Nicholas Kristof has discovered education. Health care is no longer our greatest national shame—education is. skoolboy thinks that responsible op-ed reporting can’t be far behind. Breathlessly, Kristof reports in Sunday’s New York Times that teachers are “astonishingly important.” “It turns out that having a great teacher is far more important than being in a small class, or going to a good school with a mediocre teacher,” he writes. “A Los Angeles study suggested that four consecutive years of having a teacher from the top 25 percent of the pool would erase the black-white testing gap.”
Wow, erasing the black-white testing gap in four years sounds like a pretty good deal. And just from being taught by some really great teachers! There must be some evidence of this for it to show up in the New York Times, wouldn’t you think? Some study somewhere that actually showed that black students exposed to teachers in the top quarter of the teacher effectiveness distribution for four years in a row can routinely move from the 16th percentile in the test score distribution (roughly the black average) to the 50th percentile (roughly the white average)?
Maybe that Los Angeles study will show the way. Nah, that’s just a “suggestion” by Robert Gordon, Tom Kane, and Doug Staiger that the five percentage point increment in performance from having a teacher in the top quartile, and the five percentage point decrement from having a teacher in the bottom quartile, could cumulate over time—a 10 percentage point swing for four years in a row would more than close the 34 percentage point gap between the average black student and the average white student.
The problem is, as eduwonkette pointed out last summer, Brian Jacob and his colleagues have shown that these effects do not cumulate. Only about 20% of the effect remains after a single year, and only about 12% after two years. After two years, then, the 10 percentage point swing is down to about 1 percentage point.
It gets worse. The notion of a “great teacher” identified via value-added effectiveness implies that we can identify who these teachers are, and they’ll always be great. The reality is, however, that the vaunted value-added methods show that a teacher who is “great” one year may not be so hot the following year. In a recent National Center on Performance Incentives report, Dan McCaffrey, Tim Sass and J.R. Lockwood find that the year-to-year correlation in teachers’ value-added scores are in the range of .20 to .30. This also implies that a teacher whose students gained five percentage points in one year might have students only gain one or two percentage points the following year. Better than chance, to be sure, but what is a matter of chance is whether you get a teacher when she or he is having a good year or a bad year. And the likelihood of big cumulative effects from exposure to “great teachers” just isn’t there.
Sorry, Mr. Kristof, no magic bullets in identifying and rewarding “great teachers” who can effortlessly close the achievement gap. Now, can we get back to talking about teaching instead of teachers?