Put Up The Proof, TFA

Anthony Cody points us to a remarkable claim from Teach For America's Wendy Kopp:

In this interview, Wendy Kopp states: "On average, our corps members stay in the classroom for eight years." Honestly, I find this claim incredible.

I have firsthand experience with Teach For America. I worked in high poverty schools in Oakland, California, for 24 years. In my last four years (2007 to 2011) I led a mentoring program that was developed to provide support to the many new science teachers we had in the District. The District was relying on TFA and a program affiliated with The New Teacher Project to fill vacancies. Unfortunately this simply fed into the cycle, because these novices did NOT last an average of eight years. In fact, three years after they started, 75% of them were gone. That meant we were constantly training the replacements for the ones that were leaving. While our mentoring program was somewhat successful, we struggled because so many of the TFA teachers came with the intention of only teaching two years.

I have reviewed the research on Teach For America retention, and I have never seen evidence that they have an average service record of eight years. I would like to see some substantiation of this claim. If you follow some of the links of TFA's research page, such as this one, you will find that in Memphis, Tennessee, merely 8.9% of TFA teachers are still on the job in year four. 

[...]

The claim that TFA teachers last an average of eight years in the classroom is simply astounding. If this can be substantiated, I will reappraise my critique of the program, though I will have to wonder why our experience in Oakland (and that in Memphis as well) was so different.

When I had a chance to ask Heather Harding, research director for TFA, some questions, she said 35% of TFA teachers return for a fourth year. This number is higher than we experienced in Oakland, and higher than the 8.9% we see in Memphis, Tennessee. But it is much less than we would expect if, as Wendy Kopp asserts, the average TFAer lasts eight years in the classroom.

Let's doodle on the back of the envelope a bit.

According to Gary Rubinstein, the attrition rate for TFA is around 11%, meaning one in ten TFAers do not even stay for their full two-year commitment. Is Kopp including them in her rosy numbers? Let's assume she is not, and we're just talking TFA alumni who met their initial two-year commitment.

There have been some conflicting reports about TFA's attrition; most of the confusion seems to come from the question of whether or not those who leave their initial TFA assignment continue teaching at a different school. It's worth pointing out that Kopp says her mission is to serve low-income students; if the TFAers who stay in education are moving to affluent school districts, that would seem to contradict her stated goals.

Given that caveat, what do we know about TFA teachers' attrition rates? This research brief on TFA states “all one can say with certainty is that in 2007, at least 16.6 percent of those recruited by Teach For America were teaching in a K-12 setting beyond their two-year commitment.” The same brief cites studies that show attrition rates as high as 85% in some cities; again, we don't know if the TFAers left their initial assignments to go to other teaching jobs.

But let's be optimistic: this study claims 40% of TFAers leave teaching after their initial two-year stint; another 25% leave after four years. What would that mean for Kopp's claim? Roughly, 40% of TFA's corps would have to stay on the job for 14 years, and another 25% would have to stay on for 12 years. That's really the only way the "average" length of a TFA education career can be eight years: a whole bunch of TFA alumni would have to stay much longer than the average to balance out those who leave immediately after their initial commitment is up.

Is this even possible?

TFA was founded in 1990; in that year, they placed 500 teachers. Here's a graphic from their website that shows the size of the TFA corps over the years:

There are 28,000 TFA alumni. Assume 40% left after their two-year commitment; that's 11,200 teachers. You need to balance this number with 11,200 teachers who taught an average of 14 years to come up with an average of 8 years for the entire corps.

You could, of course, have fewer long-serving teachers, if you have more teachers who taught longer than 14 years. But TFA only started in 1990; the longest possible career for a TFAer is 23 years. That first class had 500 members. The graph above clearly indicates that the class size did not grow substantially until after 2000; every year in the 1990's looks as if it has fewer than 1,000 teachers per class. But again, let's be wildly optimistic and say that there are 10,000 TFAers who could potentially have taught for 14 years. What does that mean?

Even if every TFAer who could have taught 14 years taught an average of 14 years, that still wouldn't be enough to balance the many TFAers who leave after their initial two-year assignment. Again, you need 11,200 teachers to teach an average of 14 years just to balance out those who left after two years.

And I'm not counting the number of teachers who leave after three or four or five or six or seven years; you need to balance them out as well with more teachers who taught longer than the eight-year average.

This is, on the surface, a very hard claim to swallow. It seems very unlikely that Kopp's claim is accurate, unless she is willing to refute the studies that show high rates of TFA members leaving the profession, or she can show TFA teachers who stay in the field have very long careers. Considering that Kopp herself says only one-third of TFA alumni have remained in teaching (including, I assume, those who are meeting their current two-year commitments), her claim about the length of TFA alumni careers looks very suspect.

TFA is now receiving substantial support from the federal government, so this is a matter of transparency and accountability to the taxpayers. Wendy, let's see the data.

ADDING: It's really remarkable how little we can say definitively about TFA and teacher attrition. And yet the feds gave them a nice chunk of change. Didn't anyone think to do a cost-benefit analysis before handing over the dough? Wouldn't some basic statistics about the effectiveness of the program go a long way toward guarding the interests of taxpayers?