Wednesday, July 28, 2010

Better teachers and smaller classes make a huge difference

Mr. Chetty and his colleagues — one of whom, Emmanuel Saez, recently won the prize for the top research economist under the age of 40 — estimate that a standout kindergarten teacher is worth about $320,000 a year. That’s the present value of the additional money that a full class of students can expect to earn over their careers. This estimate doesn’t take into account social gains, like better health and less crime.

That's from a New York Times article summarizing impressive new research by Raj Chetty, John Friedman, Nathaniel Hilger, Emmanuel Saez, Diane Schanzenbach and Danny Yagan.

Here's a graph of their key results:

Students were randomly assigned to classrooms over 20 years ago, making this a valuable experiment. The results reflected a combination of teacher quality, smaller classrooms and peer effects.

We would be better off if we paid our teachers a lot more, and attracted more of the best and the brightest to this profession.


  1. teachers think they're helping(working for intangibles) so we get to pay them less. examine the difference in the motivations of those in high paying professions like business mgmt and marketing: more "help yourself" professions.

  2. The problem is that managing a kindergarten class has more to do with your personality than your education or skill level. If you increased the pay of kindergarten teachers to be competitive with doctors and lawyers, you might see more intelligent and highly educated people enter the field for the paycheck, but the quality of the teaching might not be commensurate with the IQ-bump of the teacher. Higher pay is probably not the answer, but more flexibility in style and more pay INEQUALITY (in contrast to the union system that acts to help all teachers, despite their merit) would probably help.

  3. It is an early release of research.

    Slide 44, figure 6d, is troubling. It represents the basis of the statement that better class quality increases earnings.

    The positive earnings effect points, above expected trend line (which is a positive earnings effect in a negative earnings result) or positive earnings, span almost the entire range of class quality, from poor to good, from negative to positive class quality.

    One would expect the trend line to parse the plotted points, with points above and below through the entire range of class quality.

    The negative effect is clustered around the zero class quality point. To me the plot suggests teacher quality is a residual variable capturing a missing, hidden, unmeasured variable responsible for the positive earnings effect. With the additional unknown variable, the trend line would parse the points and teacher quality would probably diminish in importance, if not drop out altogether.

    There are common statistical tests which could negate my comment and I assume before publication, the tests will be checked. Until then, I will keep an open mind about teacher quality.

  4. Please note that the study did not randomize for birth order or age of student upon entry. They have been shown to have a positive correlation with academic success in other studies. I assume this was because the fact of this correlation was not known when placement was done. So we don't know if they have isolated a variable after all.

    Also I would point out the inferences drawn from the study confuse absolute and relative benefits: even if you fully credit the study, the advantage is a relative one, which requires that there be some who do not have the advantage. If everyone has the advantage, it isn't an advantage (excluding against those who immigrate after kindergarten, or are homeschooled). If there is no relative advantage, you could pay the teachers almost any wage, with no effect on outcomes. So, even if you credit the study, the inferences being drawn are excessive as there are diminishing returns to scale.