Until now, I`ve never been able to grasp
what kind of picture they had in their heads when they
decided that the way to close the
black/Hispanic-white/Asian Achievement Gap was through
more school testing.
Mandating frequent testing of K-12
students to solve the problem of
inequality always struck me as much like trying to solve
the problem of height inequality by requiring that everybody
play a lot of
basketball. That wouldn`t make
short people taller—it would just make their shortness
But now, I think, I`ve finally stumbled
upon the wacky analogy unconsciously underlying the
conventional wisdom about how more
testing would Leave No Child Behind. So I`m going to
take some time to explain the mental framework behind so
much of mainstream school reform thinking, as exemplified by
popular soundbite on
"the soft bigotry of low expectations".
The great thing about espousing the
conventional wisdom, as Kennedy and Bush did, is that you
don`t have to justify it very much. You just handwave
problems away. And, because you aren`t questioned
you have in mind for how your law is supposed to work,
it`s hard for skeptics to instill any doubts.
Recently, the Obama Administration has
been waking up to the fact that the NCLB mandate requiring
every public school student in America to score
(on a scale running from
"below basic" to
both reading and
tests by 2014 has left them with a big mess.
Barack Obama has been campaigning to
change the law, because up to 80 percent of schools will
otherwise be officially found
year and therefore must be punished. He complained last
"That`s an astonishing number. … We know that four out of
five schools in this country aren`t failing. So what we`re
doing to measure success and failure is out of line."[Obama
Urges Education Law Overhaul, By Helene Cooper, NYT,
March 15, 2011]
Why the sudden surge in schools classified
under the NCLB?
while the people who wrote the NCLB might have been naive
about testing, they weren`t at all naive about the politics
of blame. Accordingly, they required only modest improvement
in test scores during the early years of NCLB. The goal of
100 percent proficiency was cleverly backloaded.
This has created a hockey stick-like
profile of required performance that zooms up to 100%
proficiency at the last moment—like the student who tells
his parents at 8pm they must go to the store
now because, he
just remembered, his big science fair project is due in the
this way, most of the penalties wouldn`t kick in until the
second decade of NCLB`s existence—when, by the way, Kennedy
is dead and Bush retired.
Amazingly, however, it now turns out that
passing a federal law requiring every student in the country
to be "proficient" at reading and math
really doesn`t mean they will be, in fact,
"proficient". (With students,
failure is always an option.)
Why did the public policy mainstream place
so much faith in the power of K-12 testing to eliminate
achievement inequality? Why did they think that it was
feasible for everybody (or practically everybody, for that
matter) to become
"proficient"? Why did they think it was crucial for
everybody to pass—but not very important for some people to
do much better than passing?
been stumped by these questions for a decade.
At last, however, I`ve realized there
is at least one
common kind of test that
rather like what the conventional wisdom expected from K-12
everybody (outside a few cities) takes this particular kind
of test. And the great majority of them eventually buckle
down and pass it.
passing is all that`s required. While high scorers on the
SAT get into fancier colleges and are more likely to go on
to postgraduate studies, high scorers on this other kind of
test aren`t treated much differently from those who just
is this widespread test that provides the mental template
for NCLB and for so much else in the school reform brouhaha?
interests of dramatic suspense, I`m not going to reveal it
yet. See if you can guess!
a fundamental distinction about testing that is poorly
understood. Tests can be thought of as measuring either:
Relative performance versus other test-takers; or
Absolute performance against some predetermined level of
school reform rhetoric assumes that the latter is how tests
inevitably work. That`s why we hear constantly about how we
must make the standards more rigorous to raise performance.
of an absolute test is much easier to give a pep talk about:
"Every single one of you must learn how to use the quadratic
formula! It will be hard, you will lose sleep studying it,
I know that, in the end, each and every one of you can
and will do it!!"
The problem with our thinking about
education in modern America is
not that adults
address children in this manner. That`s good. The problem is
that adults are supposed to talk to other adults about
education as if they, too, were children.
A grown-up conversation about school
reform ought to mention that the most
for K-12 achievement tests are college admissions tests,
such as the SAT and ACT. And those are relativist rather
than absolutist tests. The SAT clearly doesn`t make people
more equal. The SAT is deliberately designed to leave many
and send a few far ahead.
The SAT doesn`t have a passing score to
push everybody toward any
minimum competence. Instead, the SAT elaborately
distinguishes among students for the benefit of exclusive
more or less works at what it says is does.
Indeed, a large fraction of the most
successful and enduring tests—such as the
ACT, SAT Subject Tests,
GRE, LSAT, MCAT, DAT, and GMAT, military`s
AFQT—are all built upon the assumption that human
performance is distributed relativistically upon a bell
curve. A median score and a standard deviation are
determined. Thus, if the median is, say, 500 and the
standard deviation is 100, somebody who scores a 600 ranks
at the 84th percentile.
Among famous education exams, the
Advanced Placement test started off absolute—scoring 3
on a 1 to 5 scale was
"passing". But colleges have been gradually relativising how they
treat scores—e.g., a 5 on U.S. History might get you credit
for two semesters of history, a 4 gets you one semester, and
3 or less nothing. It`s up to each college: MIT only blesses
5s and Caltech doesn`t give any credit. Moreover, less than
ten percent of 17-year-olds take any single AP test in a
year, so the AP isn`t much of a model for the mass of
Universities use SAT-type test scores as
each sees fit.
Caltech, for instance, typically wants higher scoring
students than the adjacent
Pasadena City College, the junior college that my father
attended in the 1930s.
Scoring upon a bell curve is both
mathematically elegant and pragmatically useful, which
is why it`s so widely used. Psychometricians feel much more
confident about what they are measuring when told to devise
tests that are explicitly relativist.
But there are some
assumptions behind bell curve scoring whose implications are
not at all popular. When
Richard Herrnstein and Charles Murray spelled out these
out in epic detail in 1994 in
The Bell Curve, they didn`t make themselves the
toast of the town. ["The
Bell Curve" and its critics, By
Commentary, May 1995]
The point of relativistic tests such as
the SAT is not to make sure that every student knows what he
or she needs to know: it`s to find out who is best. Nor
these test designers claim that administering their test
will make the students smarter. In fact, the designers worry
when scores go up that perhaps somebody is
gaming the test.
Of course, relativist tests do top out at
some score. But that`s merely for convenience and
cost-effectiveness. It`s obvious that an 800 on the SAT
doesn`t necessarily represent the ultimate in human
intellectual proficiency. What would
John Updike have scored on the SAT Verbal if the test
had been 48 hours long? 1050? What would
John von Neumann have scored on the SAT Math? 1100?
In turn, it`s hard for those of us who
grasp these basics of psychometrics to realize that people
at the Kennedy/Bush level of
intellectual sophistication and/or
substance abuse don`t find the logic of bell curve tests
intuitive at all.
Further, to reach Kennedy/Bush levels of
success in politics, you can`t go around saying things like,
"Well, obviously, your SAT score shows you aren`t smart enough to get
into Caltech, so you`d better come up with a
practical plan for your life."
today, you probably
shouldn`t even think
Instead, in public life, you get rewarded
for uplifting demagoguery. When Bush attributed poor
performance in school to
"the soft bigotry of
low expectations", he most likely
quite sincerely meant it.
conventional wisdom espouses the notion that there are
academic accomplishments that every student should have
because they are crucial for his future. For example, here`s
the United Federation of Teachers explicating this
"Algebra 2 is often
described as a `gateway course` because it correlates so
closely with college success. Students who complete Algebra
2 are twice as likely to earn a bachelor`s degree as
students who do not, and passing Algebra 2 reduces the gap
in college-completion rates between African American and
Latino students and their white peers by half."[Beyond
high school graduation: What the data tell us,
by Maisie McAdoo, UFT.ort, April 1, 2010]
in my experience, techniques that are reserved for Algebra 2
aren`t going to be used on the job by the great majority of
workers, even among college graduates. Nevertheless, I don`t
doubt that success
in Algebra 2 in high school does correlate with success
in college—even in classes that don`t use Algebra 2 at all.
That`s because Algebra 2 measures logic, powers of
abstraction, and work ethic, all of which are
good things to have at college.
aren`t supposed to think like that.
You are supposed to think like this: Success in life
correlates with graduating from college, which correlates
with success in high school Algebra 2. Therefore, knowing
Algebra 2 makes people a success in life—so the public
schools must teach everybody Algebra 2!
Unfortunately, not everybody who takes Algebra 2 learns
Algebra 2. Which is why the UFT is obliged to go on:
2008, the Alliance for Excellent Education reported, of
90,000 students who took an end-of-course Algebra 2 exam,
the average score was 27 percent."
ensuing arguments in mainstream public policy discourse are
over whom to blame: poverty, teachers, racism, government
schools, parents, YouTube, or whatever.
common test is the closest model for school reform`s
conventional wisdom? What test is the opposite of the SAT in
that all the emphasis is put upon achieving a passing score?
That`s right, you guessed it:
the driver`s license
When the conventionally-minded imagine
that K-12 tests will bring about equality, they are assuming
that these tests will work more or less like the test your
teenager takes down at the
Compare it to the SAT:
driver`s test is supposed to be scored in an absolute
fashion, not relative to what everybody else is doing.
Given enough tries, most people eventually pass.
whole point of driver`s test takers is to reach the
minimum level of competence to be allowed to drive. The
passing score is intended to be good enough.
There are very few rewards for acing the driver`s test.
If you get a perfect score on the driver`s test, you
won`t get a letter from
imploring you to try out to
one of their racecar drivers. NASA doesn`t invite
you to enroll in astronaut school.
Granted, for some people the driver`s
license exam is a
stepping stone to harder license tests, such as
for driving an 18-wheeler. For most people, though, it`s
the beginning and end of the line.
the government makes the driver`s test harder, teenagers
will study more for it.
impression is that the driver`s test is more difficult than
when I breezed through it in the lackadaisical 1970s. The
average age when young adults get their first driver`s
license has gone up in many states. But for teens, this is a
high-stakes test. So many work hard at studying for the
written part and practicing for the behind-the-wheel part.
Obviously, once you come out and articulate this mindset,
then the driver`s license test sounds like a pretty dumb
analogy for school achievement tests. Driver`s licenses are
absolute, school achievement tests relative—hence all the
concern about The Gap.
It`s easy to develop a more sensible goal
than the NCLB`s implicit intention of raising black and
Hispanic average performance by about a standard deviation
simultaneously not letting whites and Asians improve
(because that would merely perpetuate The Gap).
Of course, given the nature of
the IQ Bell Curve,
it would also leave whites and Asians would still better
equipped—but is our objective improving everyone`s
potential, or equality?
A good question.
However, we needn`t
face it—because any common sense goals in education policy
are unlikely as long as the conventional wisdom is protected
from serious questioning.
end note: Ironically, both Ted Kennedy and George W. Bush
had a lot of time during their adult years to think about
George W. Bush had his license suspended
for either one month or two years (sources differ) after his
drunk driving arrest in 1976.