In my last post, I questioned the idea of the senior, burned-out teacher being a major policy concern, simply on the basis of the fact that there really aren't that many senior teachers to begin with.

But what does the literature say about relative teacher effectiveness late in a career? What do we know about teachers who enter the late stages of their professions, and their effect on student outcomes? Is there a body of evidence that supports the contention that we need to clear out a lot of dead wood in our schools?

The short answer is: no, not really.

- Parents Across America has a short list of studies showing positive gains in student achievement correlating to teacher experience long through a teacher's career.  Among them is a study of Project STAR data from Tennessee by Raj Chetty and John Friedman (yes, those guys), which found positive correlations in Kindergarten between teacher experience (up to 20 years) and test scores.

- In the same post, Leonie Haimson quotes a study of reading gains (behind a paywall):

“A teacher who had been teaching at a particular grade level for more than 5 years was positively and significantly associated with increased student achievement (effect size=.27)…grade level experience was sizable compared to race (minority status effect size = -.33) and SES (economically disadvantaged effect size= -.08)….Teachers constantly improved teaching effectiveness until the 21st year and declined beyond that.”

Hmmm... "declined beyond that" doesn't sound good -- until you get to p. 227, where you find that teachers "peak" at their 24th year of grade-level experience, and teachers at their 40th (!) year still outperform their third year colleagues.

-  In a brief from CALDER, Jennifer King Rice brings together data from several studies to show a mixed bag of results for senior teachers. In lower-poverty schools, teachers with 28 or more years of experience perform near or above their colleagues who have at least three years of experience (p. 5). The senior teachers significantly outperform the Year 3-5 teachers in reading gains; math is much closer. Higher-poverty schools show vacillating effectiveness depending on the bin for experience, making it difficult to draw any definitive conclusions about senior burn-out.

- Rice cites Boyd et al, who show no appreciable decline in elementary and middle-school math teacher effectiveness (in New York City) between teachers with 6 to 10 years of experience and teachers with more than 20 years of experience.

- Harris and Sass have done some interesting work on teacher effectiveness using Florida data. One CALDER paper breaks down experience in several bins and compares value-added measures of teacher effectiveness in math and reading. They generally find "positive, but mixed" effects from experience. But as to senior burnout, the data is hardly conclusive: when fixing school and student effects (Table 3, p.39), only in Grade 9/10 math do novice teachers outperform those with 25 years or more of experience. Given the non-random assignment of courses in high school and the way high school content aligns with standardized tests, I'd approach any conclusions from this with great caution. (This is a real knotty paper -- I'm giving it short shrift by encapsulating it this way.)

- Helen Ladd gave a presentation, based on North Carolina data, where the effectiveness of elementary teachers dips slightly after the 27th year (slide 5), but still outpaces those with less than 12 years of experience. High school teachers with more than 27 years are outperformed by all other experience groups, but the difference with novice teachers is quite small (0.007 SD) and, again, I'm left to wonder if non-random assignment to courses, combined with those courses' alignments with End Of Course tests, is an issue.

- Chingos and Peterson also look at some Florida data. Their "two-stage" model has quite a number of caveats attached to it, but it is the only one appropriate to looking at teacher effectiveness past 20 years. Most of their figures on pages 36 and 37 do show a slight decrease in value-added scores after around the 27th year (the Grade 6-8 reading figure is a real anomaly). But there's no evidence (save that middle school reading figure, which is so different from the others) that novice teachers are any better than senior teachers; quite the contrary in the elementary years.

So what are we to make of this? Well, TNTP implies that there's some evidence that senior teachers do burn out. I'd say, however, that evidence is, at best, mixed, and not even close to the evidence we have that teachers gain most in effectiveness within their first few years (even TNTP concedes that point).

Further, there's no body of evidence that novice teachers consistently beat out senior teachers. Yes, it's possible that teachers late in their careers do not get the test score gains they may have earlier -- but any idea that replacing senior teachers with novice teachers should be a major policy thrust is simply not borne out by the facts we now have.

Further: all of this evidence, mixed as it is, is based on a very limited subset of the entire curriculum and types of teaching positions. And it's all based on standardized tests. And the practical effect sizes are quite small.

One thought I had when reviewing all this was that senior teachers may be more inclined, at the late stages of their careers, to look at test score gains as less meaningful for them or their students. Maybe senior teachers feel they have the freedom to focus on things they find more important to their students' educations than test scores gains.

Here's one veteran's perspective:

I have been teaching middle-school math for 20 years. I feel like a veteran of a war. Almost like the veteran sergeant in an old war movie who is still standing after numerous campaigns and is there to greet the new group of raw recruits. I have seen administrators come and go. I have seen curriculum ideas come and go. I have seen all manner of policies come and go. I have taught students who were considered unteachable. That being said, I feel that I am at the top of my game this year.

In an effort to align my instruction with the Common Core State Standards, I am incorporating Singapore Math into my daily lessons so that students who have poor number sense can become better math students. We don’t have a textbook – I am using my years of experience and knowledge to pull material together that I feel will best serve my students. I use daily assessment to guide my teaching. I share data/test results as feedback with my students when they take yet another standardized test. I have control of my classroom and can cover the material I need to cover in the 50 minutes allotted without spending half of the time doing crowd control.

There are lots of new teachers who are great to work with. They have a lot to offer and their input is appreciated and applauded. But there are new teachers coming in who believe that just because I am old enough to be their mother in some cases, I have no value. I am more than ready to listen to their ideas. If they are good ones that I feel will benefit the students – great. But if I don’t greet their idea with pomp and fanfare (because we’ve already tried it with little success), it doesn’t mean I’m burned out.

But I am tired of bright new teachers of whatever age, coming into my building and just from looking at me, assume that it’s time to shove me to the side because they’ve been brainwashed that veteran teachers have no value in this “grand new world of education.”

Amen. Senior teachers should be looked at sources of wisdom and history, just like in every other profession. Yes, undoubtedly some do burn out -- just as in every other profession. That doesn't mean, as reformy pundits too often suggest, that clearing out veterans to make way for novices is a reasonable strategy for improving student outcomes: we just don't have the evidence to support that policy.

Let's talk about pension incentives and senior teachers next...

A pressing education policy concern? Not hardly.