There is a familiar saying that goes, "those who can, do; those who can't, teach."  I've never agreed with that cynical axiom, which seems like an unfair bit of sniping against a truly noble profession.  But I do largely agree with the less familiar addendum to the saying, which claims that "those who can’t teach, teach teachers."

It is commonly said that there is a crisis in American education.  No doubt it is that perception which underpins the existence of this very site.  For some, that perception is enough to justify an assumption of incompetence in the average teacher.  But I would venture to guess that the greater share of failure generally ascribed to teachers is, in fact, the failure of those structures tasked with making our teachers better.

In my young freelance career, I have had the opportunity to write and edit various columns and academic works in the areas of education and educational policy.  I also have the pleasure of counting as my best friend a teacher whose several years of experience already range from primary school to college instruction.  Several other people with whom I am at least incidentally acquainted are teachers or aspire to be.  On the basis of all of this, one observation continues to strike me above all others: that there is an absurd disconnect between the realm of education policy and the realm in which actual classroom teaching goes on.

Though my observations on the issue are various, they are also less than comprehensive, so I can hardly begin to imagine how many resources are wasted every year on the gathering and analysis of trends that have shockingly little to do with educational outcomes or the performance of teachers on the ground.  In my experience surveying graduate-level research, I see work investigating methods for how to improve students' self-esteem, how to increase engagement by the surrounding community, what government profiles are conducive to educational reform, how consultants can foster change in the organizational structure of particular schools, and so on.  What I see precious little of are studies indicating how educators can teach better or how students can learn more.

To give one particularly egregious example, I once encountered a published study on the topic of city governance models and their effect on public education.  The paper stated plainly in its executive summary that the research provided no evidence that governance structures had any influence whatsoever on student achievement.  It said this upfront, and then the paper went on for another 130 pages.  To my mind, if your area of study is educational policy and you strike on a particular subject that has no influence on educational outcomes, then that is all you need to say about the subject.

Naturally, though, over the course of more than 130 pages, the governance paper went to great lengths to argue for its own relevance.  It explained how efficient governance structures could increase public commitment to education, increase funding, and increase stability.  Frankly, though, I don't think that a stable, well-funded public school with poor student performance is any better than a chaotic, de-funded school with equivalent performance.  In fact, I would hope that a school with poor performance would lack internal stability as it struggles to restructure, retrain, and experiment in order to find alternatives that work.

No topic of study within the areas of policy and consulting should have any bearing on classroom structures or government mandates if it isn't connected to definite improvements in teacher strategies or the academic performance of individual students.  To the extent that there can be said to be a need for sprawling academic departments in educational policy and educational consulting, those are the only topics that account for that need.

It is frightening to think that there are great masses of research ostensibly to be used to influence broad educational policies, which have little or nothing to do with improving student performance.  And governments actually both implement and reverse the associated policies as new research emerges to support competing speculations about what second-order changes might remotely contribute to different academic outcomes somewhere down the line.

It is a discomforting fact of education that each generation of students is also a generation of guinea pigs.  Policies change because we guess that the new ones will be better for students, but we don't really know until years later when a greater or lesser number of those students have failed.  And even then we don't know which specific set of factors to praise or blame, so we try them in different combinations.

Upsetting though that fact is, it is unavoidable.  But so long as students must be guinea pigs, I would much prefer that they be guinea pigs who can be observed directly, and not analyzed as an arcane set of numbers and data points.  Inside a classroom, teachers can respond to the needs of each student day-by-day, and if she's effective at doing so, the improvements emerge independent of governance structures and manipulations of self-esteem that are unrelated to performance.  Yet policymakers seem obsessed with finding explanations that wholly transcend the space between each teacher and her students.  And they evidently believe that every guinea pig can be perfected at once by only changing the maze they're made to run through.

This is all part and parcel of the severe addiction in the United States to educational data, and it misses the obvious solution to the problem: if students are performing poorly, teach better.  Of course teacher performance is not the only factor influencing outcomes.  Classroom culture matters; parental and community engagement matters; funding matters; student background matters.  But none of these considerations are in themselves grounds for altering the policies that actually govern how teachers teach and how administrators evaluate them.

The same addiction to data that promotes irrelevant policy studies is also the foundation of an onerous system of standardized testing, which insists that quantitative data alone can determine the value and competence of each teacher.  This remote sort of evaluation recently named Anderson School teacher Carolyn Abbott the worst eighth grade math teacher in New York City despite the fact that her seventh graders scored in the 98^th percentile and the eighth grade tests on which she was being evaluated were comprised of material that her advanced students had learned in fifth or sixth grade.

Context matters.  If policymakers believe they are able to influence academic outcomes across classrooms, districts, and states by changing anything other than the level of talent in their teachers, I expect that they'll be consistently frustrated in their efforts, though not as frustrated as their teachers are.  It's no wonder that the study of governance structures' effects on public schools found no correlation to improved academic performance.  The space where government and community meet the school is not the space where teaching happens.  It happens in the classroom, and it is there that change must be both sought and implemented.  In absence of that, all the rest is just frills.

The Irascible Professor comments: The IP agrees with most, but not quite all, of what Edward has said.  Certainly, here in the U.S. we collect enormous amounts of data related to K-12 education.  And as Edward notes a good deal of that data is seriously flawed.  There is an old saying in the data analysis game; namely, "garbage in, garbage out."  Poorly structured studies and evaluation tools produce information that basically is worthless.  The Carolyn Abbott case clearly supports that notion.  Unfortunately, much of the data collected by education researchers is flawed because they seldom employ rigorous research models and techniques.  Too often variables are poorly constrained and inferences are drawn from populations that are not representative of all students or teachers.

But, Carney's assertion that the quality of the teacher is the most important factor in student learning is supported by many reasonable studies.  So, at least those data are relevant.  We know that high quality teaching improves student performance.  One thing that we need to know is how to attract better teachers to the profession.  However, it is unlikely that we ever be able to staff every classroom with a superbly talented teacher.  So, another very important question is how do we improve the performance of the "average" teacher.  Politicians have tried to do this by wielding the stick; namely, using such measures as calculating the "value added" to standardized test results.  Unfortunately, this is a seriously flawed approach because while it's possible to have a standardized test, it is not possible to have a standardized student.  While K-12 class sizes are not as small as we might like them to be, they certainly are small enough to have significant variations in both the average intelligence and the average motivation of students from year to year.  Value-added testing doesn't always account for these factors sufficiently.  In addition, standardized testing doesn't necessarily help teachers to be be better at teaching.  More often it motivates teachers to learn how to be better at preparing their students to be better test takers.

The Irascible Professor invites your  .