Dialogue With Dylan Wiliam on Teacher Quality: Part II
The examples you described are compelling, and descriptive of the kinds of education we would want all students experiencing. The question, of course, is how we get there.
For many years, I went along with the argument that smarter people do just about any job better. This is a clear finding from the intelligence research, although very unpopular with those who like to believe that IQ tests measure only the ability to do IQ tests. However, the expertise research seems to show clearly that almost anything that predicts quality of performance at the initial stages diminishes greatly in predictive power over time. A recent study found that initial teacher quality (as measured by student progress in the teacher's first year of practice) accounted for less than 5% of the variation in teacher quality five years later. Of particular interest to me is the fact that the best teachers (top quintile of value added) did not improve their productivity over the first five years of their careers, while the least effective (bottom quintile) improved radically.
On the other hand, I do not doubt that you are right in saying that the kinds of teachers you describe are successful at least in part because of their general intellectual capability, and their profound grasp of the subject.
I think there are at least two resolutions of this apparent contradiction. The first is that you are talking about particular individuals, at the extreme end of the competence scale, and therefore analyses of the generality of teachers are unlikely to apply. But the second possibility is that the deep intellectual power that you observed was the result of many years of deliberate practice, driven by a passion to improve. The evidence from the expertise research is that deliberate practice is not inherently enjoyable, so the only individuals that can rack up 10 years of maximal deliberate practice are those who push themselves to improve even when no one else is pushing them. This may come from a passion for young people's success, or from a passion for the subject, and of course, ideally, both. So I would say that what you saw in Harvard-Westlake and Seattle was the result of rather idiosyncratic routes to excellence that should, of course, be expanded as much as possible, but is unlikely to produce the two million expert teachers we need in the United States. In other words, you are right, but I cannot see a way that your observation leads to a substantially improved teaching force within the next 15 to 20 years, which I take to be the crucial horizon in terms of the continued economic prosperity of the United States.
Picking up the other points in your email, I agree that there is no way to improve expertise in teaching without attending to significant shortcomings in subject matter knowledge, but my belief is that this comes about through deliberate practice (and of course, at this point I would use what philosophers have described as a "No true Scotsman move" to regard any practice that did not improve subject matter as failing to satisfy the requirements of deliberate practice). I am certainly not against, in this regard, giving teachers direct instruction on deep principles of mathematics. Indeed, in one study in the 1990s we found that the best predictor of the mathematical progress made by elementary school students in mathematics was not the highest level of mathematics course taken by the teacher but the extent to which the teacher saw mathematics as "connected"—in other words, that they saw fractions, decimals, percentage, and ratio not as separate topics but as different approaches to dealing with non-integral numbers.
As for general intelligence, I would be delighted in the United States if, like Singapore and Finland, we could get the smartest people in the country wanting to be teachers, and we should of course do everything we can to work towards this. After all, IQ does explain some variation in expert performance even after ten years, but it's a long-term project. So we need to work with the teachers we have as well. This does, as you say, involve raising difficult issues about teachers' subject knowledge. One way we have found to do this is through the use of what I call "hinge-point questions," which are questions designed to check on a classes' understanding of a key point before moving on. The idea is that these hinge-point questions encapsulate what we know about student misconceptions, which of course many teachers share, and by presenting "answer-choice rationales" (the reason someone might have chosen a particular incorrect response) we engage in a form of "just-in-time" teacher subject knowledge development. Of course, all this assumes that teachers want to improve, which I would address by changing teachers' contracts to require them to demonstrate improved practice in order to keep their jobs. Some details of how this can done are in a chapter I wrote for the Institute of Public Policy Research recently.