Class Size vs. School Vouchers for Minority Student Achievement

Introduction

The continuous achievement gap between African-American students and their white peers is a major problem in American education. The gap in fourth-grade reading on the National Assessment of Educational Progress (NAEP) diminished in the 1970s, but since the early 1980s it has remained largely unchanged.

Evidence from many studies suggests that African-American students may benefit more than other students from improvements in educational quality. For example, class size effects have been greatest for African-American children.

This paper focuses on the effects of two programs — voucher programs and programs that reduce class size — on the achievement of African-American children, and on related studies covering both minority and non-minority students. Evidence for the effects of both programs is presented. In many studies across many locations, both voucher programs and class size reduction have been found to significantly increase student performance. The policy implications of these findings are discussed to determine whether class size reduction is a better alternative than voucher programs for increasing student performance.

The impact of class size versus voucher programs on student achievement has been a widely debated issue for years. Educational vouchers were proposed by Milton Friedman in the 1950s as a means of improving the quality of elementary and secondary education (Molnar, 1998). In recent years, the use of vouchers has resurfaced, particularly as a method to help minority students. Vouchers provide public money for students to attend private schools.

Several experimental voucher programs have been introduced in the past decade. One of the oldest and largest is a pilot program begun in 1990 in Milwaukee, Wisconsin. Initial analyses of this data suggested little or no improvement in the quality of education as the result of a voucher system (Molnar, 1998). These results, however, are quite controversial and are the subject of considerable debate. More recent pilot programs in New York City and Cleveland have provided additional evidence for this debate. Preliminary, and equally controversial, results from the New York Choice Scholarship Program suggest that a voucher system has produced modest improvements in test scores for low-income students who transfer to private schools as a result of a scholarship program.

Advocates of smaller classes argue that students learn better in smaller classes where the teacher has more time for each child. However, until recent years there was little conclusive evidence that smaller classes are more effective, so the idea was set aside in favor of other programs, such as the voucher program. As a result, many public school students — particularly those in inner-city schools — have been crowded into classes of up to 40 students.

Recently, the issue of small class size attracted a great deal of media and political attention as numerous studies presented clear evidence that smaller class size really does work (Taylor, 2002). In the early grades, all children learn better in classes of 15 to 17 students, and the improvements are even greater for poor children in inner-city schools. According to recent studies, the following evidence demonstrates the need for smaller class size:

Long-lasting effects: According to Harvard economist Frederick Mosteller, STAR — a Tennessee class-size study — is "one of the great experiments in education." STAR researchers analyzed the achievement of K–3 students who were randomly assigned to classes of 13 to 17 students, demonstrating that students in small classes performed much better than students in regular classes in math and reading, every year and in all grades. The small classes had the greatest impact on the scores of children in inner-city schools. Researchers who continued to follow students who participated in the experiment discovered that the benefits of small classes in the early years last at least into high school, long after students return to regular-size classes.

Half a year difference in fourth-grade math: Harold Wenglinsky, a researcher who analyzed fourth- and eighth-grade math scores on the 1992 National Assessment of Educational Progress, found that "fourth-graders in smaller-than-average classes are about half a year ahead of fourth-graders in larger-than-average classes," with the "largest effects" for low-income students in urban areas. The effects are even stronger for urban eighth-graders.

Voucher advocates argue, however, that voucher programs are a better way to improve student performance (Taylor, 2002). Princeton professor Cecilia Rouse (1998) conducted a thorough study of performance in Milwaukee elementary schools — both public schools and those accepting vouchers. She found that public school students in certain special schools where the pupil-teacher ratio was reduced to 17 to 1 progressed as well as voucher students in math and made "substantially faster" progress in reading than voucher students. These were low-income, minority children — the same profile as students in the voucher schools.

It has been noted that good schools — whether public or private — have much in common. Successful schools usually share a combination of the following characteristics: high academic standards and a challenging curriculum for all children; a safe and orderly environment; qualified teachers; and parent involvement (McGarvey, 2001). Recently, researchers have added another ingredient to this list — small class size — supporting a long-held theory that small class size makes a significant difference.

According to Rouse (1998), school leaders and other stakeholders should move beyond the empty "public vs. private" debate and examine the school-level factors that actually explain student achievement. In a 1998 paper, she explores the effects of class size by comparing the achievement of Milwaukee voucher students with students in three types of Milwaukee Public Schools (MPS): regular schools, magnet schools, and schools participating in the Preschool to Grade 5 Grant Program (P-5 schools). P-5 schools, which enroll approximately 25% of all MPS elementary students, serve "predominantly minority and extremely disadvantaged" children and receive additional state funds that have enabled them to lower their pupil-teacher ratio, on average, to 17 to 1.

Literature Review

Rouse's report found the following:

Students in the P-5 (small class size) public schools demonstrated "substantially faster gains in reading" than those in the regular public schools, the public magnet schools, and the voucher schools. Students in the P-5 schools also made faster math gains than students in the regular public schools and the public magnet schools, and achieved the same math gains as students in the voucher schools. Even though average class size in the P-5 schools was larger than in the voucher schools — 17:1 versus 15.3:1 — P-5 schools outperformed voucher schools in reading and were equal in math.

These findings suggest that small class size has a positive effect on student achievement, and many other researchers support these conclusions (McGarvey, 2001). For instance, Word et al. (1990) revealed that Tennessee K–3 students in small classes substantially outperformed students in larger classes in both math and reading every year, at all grade levels, and across all geographic areas. A follow-up study (Nye et al., 1995) demonstrated that these benefits lasted through at least eighth grade, with achievement advantages particularly large for minority students.

The Wisconsin SAGE evaluation (1998) showed that Wisconsin Student Achievement Guarantee in Education (SAGE) students "enjoyed significantly greater improvements in test scores in reading, language arts, and math" than their non-SAGE peers, with the greatest gains for African-American boys (McGarvey, 2001). Results so far "are consistent with the Tennessee experience."

In a 1997 study of more than 200 school districts, Harold Wenglinsky revealed that increased teacher-student ratios (smaller class size) significantly raise average math achievement: "Fourth graders in smaller-than-average classes are about a half a year ahead of fourth graders in larger-than-average classes," with the "largest effects" — three-fourths of a year ahead — for low-income students in urban areas.

As researcher Alex Molnar (1998) stated in a recent summary of voucher and class size research: "No strong evidence exists that participation in a voucher program increases student achievement." On the other hand, "There is no longer any argument about whether or not reducing class size in the primary grades increases student achievement. The research evidence is quite clear: It does." In this light, it appears that small class size results in greater academic achievement than vouchers.

Princeton University researcher Cecilia Rouse, whose findings are often cited by voucher supporters, conducted a study in 1998 comparing Milwaukee's voucher schools with the city's P-5 schools — public schools with small class sizes and additional targeted funding (Pathak et al., 2004). "The results suggest," Rouse stated, "that students in P-5 schools have math test score gains similar to those in the [voucher] schools, and that students in the P-5 schools outperform students in the [voucher] schools in reading." Rouse added: "Given that the pupil-teacher ratios in the P-5 and [voucher] schools are significantly smaller than those in the other public schools, one potential explanation for these results is that students perform well in schools with smaller class sizes [emphasis in original]." These findings suggest that improved test scores for some voucher students may have been the result of smaller class sizes.

Princeton University researchers Alan Krueger and Diane Whitmore (2001) compared the effect of attending a smaller class to the effect of receiving a private-school voucher, basing the comparison on data from African-American voucher students. They found that students who had attended small classes "improved their test performance by around 50% more than the gain experienced by black students who attended a private school as a result of receiving a voucher."

According to Pathak et al. (2004): "The African-American voucher students were learning in the very educational climate that many policy analysts have long sought for public schools — a climate that is incredibly difficult to create when a state diverts substantial tax dollars to vouchers."

One of the most publicized reports on this topic is "Test-Score Effects of School Vouchers in Dayton, Ohio; New York City; and Washington, D.C.: Evidence From Randomized Field Trials," developed by a team of researchers including Paul Peterson of Harvard University (Molnar and Achilles, 2002). Shortly after the voucher report was published, former U.S. Secretary of Labor Robert Reich wrote an article in The Wall Street Journal calling for "progressive vouchers." Referring to the latest voucher study, Reich stated: "Evidence mounts that vouchers do work for kids who use them."

Reich's statement was supported by little evidence. Nearly all of the research on the achievement impact of publicly funded voucher programs is drawn from the Milwaukee Parental Choice Program and the Cleveland Scholarship and Tutoring Program. While the Cleveland evaluation is ongoing, the Milwaukee data are now several years old. Additionally, the 1995 legislation that expanded the Milwaukee program to include religious schools also removed its evaluation component, so no new data will be available in the near future. Thus, most of the evidence about the impact of voucher programs comes from privately funded voucher programs, including those in Dayton, New York City, and Washington.

In Reich's report, the Peterson team presented results from Dayton, New York, and Washington averaged across grade levels, along with combined averaged results from all three cities. Because averaged results tend to conceal inconsistent findings, they may make the reported achievement impact appear more generalized than it actually is.

Even the averaged results show inconsistency. For example, according to the Peterson team, merely being offered a voucher makes a notable difference in reading — but not math — achievement for African-Americans in New York City. In Dayton, the offer produces no significant result. In Washington, the offer makes a difference in African-American test scores in both reading and math. Actually switching to a voucher school in New York results in a reading advantage for African-American students; in Washington, there is an advantage in reading but not math; and in Dayton, switching to a voucher school resulted in no achievement advantage at all.

According to a five-year Indiana University study, students in the Cleveland voucher program performed at the same level as their public school peers (Metcalf, 2004). After tracking the test scores of 6,000 students from kindergarten through fourth grade, the researchers found no difference in performance. The results were statistically adjusted to account for the lower test scores of low-income and minority children, because the voucher group had a greater proportion of white and affluent children. "After you adjust for minority status, there's no difference," said the study's chief author, Kim Metcalf.

The study also reported high parent satisfaction with the program and a desire among poor parents to give their children the same opportunities that more affluent families have. "It seems to reflect a desire to help their children develop ways of interacting, behaving and thinking that will help them to be successful, productive and self-sufficient adults," Metcalf (2004) said.

Marc Egan of the National School Boards Association argued that voucher proponents failed to deliver on their promises of providing a better education by moving children out of public schools. "They're nothing but a diversion from really improving public schools," he said (Metcalf, 2004).

Students who entered the voucher program from public schools had the same income and ethnic characteristics as the overall public school population. However, private school students who received scholarships were more likely to be white and affluent. The program may do well at offering scholarships, the report stated, "but it may not be completely effective in attracting and retaining students from the very low-income, African-American families that it originally targeted."

Vouchers have been described as a sideshow distracting from the real issues of education reform: academic standards, accountability for student performance, meaningful public school choice (including charter schools), teacher quality, and equitable and sufficient investment (Rotherham, 2002). Voucher initiatives essentially continue a debate that allows both established educational interests and voucher proponents to avoid the thornier questions of school reform by focusing on a simplified campaign of slogans.

Aside from the basic methodological questions that threaten the validity of many voucher findings, the literature on vouchers and academic achievement amounts to an inconsistent body of results that offers little guidance to either parents or policymakers. As Jay P. Greene once stated in Education Week (Molnar and Achilles, 2002): "Interest groups have learned that they can successfully check research contrary to their goals by producing their own studies, no matter how lousy, to sow confusion among policymakers, journalists, and the attentive public about what to believe."

Despite the preliminary character and inconsistent outcomes of voucher research, voucher advocates claim that the achievement evidence is clear, systematic, and compelling. Some compare the power of this research and the magnitude of the voucher findings to the results of class-size research — most frequently the Tennessee Student-Teacher Achievement Ratio (STAR) experiment.

The STAR project, which took place in Tennessee from 1985 to 1989, was a longitudinal, randomized experiment involving 11,600 students in grades K–3. STAR aimed to determine the effect of reducing class size in grades K–3 on student achievement and social development. STAR students were monitored throughout their K–12 education and have continued to be tracked in their postsecondary pursuits. As such, STAR serves as a large-scale, longitudinal analysis of the impact of reducing class size in the primary grades.

A variety of studies have been conducted based on the initial STAR findings. Krueger and Whitmore (1999) found a relationship between participation in a STAR small class and the rates at which students took college-entrance exams. Follow-up studies of STAR students in grades 4–12 revealed that they outperformed students who had been in normal-size classes on tests in all subject areas. Students enrolled in small classes from kindergarten to grade 3 outperformed regular-class students on all tests every year, and the gap between small- and regular-class test results grew each year. For example, an average STAR student who was in a small class for grades K–3 would outperform a regular-class student at grade 4 by 6.6 months, and by 8.7 months at grade 8 — five years after leaving the small-class environment.

Positive voucher academic effects are sometimes compared to the total STAR effects rather than to the STAR effects for minority students specifically. For voucher studies, in which the majority of students are African-American, the relevant comparison is to the performance of STAR minority students. For those students, the STAR effects were nearly double the overall effects. Furthermore, it appears that voucher students who perform at a higher level than their public school peers usually attend smaller classes in smaller schools.

A RAND Corporation study of American education, Improving Student Achievement, found strong associations between student achievement and both higher public pre-kindergarten participation and reduced pupil-teacher ratios in the lower grades.

Some voucher advocates argue that no one has demonstrated that vouchers harm public schools. However, this argument is insufficient. Sound public policy should be constructed using the most accurate data available so that public schools can provide the best possible education to the students they serve.

Conclusion

The body of research reviewed in this paper consistently demonstrates that class size reduction produces stronger and more reliable academic gains than school voucher programs, particularly for minority students. Evidence from the Tennessee STAR experiment, Wisconsin's SAGE program, Milwaukee's P-5 public schools, and direct comparative analyses by researchers such as Krueger and Whitmore all point in the same direction: students in smaller classes — especially low-income and African-American students — outperform their peers in larger classes, and they outperform voucher recipients as well. Voucher research, by contrast, remains methodologically contested, geographically inconsistent, and limited in scope. Policymakers seeking to close the achievement gap would be better served by directing resources toward systematic class size reduction in the early grades than by expanding voucher initiatives whose benefits remain unproven.