Study

Randomized controlled trial of 1,960 families of low-income public school students in grades K-4, selected through the lottery described above.  Families were randomly assigned to a group that was offered the vouchers or to a control group that was not.

Results after three years:

76% of those families offered vouchers opted to use them for at least one year.  What follows are the outcomes for all students offered vouchers, including both those who used them and those who did not, compared to the control group.

• For the overall sample, vouchers had no significant impact on their math or reading test scores.
• For the subgroup of African-American students, vouchers may have had a small positive impact on their combined math/reading test scores (the voucher group scored in the 27th national percentile, versus the 22nd percentile for controls).  However, the size and statistical significance of the effect depends on the definition of “African-American” and other factors.  Also, such subgroup findings sometimes appear by chance; further studies are needed to determine whether this effect holds up.

Discussion of study quality  (click here for a glossary of terms)

• This was a large study with a three-year follow-up.
• The study measured outcomes using an intention-to-treat analysis.
• The study assessed reading and math achievement using an objective, well-established test (the Iowa Test of Basic Skills).
• The study evaluated a voucher program that is fairly typical of voucher programs for low-income public school students, thus bolstering the policy-relevance of the findings.
• Prior to the intervention, the voucher and control groups did not differ in the preponderance of characteristics.
• Study Limitation: The study had moderate attrition (test scores were obtained for 67% of students at the three-year follow-up).  Statistical tests suggest that the attrition did not cause any observable differences between the voucher and control groups that might undermine the randomization.  However, the study results should be treated with caution because the attrition may have resulted in unobservable differences between the two groups, leading to inaccurate estimates of vouchers’ impact.

Sources (click on linked authors' names for contact information)

Mayer, Daniel P., Paul E. Peterson, David E. Myers, Christina Clark Tuttle, William G. Howell.  “School Choice in New York City After Three Years: An Evaluation of the School Choice Scholarships Program.”  Mathematica; February 19, 2002.  Click here for a link to this study.

Myers, David E. and Daniel P. Mayer.  “Comments on ‘Another Look at the New York City School Voucher Experiment.’”  Mathematica; April 1, 2003.  Click here for a link to this study.

Krueger, Alan and Pei Zhu.  "Another Look at the New York City School Voucher Experiment."  American Behavioral Scientist, vol 47, no. 5, April 2003, pp 658-699.  Click here for a link to this study.