Eric Hanushek on the "teacher effectiveness gap"
This is a guest post from Eric Hanushek, the Paul and
Jean Hanna senior fellow at Stanford University’s Hoover Institution, in response to Mike
Petrilli’s essay, “The
‘teacher effectiveness gap’ was just a myth: 3 implications.”
Let me try to put some of the issues raised by Mike Petrilli’s recent post in perspective. Much of the research has found substantial
variation in teacher quality within all schools. It is difficult to
ascertain how much variation there is between schools, but I don't think
answering that question is key to policy.
- We want to improve the quality of teachers
everywhere—which in my opinion calls for weeding out the ineffective teachers
- Even if little of the variation in teacher
quality is between schools, it does not eliminate concerns about what is
happening in disadvantaged schools.
- A recent EdTrust
West paper—which is great and which tried to analyze the issues in a
serious way—finds some substantial differences in average quality (biased
against disadvantaged students) in Los Angeles—so if a serious analysis of New
York City finds no bias, we will still be left with policy issues outside of
- It would not make sense to attempt to
redistribute good teachers from middle-class to low-income schools, but we can
still pursue policies that try to hold top teachers in poor schools. Thus,
direct incentives (tied to effectiveness of teachers) would make sense in poor
schools. That would be a direct way to increase average quality. Moreover,
apropos the recent LA decision, it might be easier to break the bad contract
provisions in schools serving predominantly poor kids.
- Even if average teacher quality is the same
across middle-class and poor schools, the poor kids in general will score lower
because they come with less average inputs from family and neighborhoods—and we
have to deal with that as a nation.
- It does not take indirect evidence from the
New York Times, crossed with
Marguerite Roza’s data, to infer that teacher characteristics are not good
proxies of effectiveness. We have much more extensive and persuasive
evidence on this.