

Things that Shouldn’t Make You Go Hmmmm

After going through the many ways in which teacher metrics have improved over a seven year period, the authors scratch their heads:

What explains the apparent rise in academic competency among new teachers? As we show, the SAT scores of those seeking and finding employment in a teaching job differ in different years. It is possible that alternative pathways into the teaching profession have become an important source of academic talent for the profession. Unfortunately, we cannot explore this issue in any depth because the way in which teachers were asked about their preparation has varied over time. Regardless, alternative routes are unlikely to be the primary explanation for the changing SAT trends given that, with a few high-profile exceptions like Teach for America, alternative certification programs are not highly selective.

Wow, introduce alternative certification and before I have a chance to get huffy, walk it back. Damn skippy alternative certification programs are not highly selective. Many are set up specifically to recruit URM teachers, and in many ways “alternative certification”, outside of TFA, is a proxy for black and Hispanic teachers.

The 1998 Higher Education Act required ed schools to prove that 80% of their candidates passed all licensure tests. If I understand the politics correctly, the law was intended to force ed schools to spend more time covering content, which is absurd. Ed schools responded predictably—who the hell wants to teach 6th grade math in college?—requiring candidates pass at least one licensure test to be eligible for admission. A couple years later, NCLB’s “highly qualified teacher” criteria led to a wholesale increase in elementary and middle school content knowledge requirements, reflected in much tougher licensure tests.

So the 1998 Higher Education Act, coupled with the tougher licensure tests of the NCLB era, led to a tremendous decline in black and Hispanic teachers overall, and their virtual disappearance from education schools. Alternative certification programs sprung up as a way to bring in more URM teachers—they aren’t bound by the 1998 law, so they can bring in candidates and then spend most of the training time teaching them the content to pass the test.

Have you been paying attention? Because it’s pretty friggin’ obvious why the “academic competency” of teachers has improved. I’ve been writing about this forever. THE LICENSURE TESTS ARE HARDER. I shouted this back in March (It’s the test, Zitbrains!) and at least twice on the Clarence Mumford case, and I don’t know how to holler it any louder.

This authors mention the licensure tests twice, but never for the right reason. They never once consider whether the tests might be a source of the increase. In fact, they never seem to realize that their report is largely redundant, since ETS covered the same ground six years ago.

Meanwhile, despite this big boost in teacher “academic competency”, which I’ve been writing about for two years, we aren’t seeing a corresponding huge boost in student academic outcomes, and all research continues to show that, at best, the link between teacher cognitive ability and student outcomes is twitchy and unreliable. All research continues to play the Reverse Drinking Game and ignore student cognitive ability. Math professors assure us that the only difference between “math people” and everyone else is effort, and that anyone with an IQ of 70 can learn algebra.

So, here’s what I think, but can’t prove: our teachers are pulled roughly from the same pool as always, which is the 35-50% for elementary and special ed teachers, and 50-75% for secondary content teachers. But the bottom quintile or so is gone because of higher licensure standards, so the average has increased. This has resulted in far fewer black and Hispanic teachers, particularly black teachers. Existing black teachers are also being forced out of the profession by new requirements (hence the Mumford impersonation fraud ring).

Remember, anyone who pushes for improved teacher qualifications is saying, in effect, we need fewer black and Hispanic teachers. And, as the recent TFA study’s big takeaway shows, all you get for largely eradicating black and Hispanic teachers is, maybe, .07 of a standard deviation.

Just today Dara Zeehandelaar commented on my blog:

The point is that students might benefit if traditional certification programs were more selective in who they admitted (for example, by admitting students with higher GRE scores, higher undergraduate GPAs, undergraduate or even graduate coursework in the content area in which they want to teach, professional experience, etc. — the same things that TFA looks for). “Selectivity” in this case has nothing to do with race.

She said this with a straight face, too.

Sorry, Dara. Selectivity in this case has everything to do with race.