Is TFA scalable?

TIME with five TFA myths. Note: The writer is clearly biased by his friends and co-worker's ties to the organization. He only admits this late in the article.

I agree that four of them are myths. TFA teachers are, on average, better than the alternative (mostly full-time substitutes or uncertified teachers). TFA math teacher are better than average teachers too. That is the most important one.

But number four, that TFA isn't scalable, may well be a myth. The writer even seems to grant this at the end of his discussion. I did a calculation once, however, that suggested TFA was effect despite its out of control costs. Here's how:

Let's say we want to scale TFA so that it's teaches can teach the 20% of American students who live in poverty. We have 3 million teachers so we'd need to train 3*0.2 = 0.6 million = 600,000 teachers. Say each teacher teaches for 20 years on average (a mix of short-term busts and long-term commits). That means we have to train 600,000/20 = 30,000 new teacher each year to keep a steady supply of teachers.

At 30,000 teachers * $20,000 / year in training costs * 3 years of training = $1.8 billion. That's less than half of what Race to the Top cost. But it wouldn't come close to erasing the achievement gap. TFA teachers only increase scores by (I think) 3 percentile points. The gap is about 35 percentiles. So TFA would have to become vastly more effective before it even makes sense to talk about erasing the achievement gap. That said, if they could do it at the cost by unit of achievement gap that they are doing now, it would cost just 4% of the defense budget or $21 billion per year.

TFA is scalable in the sense that it is not unreasonably expensive and there are probably enough quality teachers to fill the ranks. It isn't a scalable in the sense that it can erase the achievement gap just by expanding.