According to Teach For America, the average teacher in this country is pretty bad. The main deficiency of the average teacher, they say, is that the average teacher in this country does not believe in kids as much as Teach For America teachers do and therefore have lower expectations which then become a self-fulfilling prophecy.

There have been a lot of studies about TFA over the past 25 years. Generally they use end of the year state test scores and compare TFA teachers to other teachers who have the same amount of teaching experience. Back in 2004 Mathematica did a study which concluded that TFA teachers were slightly more effective at teaching math and slightly less effective at teaching reading compared to comparable non-TFA teachers.

Most of the studies have what might be a critical form of bias called ‘survivorship bias.’ What this means is that these studies don’t include the teachers who quit midyear before their students get to take those assessments. And, yes, there are also teachers in the comparison group who quit, but for the studies I have looked at, I haven’t seen any try to adjust for this. In some TFA regions, over 50% of the first year teachers don’t make it through their first year. I have to think that those teachers caused a lot of damage to their classes (and to themselves) before they quit but that damage is not factored into the studies, I believe.

Recently a large ‘meta-analysis’ was released in which the American Institutes for Research (AIR) somehow put together the results of 23 different studies over 24 years and came up with some conclusions based on this averaging. According to the TFA press release which had the title ‘Meta-Analysis Shows Teach For America Teachers Have Consistent and Significant Impact on Student Learning’, the results were positive.

I clicked on the link to the study and found that there were more conclusions than that. One thing about these studies, I’ve noticed, is that sometimes they average the results of the first year corps members, the second year corps members, and the alumni. So in the early days there were a lot more corps members than there were alumni teachers. But now there are more alumni teachers, I would expect, than there are corps members. So it is hard to know exactly how much to weigh the different groups. On average, the amount of time that a typical TFA teacher stays in the classroom is probably about 2.5 years (I’m just guessing, but it is what I think). Whatever the actual number is, it would be most accurate in judging TFA to weigh the alumni based on that statistic.

But fortunately this ‘meta-analysis’ does do a breakdown of TFA corps members (at least the ones who did not quit during their first year) and alumni. Here are the two charts that summarize their conclusions:

As you can see, the alumni bring up the results a lot. If those alumni are disproportionately represents in the averaging, it would skew the results.

So looking at these result and trying to be ‘data driven’ what are the important things that jump out?

Well the first thing is that before the pandemic, TFA training was mostly in person but during it they went remote and then when the pandemic ended the TFA training became a hybrid model with a lot of training being done online. So this could explain the big -0.13 in the postpandemic row of the math chart and the -0.04 in the postpandemic row of the ELA chart. If TFA considers this study to be authentic enough to spread on social media, they should take these postpandemic results seriously and restore the in person training. (Oh, and you can’t say that everyone did worse postpandemic because these negative numbers are in comparison to the non-TFA teachers with similar experience.)

Other things to notice, in general, even including the alumni data, elementary results are just slightly positive meaning that the TFA teachers are about the same as the non-TFA teachers with the same number of year in elementary schools. I’ve always had concerns about TFA placing teachers in elementary schools. At least with a middle or high school a teacher who is really struggling can only do so much damage in the one subject they are teaching. But with the responsibility of being the only teacher in an elementary setting, it is probably too much for a short summer training. In my opinion, TFA should have two tracks: A two year track for people in secondary schools and a three year track for people who want to do elementary where the first year they are teachers-in-training. Seeing the results of this ‘meta-analysis’ would support this idea.

Also I think it is fair to mention that in the four different group (corps member ELA, alumni ELA, corps member math, alumni math) the corps member ELA is actually a negative -0.01. Yes that’s the smallest negative you can have so it is basically 0, but I’m sure if all four numbers were positive, TFA would be eager to point that out so I’m pointing out that one of the 4 is actually negative.

So I don’t think this study is something for TFA to really write home about. Because TFA only highlighted the numbers where the alumni are averaged in (possibly disproportionally) with the corps members, two of the three citations in Elisa Villanueva-Beard’s post have “perform better” rather than “perform similarly” though it would be more accurate that say that corps members do not “perform better” than other non-TFA novices in any of the categories according to the conclusions of the study. And if, as TFA, would want you to think, the average teacher isn’t very good, then the average TFA corps member must not be very good either.