Teach For America has been the subject of much debate since its inception in regards to one question: Are TFA teachers, entering the field of education with only five weeks of teaching experience, better or worse teachers than other novices? More specifically, what type of impact do corps members have on their students' educational outcomes?

Three key studies have set out to answer this question. The most sweeping study looked at students and TFA corps members nationwide (Mathematica xiv) while the other two focused on students and teachers in Houston, Tex. (Darling-Hammond et al., CREDO 11).

The first national study

The national study, performed by Mathematica Policy Research, found that:

Corps member success in Houston: a study

Another study was conducted by the Houston Independent School District (HISD) in conjunction with a Stanford University think-tank. The Houston school district has hired between 20 and 40 fourth through eighth grade Teach For America teachers each year between 1996 and 2000 (CREDO 29).

This study found similar results: "On average, the impact of having a TFA teacher was always positive," and "produced significantly better student outcomes than non-TFA teachers" (CREDO 12). In a finding that may not please TFA's leadership, however, the CREDO study found that "the differences between the average TFA teacher and the average non-TFA teacher, while always positive, are not statistically significant" (CREDO 12).

Darling-Hammond, et al. Study

The only study to conclude that Teach For America teachers have a negative impact and are not as well prepared was released in 2004 by a Stanford University professor. It examined student test scores during the beginning and end of each year for students who were in classes taught by uncertified (Teach For America) teachers compared to certified (non-TFA) teachers (Darling-Hammond et al. 14). Only fourth and fifth grade teachers in Houston were examined, according to Teach For America's response.

On one type of standardized test result, the study found that Teach For America teachers "had a significant negative effect on student scores in math (Darling-Hammond et al. 19-20). It also found that corps members had "significantly negative [results] in both reading and math."

The Stanford study concluded that certified teachers who had a traditional education and were certified before teaching were more effective than the corps members.

Teach For America responded strongly to this study, finding fault with its lack of peer review, small sample sizes and large-scale conclusions. It says that this study did not compare the effectiveness of TFA teachers to those who are not in each school. Three other independent studies found fault with the methods used by the Stanford researchers. They claim that the study draws conclusions that are too broad based on a very small study.

Conclusions

It is quite clear that Teach For America teachers are answering a critical need in our country - they are providing their services to cities that need teachers the most.

The first two studies show that Teach For America teachers are indeed strong teachers despite their lack of long-term training, although significant gains in student achievement can't be expected. While their students often perform slightly better in math, their gains in reading are not significant. The last study even concludes that TFA teachers' students actually perform worse on reading and math tests.

But both positive studies agree: The program supplies the nation's most in-need school districts with high-quality teachers.

"Teach For America is a cost-effective way to attract teachers to low-income schools," concludes the Mathematica study (Mathematica 64). It concludes that schools can expect to increase their students' math scores and keep their students' reading scores the same with very little additional costs or risks.

In other words, the data show that there is "little risk" in hiring TFA teachers because they have been shown to keep students on the same levels, at the very least.

This site was created by Andy Lewis for MMC 3260