Email Status

Volume 28, Number 3
May/June 2012

Class Size Revisited—Again

Limited research spurs experimentation with staffing


In a world filled with terms like “value-added” and “adequate yearly progress,” many parents still rely on a far simpler data point to assess a school’s quality: the number of students in the classroom. They feel, strongly, that their child stands to get more personal attention in a class of 17 students than in one of 27 and that smaller classes are less likely to get out of hand. According to a 2007 poll in the journal Education Next, more than three-quarters of parents would rather shrink classes than pay teachers more.

Smaller classes are wildly popular among teachers, too, for similar reasons. They say that with fewer students, they can control them better, give more personal attention, and assign and grade more complex work. Nadia Zananiri, who has about 35 students in her Advanced Placement world history class in Miami Beach, Fla., says, “Something is going to have to give . . . You can only humanly grade so many essays. This is a writing-heavy class, so the options are to assign fewer assignments or give less feedback.”

Yet with only limited research bolstering the case for class-size reduction, legislative mandates are fizzling out, and the topic has been all but dropped from public debates about how to improve schools.

Budget pressures are partly behind the shift. Shrinking classes means hiring more teachers, and teacher personnel costs already typically constitute 60 percent of a school’s budget. Wisconsin adopted a class-size program for primary grades in the 1990s, but districts have had trouble coming up with matching funds to support it, and caps have been raised from 15 to 18 students. In Georgia, where districts couldn’t afford to hire more teachers or reallocate building space, the state board of education last year gave districts flexibility to override a mandate that capped elementary classes at 23 and high school classes at 32. In Texas, which adopted a class-size law in the 1980s, one in five school districts was given permission this year to sidestep the 22-student restriction in elementary schools.

The last gasp for statewide programs could well be in Florida. Voters there amended the state constitution in 2002 to limit core academic classes to 18 students in preK–3, 22 in grades 4–8, and 25 in high schools, a move that has cost the state more than $21 billion so far. Beginning in 2010, the state began levying financial penalties on districts for each student over the cap.

Ballot initiatives to ease the amendment’s requirements have failed, but in 2011 legislators reduced the number of classes counted as part of the core, exempting, for example, foreign language and Advanced Placement classes from the restrictions.

Nationally, the movement to reduce class size is losing proponents, who now say that what matters more than the size of the class is the quality of the teacher in front of it. Secretary of Education Arne Duncan, for example, has acknowledged what research shows: smaller classes under specific conditions do seem to benefit some young students. Yet he said last year that “class size is a sacred cow, and we need to take it on.”

Critics argue that to improve teacher effectiveness, it makes more sense to invest in differential pay or training. But the field lacks specific analyses comparing the cost-effectiveness and benefits of smaller classes with other initiatives.

“Class-size reduction in the early elementary grades is a costly reform that’s only viable when you have qualified teachers and adequate space for the additional classrooms,” says Adam Gamoran, a professor of sociology and education policy at the University of Wisconsin–Madison. “I’d like to see studies of class size that were experimental, that occurred in a varied set of contexts, and that paid careful attention to cost, so they could be compared to the cost of other interventions.”

While they aren’t engaged in empirical research, some innovators are reframing the either-or question: Are there ways to respond to parents’ and teachers’ wishes about class size without breaking the bank?

Does Class Size Matter?
When people say that research shows that class size matters, they are nearly always referring to the Student/Teacher Achievement Ratio study (Project STAR), based on research conducted in Tennessee from 1985 to 1989 (see sidebar “Research on Class Size”).

Project STAR is the only randomized experiment in class-size research—meaning that, similar to a medical drug trial, students were separated into treatment and control groups.

Project STAR compared K–3 students taught in classes capped at 17 with those in classes capped at 25. On average, children in smaller classes performed better on math and reading tests, an edge equivalent to about 15 percent of a year of learning. Studies using other types of research methods have generally found more mixed effects on student scores. And few state programs have yet to demonstrate results on par with Project STAR, primarily because they could not replicate the research conditions.

Researchers have also had a hard time explaining what exactly went on in the smaller Tennessee classes to cause the gains. Studies have, in general, not found that teachers in the smaller classes used significantly different teaching practices. Perhaps, it’s thought, students were simply more engaged in the smaller classes because teachers were able to develop closer relationships with them.

That analysis of teacher practice, researchers say, represents a crucial gap in class-size research. Beth Graue, a professor of curriculum and instruction at the University of Wisconsin–Madison, has found that implementation of her state’s class-size initiative differs vastly, depending on teacher beliefs and the quality of resources available to support smaller classes. To meet the parameters of the program, Graue found, some Madison schools simply put two classes of 15 students taught by separate teachers in the same classroom; others assigned two teachers to a room of 30, where one teacher tended to focus on clerical duties, while the other teacher was instructing. Smaller classes often made their work more manageable, but the teachers struggled to identify and use different pedagogical approaches that took advantage of the smaller group size.

Graue concluded that class size is better thought of as a condition that can take many forms rather than as a single intervention. “You have to try to figure out why you want to have smaller classes, in what context, for whom, and then what’s going to happen with instruction that’s going to be different,” Graue says. Instead, she says, policymakers simply cut the numbers and “assumed there’d be a magical transformation in the classroom.”
New Experiments
The flexible thinking that Graue advocates is at the center of a handful of small, local experiments with class size. Rather than reducing class sizes across the board, these initiatives create smaller groups for certain teachers, subjects, or situations and pair them with teacher training.

The Generation Schools Network, a Denver-based nonprofit, is launching several middle/high schools this fall with a novel schedule that shrinks the size of core classes without hiring many new teachers. English classes will have about 20 students and math classes 25, compared with as many as 40 in some local high schools.

Teachers’ time will be staggered in a kind of relay. In the morning, core-content teachers teach a double block of small classes. In the afternoon, a different group of instructors takes over electives, which generally are comparable in size to those in other schools. While they do so, the first group of teachers gets common planning time so that they can devise more engaging instructional activities to fill the longer core classes.

Furman Brown, Generation Schools’ founder and director, believes the investment in smaller class sizes is worth it, but only when combined with major shifts in teaching. “We’ve said teachers need more time with their kids,” Brown says. “You can’t expect to change the class size and in just 35 and 40 minutes expect things to change.”

Another group of schools is trying a model that attempts more personal instruction without actually relying on smaller classes across the board. At the nine schools run by Rocketship Education, a California charter chain, an extended school day permits each student to have two additional hours a day of instruction. Regular classes are about the same size as they are in other local public schools, says Rocketship CEO John Danner, but smaller groups are used to help struggling students during extended learning time, either through one-on-one practice using online curricula or through tutoring of four to five students at a time by paraprofessionals.

Savings are invested in teacher professional development. Analyses of results from the charter group’s quarterly achievement tests, Danner says, show achievement on the rise. Danner believes that’s a result of increased attention on each child’s specific areas of weakness and the increased support for teachers.

This spring, five other sites selected by the Bill & Melinda Gates and Carnegie foundations and the consulting firm Public Impact will be piloting additional ideas that play with class size and instructional time. Public Impact codirector Bryan Hassel suggests that teachers who are especially effective at increasing student learning take on a few additional students—not to exceed the size of classes in top-performing nations, generally no more than 35 students—in exchange for extra compensation. Schools would then reserve smaller classes for novice and struggling teachers.

“Teacher quality varies so much; the quality of the teacher giving you that personal attention is still really important,” Hassel says. “As long as we have the rigid idea that the overall class size everywhere has to be 21 students, you can’t test those ideas.”

Not all the ideas are likely to fly with teachers or their unions. While Generation Schools have won plaudits from both the New York City school district and its union for its focus on professional development, teacher unions have been wary of how the Rocketship approach reduces the number of teacher positions. They’ve also been among the fiercest critics of attempts to weaken the Florida mandate.

Debate about such new experiments can seem disingenuous to teachers. For instance, Zananiri, the Florida teacher, suspects that some students not prepared for Advanced Placement were put into such classes simply to avoid class-size caps that no longer apply to honors courses. She had as many as 50 students early in the school year and contends that administrators addressed the issue only after repeated complaints.

“That was like a dream, having a class size of 25,” Zananiri says of her one year under the class-size cap. “Teachers love having a manageable size.”

Stephen Sawchuk is an assistant editor for Education Week who covers the teaching profession and writes the “Teacher Beat” blog.

Also by this Author

    For Further Information

    For Further Information

    G. M. Bohrnstedt and B. M. Stecher, eds. What We Have Learned About Class Size Reduction in California. Sacramento, CA: California Department of Education, 2002. Available online at

    R. G. Ehrenberg, D. J. Brewer, A. Gamoran, and J. D. Willms. “Class Size and Student Achievement.” Psychological Science in the Public Interest 2, no. 1 (2001): 1–30.

    J. D. Finn and C. M. Achilles. “Tennessee’s Class Size Study: Findings, Implications, Misconceptions.” Educational Evaluation and Policy Analysis 21, no. 2 (1999): 97–109.

    E. Graue, K. Hatch, K. Rao, and D. Oen. “The Wisdom of Class Size Reduction.” American Educational Research Journal 44, no. 3 (2007): 670–700.

    C. Milesi and A. Gamoran. “Effects of Class Size and Instruction on Kindergarten Achievement.” Educational Evaluation and Policy Analysis 28, no. 4 (2006): 287–311.

    A. Molnar, P. Smith, and J. Zahorik. “Evaluating the SAGE Program: A Pilot Program in Targeted Pupil-Teacher Reduction in Wis­consin.” Educational Evaluation and Policy Analysis 21, no. 2 (1999): 165–177.