Teach For America (TFA) is a nation-wide program where college graduates serve two-year terms as teachers or principals at underserved schools. The organization was founded in 1989 by Princeton graduate Wendy Kopp (“Research,” n.d.). Since then, over 50,000 people have participated in the program (“Who we are,” n.d.). In 2010, the nonprofit received a $50 million Investing in Innovation (i3) grant from the US Department of Education to expand the program (Clark, Isenberg, Liu, Makowsky, & Zukiewicz, 2015). TFA maintains a page on their website detailing (favorable) publications on the effects of the program; the organization claims that one study (Backes & Hansen, 2015) “found suggestive evidence that corps members had an impact on several non-tested outcomes. Students taught by corps members in elementary and middle school were less likely to miss school because of unexcused absences and suspensions than students taught by non-Teach For America teachers in the same school” (“Research,” n.d.). However, this conclusion is unwarranted given the actual, less substantial findings of the study.
Compared to the authors of the TFA site, the study’s authors are more honest in communicating the realistic significance of their findings. Based off of their analysis of data on non-tested indicators of student success, they conclude elementary school students of TFA instructors showed a reduction in unexcused absences and days of suspension. At the middle school level, students of TFA members likewise experienced declines in unexcused absences and days of suspension (Backes & Hansen, 2015, p. 20). While these findings appear promising, the researchers note that only the results of elementary days of suspension and middle-school unexcused absences are statistically significant, the former only marginally. Similarly, initial analysis finds middle school TFA teachers’ students’ unexcused absences to decrease by 7% compared to their non-TFA classmates, but taking into account the higher baseline of unexcused absences among the populations that TFA members are recruited to teach, the change reduces to only 4%. The largest difference in proportions between TFA and non-TFA students was in the change in days of suspension in elementary school students; however, Backes and Hansen (2015) mention that “it is hard to know whether this particular result is replicable in different data” (p. 21) due to a low baseline of days of suspension and a high standard error in the data. With the minor significance of this analysis’ findings, TFA’s announcement that the study yielded “suggestive evidence that corps members had an impact on several non-tested outcomes” (“Research,” n.d.) is misleading.
This analysis also found that elementary students of TFA instructors experienced a 1.7% increase in GPAs compared to their non-TFA-instructed peers. The TFA site does not include this in their summary of the findings, but given that the organization cites this study in its entirety, it is still worthwhile to evaluate the merit of this conclusion. The authors define the difference in GPAs as only “at least marginally significant” (Backes & Hansen, 2015, p. 20), and the effect was not present at the middle school level. While these considerations diminish this minor difference’s significance, it is also important to note that teachers are the ones who determine these GPAs; not all of a student’s teachers will be TFA-affiliated, of course, but this variable does allow for the possibility of grade inflation by TFA members.
In an effort to demonstrate quantifiable value in their program, Teach For America evaluated Backes and Hansen’s (2015) overly generously when summarizing it on their site. The incremental and potentially bias-influenced differences in TFA- and non-TFA-instructed elementary and middle school students’ unexcused absences, days of suspension, and GPAs in reality are not as encouraging as TFA claims.
References
Backes, B. & Hansen, M. (2015). Teach For America impact estimates on contested student outcomes. National Center for Analysis of Longitudinal Data in Education Research. Retrieved from https://caldercenter.org/sites/default/files/WP%20146.pdf
Clark, M. A., Isenberg, E., Liu, A. Y., Makowsky, L., & Zukiewicz, M. (2015). Impacts of the Teach For America Investing in Innovation scale-up. Mathematica Policy Research. Retrieved from https://teachforamerica.app.box.com/s/wyuu1rpqogxmksat86mnuk16sur981le
“Research.” (n.d.) Teach For America. Retrieved from https://www.teachforamerica.org/support-us/research
“Who we are.” (n.d.) Teach For America. Retrieved from https://www.teachforamerica.org/what-we-do/who-we-are
Data/Society/Decision-making Spring 2019 