The Case for Smaller Classes

Following is an interview with Alex Molnar, director of the Center for Education Research, Analysis, and Innovation at the University of Wisconsin-Milwaukee and one of the principal investigators of Wisconsin’s SAGE class size reduction program. Molnar was interviewed by Barbara Miner of Rethinking Schools.

Question: You are nationally known as a leading proponent of reducing class sizes. Why do you think this issue is so important?

Answer: Small classes clearly increase student achievement in the primary grades. We have 15 years of research evidence that illustrates this, from Tennessee’s Student Teacher Achievement Ratio (STAR) study to Wisconsin’s evaluation of its Student Achievement Guarantee in Education (SAGE) program.

The evidence is very, very clear. In the primary grades, when students are in classes of around 15 students per teacher, they are going to gain two to three months in academic achievement over their peers in larger classes.

Can you briefly explain the Tennessee STAR study?

The Tennessee STAR study is one of the largest educational experiments in this century in the United States. And I use the word “experiment” in a scientific sense, in that the STAR study featured random assignment of students and teachers and carefully controlled experimental conditions.

The study began in 1985 with a group of kindergarten students who were in small classes through third grade. Between 8,000 and 10,000 students participated in STAR.

Those students have now entered college. Researchers have followed the students through the grades and have been able to document that their achievement gains in the primary grades continued throughout high school. Furthermore, they took the SAT and ACT tests at significantly higher rates than students who were in the larger classes.

Can you explain the SAGE program in Wisconsin?

The heart of the SAGE program is that it calls for class sizes of 15 students for every teacher in kindergarten through third grade. The program is funded by the state and targets schools that enroll the largest number of children eligible for free or reduced lunch.

The program began as a pilot in the fall of 1995, with 30 schools in 21 districts. The first year, it affected only kindergarten and first grade. In subsequent years, it was expanded to include second and then third grade.

The initial program proved so popular that the legislature and governor agreed to expand it to 79 schools in 1998. This year funding for the program was increased further, and it is anticipated that there will be somewhere between 500 and 600 SAGE schools in the fall of 2000, including virtually every elementary school in Milwaukee.

Under SAGE, the state provides $2,000 in funding for every child in the program in kindergarten through third grade. The program clearly benefits urban areas; but from its inception, SAGE has been a statewide program and has included low-income rural and suburban schools.

What have been the results so far of the SAGE program?

There have been three annual evaluations of SAGE to this point. So far, SAGE is showing very promising results. The bottom line is that SAGE students are doing better academically.

In the recent third-year report, for example, SAGE students performed statistically higher across all grade levels in comparison to a control group of students in non-SAGE classrooms.

Does SAGE make a difference in terms of the achievement gap between African-American students and whites?

One of the most encouraging results of SAGE is the effect on that achievement gap. In the SAGE evaluation, we found evidence that African-American students narrowed the gap in terms of their achievement with whites.

African-American students in both the SAGE schools and the comparison schools started off with their achievement at a significantly lower rate than white students. In the SAGE schools, African-American students tended to narrow the academic achievement gap with white students. In the comparison schools, the African-American students fell further behind.

This corresponds to the Tennessee findings and the results of follow-up studies of students who attended small classes in the STAR experiment. Consider, for example, the ACT and SAT test-taking statistics on these students. Both white and African-American students who took part in the smaller classes in the primary grades had significantly higher rates of taking the SAT and ACT. But the rate of increase for African Americans was substantially higher than that for whites. There was a closing of the test-taking gap in that respect.

In a time of tight budgets and tax cuts, how was it possible for SAGE to expand so dramatically?

The SAGE program is obviously a bipartisan success and has broad support across the political spectrum. The expansion of the SAGE program in 1997 and again in 1999 would not have happened without bipartisan support. In 1999 Democratic Senate majority leader Chuck Chvala made the expansion of the SAGE program one of his top legislative priorities, and he hammered out a budget agreement that won the support of Governor Tommy Thompson.

I don’t think there is any question that some of the people who helped start SAGE initially saw it as a powerful educational program that would outperform vouchers, in terms of promoting the achievement of poor children. On the other hand, SAGE was also supported by a lot of people who also support vouchers.

Let me give you an example. I don’t believe the SAGE program would exist except for the efforts of three Republicans – senators Tim Weeden of the Beloit area, Brian Rude from the La Crosse area, and David Prosser, who in 1995 was the speaker of the assembly. You may recall, that when SAGE first passed in 1995, the Republicans controlled the Assembly, the Senate, and the governor’s office. So the support of these Republican legislators was absolutely key.

On the other side, some voucher supporters, Howard Fuller, for example, who was then superintendent of the Milwaukee Public Schools, actively opposed the SAGE legislation.

Some have argued that in order for small classes to work best, there need to be more schools. Otherwise, especially in already overcrowded schools in Milwaukee, there will be 30 kids in a class with two teachers.

In an ideal world, I would prefer to have facilities designed for 15 students and one teacher in a classroom. And I think most parents would prefer that as well.

One of the potentially most important findings in the SAGE evaluations thus far, however, is that we have been unable to distinguish the academic performance of classrooms that have two teachers and 30 students from those that have one teacher and 15 students. If this finding continues to be robust, it obviously has important implications about the cost of reducing class size.

California instituted a statewide program of class size reduction. But there has been criticism that there was a shortage of teachers, and that good teachers left for the suburbs. Can you speak to some of those criticisms?

I’ve had a number of opportunities to talk to the people involved in the California evaluation, as well as in the California state budget office. There’s no doubt in anybody’s mind that the California program was affected by the way California approached class size reduction – with this enormous reduction all at once, without a start-up year to plan an evaluation, and without phasing in the program, as we did in Wisconsin.

There were two results of the California approach, in which the program was implemented almost overnight. One is that it placed enormous stress on the system with regard to the need for new teachers and the quality of teachers. Second, the evaluation was planned on the fly so it was very difficult to get the kind of quality data that we have from Tennessee and now from Wisconsin. The genie was out of the bottle before the evaluation was in place.

In Wisconsin, not only do a school and a school district elect to participate in SAGE, the program expanded in a measured way over five years. In California they tried to do it statewide all at once. School districts in Wisconsin have had time to plan the changes that reducing class size requires. Further, the K-12 education system overall has had time to respond to the various challenges that come up when you reduce class size.

With regard to California, it is important not to paint too bleak a picture. Some of the data on the achievement impact of the program are beginning to mirror the Tennessee and SAGE results. Also, with regard to the issue of teacher quality, it is not quite accurate to say that experienced teachers are fleeing the urban areas in California. Experienced teachers are staying in urban areas. But new teachers, the higher quality ones – at least those who have characteristics or a profile that you would think might produce a higher quality teacher – tend to go to the suburban areas.

This is for a lot of reasons. One is that the budgets and central bureaucracies in urban districts have been subject to budget-cutting for 20 years – to the point where they don’t have the capacity, the human resources, to identify people, to screen people, to review their backgrounds, and to hire the top people. As a result, the urban districts have to take whomever they get.

Suburban districts, on the other hand, are getting out into the labor market with enough resources in their central administrations to take a look at these potential new teachers and take the cream of the crop.

Some people have argued that improving teacher quality is much more important than reducing class size.

It’s not either/or. One of the things we know from Wisconsin and Tennessee is that good teachers and less-good teachers are distributed throughout the system. Yet unerringly, the SAGE classrooms turn out to be the top performers. And that suggests that the proper question is not, “Is it good teachers or is it small classes?” What I would say is, all things being equal, small classes out-perform average size classes. The extent to which having a high-quality teacher in the classroom amplifies that gain is unknown. We don’t know the answer to that.

What if the state says it is going to put $30 million in education reform. Would it be better to put it into small classes or improved teacher training?

Small classes. There’s no question about it. You can get a greater return more quickly in terms of student achievement by reducing classes.

I’m not against teacher training – I work in a school of education. The problem is, what do people mean when they talk about the achievement impact of having teachers with this or that set of characteristics – and where are the data to support the contentions? Teaching is a very complex, context-related activity. Further, some of the most profound effects of good teaching may not be identifiable using annual, standardized test data. Improving teacher quality is a very complex, long-term undertaking.

I sometimes feel that, at least on the part of some people, the argument about teacher quality represents more a political tactic to blunt the movement to reduce class size, than a commitment to improving teacher quality.

SAGE is often used as a shorthand for small classes. But can you talk about the other components of the program?

SAGE has its origins in State Superintendent John Benson’s Urban Initiative Task Force. Task force members felt that it wasn’t enough to reduce class size. The underlying understanding, which supports the task force recommendations that later became SAGE, was that there are an awful lot of children in our society who don’t have sustained, positive contact with adults. Schools can’t do everything. But what schools can do is reduce class sizes sufficiently so that the teachers have the greatest possible impact on a child’s life.

What are the other components of SAGE, beyond small classes in kindergarten through third grade? That there be a rigorous academic curriculum. That a school have extended hours both before and after school. And that there be a program of professional development and accountability. The expectation is that teachers in SAGE schools will be trained to teach effectively in small classes.

The definition of a rigorous academic curriculum was left undefined in the legislation. The reason is that the legislature wanted the SAGE program to be voluntary and subject to as much control by the local school, the community, and the teachers, as possible. The local school is obliged to tell the state Department of Public Instruction what its definition of a rigorous curriculum is, and to report to the department at the end of the year how it implemented that curriculum. Each school must also describe how it is using the required extended hours and what professional development activities its teachers have been engaged in.

Wisconsin is home to two popular reforms: smaller classes, and vouchers. In terms of documenting academic achievement, are there any comparisons?

I have conducted extensive reviews on the achievement impact of educational voucher programs and class size reduction programs. The data on the achievement impact of smaller class is is much more powerful overall than the data on the achievement impact of educational vouchers.

Based on my reading of the recent Legislative Audit Bureau report, it is no longer possible to say anything meaningful about achievement of voucher students in Milwaukee. This is because under the 1995 voucher law, there’s no possibility to gather systematic data. The voucher law, for example, doesn’t require that schools collect achievement data or report it if they do. With few exceptions, every voucher school is able to provide information that it wishes, and not provide information that it doesn’t wish to. In part, this is because when the voucher program was expanded to include religious schools, the legislature removed the evaluation component.

The current voucher law doesn’t provide an opportunity for researchers to get the kind of evidence that would be necessary to draw conclusions about the performance of the students in the voucher schools.

Quite apart from how one might feel about the wisdom of vouchers as public policy, it’s surprising to me that there isn’t a broad consensus that says we should know how well this program is working in terms of academic achievement – or how well it is reaching the families it is supposed to serve and what the impact of the program is on the students who attend the Milwaukee Public Schools. The answers to questions like these could help us understand what kind of return Wisconsin taxpayers are getting as a result of their investment in educational vouchers.

In contrast, Wisconsin’s SAGE program is held up nationally as a model of good legislative policymaking. It does a number of things well. It is voluntary. Individual schools have freedom to design their own programs. It holds schools accountable by requiring annual performance contracts. It effectively reaches the students it is intended to serve. And it has an independent evaluation that provides data for legislators and policymakers.

What we have in Wisconsin, at least in my own judgment, is a parody of the opening lines of a Tale of Two Cities: We have the best of policymaking and the worst of policymaking.