Grading Teachers, Failing Kids
What would children gain from teachers who have the highest evaluations as measured by the metric that Mayor Rahm Emanuel wants to implement for Chicago’s public school (but not charter school) teachers? This is precisely what three Columbia and Harvard economists researched in a study that has received wide attention from the media and politicians. According to the authors’ own interpretation—the sky. But if you read the study itself—nothing at all.
The study examined the incomes of adults who, as children in the 4th through the 8th grades, had teachers of different “Value Added” scores, with Value Added defined as improvement in the scores of students on standardized tests. The researchers found that the average wage and salary of a 28 year old who had an excellent teacher was $20,509 in 2010 dollars, $182 higher than the average annual pay of all 28 year olds in the study.
How does this compare to the average salary and wage of a 28 year old in this country? The authors chose not to make the comparison, but we can make it for them. They excluded from their study people whose income was higher than $100,000. As we shall see, this exclusion is problematic; but to do the comparison we must do the same. The average salary and wage in 2010 of a 28 year old who earned less than $100,000 a year was $29,041, 42% higher than the income of a 28 year old in the Columbia-Harvard study who had an excellent teacher. In other words, even if we accept the spin the authors put on their numbers, having an excellent teacher does not pull people out of poverty.
But tiny as it is, even the gain of $182 a year from having a high “value-added” teacher may not really exist. First, the exclusion of people with high incomes involved some 4,000 individuals, or 1.2% of the sample. The authors justify it by claiming that such people are outliers. But what if it turned out those high income earners had “bad” teachers? Including them in the study would have completely changed the results. Excluding a large number of the best performers from a study about the effect of teaching seems strange.
But that’s not all. When the authors examined the incomes of 30 year olds they found no statistically significant result. Why? In the study the authors explain this by the small number of 30 year olds in their sample. And in their interviews with the media and in public presentations the authors do not mention this result at all. Yet the number of 30 year olds in their sample is 61,639, and these are all
students who went to school in the same city. Is this really a small sample? To gain an appreciation for the size of the sample consider the fact that in order to estimate the unemployment rate that it publishes every month, the Bureau of Labor Statistics relies on a national survey of 60,000 households with an average of 1.95 adults in each. Surely if 120,000 people are a good size sample to study a labor force of 150 million people spread all over the country, a sample of 61,639 is a good size sample to study a population of fewer than 5 million elementary school students who all come from the same school system. By any measure the sample size is not only adequate, it is fantastically huge, and the result is not statistically significant.
The statistically insignificant results for 30 year olds may have been inconvenient for the authors for another reason. An increase of $128 a year is small by any standard, so the authors resorted to estimating a lifetime increase in earnings due to this increase. To do this they assumed that the percentage increase in income, 0.9 of one percent, which they estimated for age 28, holds for each year of a person’s working life. And perhaps this is why the authors chose to ignore the results for the 30 year olds. What their findings actually permit them to claim truthfully is that an excellent teacher increases average annual income by $128 at age 28, and that this effect disappears at age 30. But then there would have been nothing to report.
Does teacher quality matter? Of course it does, and it should, therefore, be evaluated: by peers, principals, outside experts, students and parents. But it cannot be measured, and Mayor Emanuel of Chicago knows it. Why else would the non-union charter schools be exempted from the same measurement? The evaluation method the mayor wants to implement is nothing more than a union-busting, pro-privatization trick that is based on partisan economics.
Moshe Adler teaches economics at Columbia University and at the Harry Van Arsdale Center for Labor Studies at Empire State College. He is the author of Economics for the Rest of Us: Debunking the Science That Makes Life Dismal (The New Press, 2010), which is available in paperback and as an e-book.