Voucher Threat Helps to Raise Test Scores in Florida: Choice sanctions under No Child Left Behind have little effect

Contact: Paul E. Peterson
Phone: 617-495-8312
Contact Organization: http://www.ksg.harvard.edu/pepg/
Date: April 05, 2005

Cambridge MA — In the first independent study to examine the impact of the No Child Left Behind Act (NCLB) on the test-score performance of individual students, Harvard researchers Martin R. West and Paul E. Peterson at the Program on Education Policy and Governance (PEPG) at the Kennedy School of Government found that key provisions of the Florida A+ Accountability Plan were more effective than NCLB’s at leveraging student achievement gains.
West and Peterson’s findings, presented at the annual conference of the Royal Economic Society in Nottingham, England on March 22, show that Florida’s fourth and fifth grade students make modest but significant gains in reading and math if their school is at risk of becoming a part of the state’s voucher program. Florida schools at risk of being subject to the public-school choice provisions of NCLB showed no gains.
However, the report also stresses that the research findings pertain solely to the school choice provisions as they operate in Florida (and not necessarily how NCLB choice provisions are operating more generally), and they should not be taken as a measure of the overall effectiveness of either NCLB or the Florida A+ Plan. Since 2002, average test scores of all fourth and fifth graders in Florida have risen significantly.
Under the Florida A+ Plan, students qualify for a voucher to attend private school if their public school is given an ‘F’ twice in any four-year period. Students in schools that received their first ‘F’ grade in the summer of 2002 scored 4 to 5 percent of a standard deviation higher the following year than did students at comparable ‘D’ schools not subject to the voucher threat. Though the one-year effect was relatively small, the gain would become quite large were it to continue at the same rate for the three years the schools remained threatened by vouchers.
Most of the students attending ‘F’ schools came from disadvantaged backgrounds. Significant gains were observed for African Americans, those eligible for free lunch, and those who had the lowest initial test scores. Significant gains were not observed for white students, Hispanic students, those not eligible for free lunch, and those who had higher test scores.
West and Peterson also noted that ‘D’ schools seem to respond positively to the stigma of receiving a very low grade. Schools receiving a ‘D’ grade improved by 5 percent of a standard deviation relative to comparable ‘C’ schools. The gains registered by ‘F’ schools were made on top of these improvements.
“Only 8 percent of all schools in the state were given a ‘D,’ so the stigma was readily apparent,” said West. “It appears that the schools are very capable of moving forward when faced with a clear challenge—either receiving a very low grade or when faced with a voucher threat.”
No similar gains were observed at schools threatened by NCLB’s public-school choice provisions. In Florida’s Title I schools, students may move to another public school within the district, if their school is designated as not making adequate yearly progress for two consecutive years. Students at schools under this threat in the summer of 2003 did no better the following school year than students at similar schools not subject to the threat.
Peterson points to two factors that help explain why NCLB’s choice provisions, as they are applied in Florida, have had so little impact.
“Nearly 75 percent of Florida’s schools did not meet adequate yearly progress goals in 2003,” he said. “Sky-high failure rates can undermine the accountability threat. When everyone is criticized, no one is going to take the criticism seriously.”
Additionally, Peterson noted, NCLB’s public-choice provisions are not strong enough to compel much of a response from low-performing schools. According to U. S. Department of Education data, less than one percent of students change from one public school to another when given the opportunity under NCLB.
“That does not seem to be enough of a challenge to spur local school improvement,” Peterson said.
West and Peterson’s study draws upon the individual test-score records of 900,000 elementary school students in Florida for the years 2002-2004. Impacts of accountability programs are estimated after adjusting for prior test scores and demographic background of students and the social composition of their peer group.
The study is based on data from the Department of Education, State of Florida, which has made individual test-score information available to qualified scholars who have signed confidentiality agreements protecting the privacy of individual students.
Paul E. Peterson is the Director of the Program on Education Policy and Governance at Harvard University . Martin G. West is a PEPG Research Associate. The paper is available at http://www.ksg.harvard.edu/pepg/.
Antonio M. Wendland TEL (617) 495-7976
Program on Education Policy and Governance (PEPG)
J. F. Kennedy School of Government, Harvard University, 79 J. F. Kennedy Street, Taubman 304, Cambridge MA 02138
FAX (617) 496-4428


John F. Kennedy School of Government 79 John F. Kennedy Street
Cambridge, MA 02138
617-495-1100 Get Directions Visit Contact Page