No. Twenty-Three, August 1998
Laboratory for Student
Vision for Education (SERVE)
New federal class size reduction (CSR) proposals have further fueled a growing national interest in this approach to boosting student achievement. CSR is readily understood and appeals to common sense. Many see it as a means of sending dollars directly to the classroom rather than to the bureaucracy. It is enormously popular, a factor that makes it an attractive and politically viable policy option. At this writing, at least half the states and a number of school districts have enacted or are considering some form of CSR.
Though researchers continue to debate the issue, wide agreement exists that the critical question is not whether class size can make a difference in student achievement, but how and under what circumstances it does. Highly related are questions of CSRs costs. Do they outweigh its benefits? Are other alternatives more cost effective? Finally, for those designing CSR policies, especially on a large scale, many questions must be addressed about the implementation trade-offs of differing policy options and how these may affect student outcomes.
This brief addresses each of these concerns, drawing from the experiences of a number of states and districts with some track record implementing CSR. It starts from the position that class size reduction is not a silver bullet or an end in itself. Rather, CSR is one approach that has been shown effective in reaching the real goal: improved early learning. Success depends on getting the numbers down and on policies that support schools ability to take advantage of the opportunities CSR presents
1. Do small classes in and of themselves affect student learning?
"Yes" is the answer that emerged from Project STAR, the largest, longest-lasting, and most controlled study to date on class size. After decades of inconclusive research on small classes, STAR was able to show definitively what parents and teachers have long believed: that bringing class size down in the primary grades in and of itself has positive effects on student achievement in all subject areas (see box). STARs small classes had 1217 students, while the control "large" group had 2226. Children who gained most from smaller classes were minority students and those in inner-city schools. And the benefits lasted, at least through 7th grade. Several recent smaller studies generally support STARs findings, notably in terms of gains for urban minority students. (See #9.)
2.What conditions are critical to achieving the small-class effect?
In Project STAR, Tennessee schools were "laboratories" for class size research. Certain conditions prevailed without which the positive effects of small classes may not occur. Chief among those conditions are:
Adequate supply of good teachers. No organizational arrangement, including small class size, can compensate for poor teaching. In Tennessee, all STAR teachers were state certified and qualified to teach in their assigned grades. Even among the small classes, some teachers were more effective than others; researchers have yet to study what may have caused these differences.
Sufficient classroom space. STARs participating schools had no problem finding appropriate space to create enough classrooms for the reduction in numbers of students per teacher.
A representative student mix in each class. In STAR, the mixture of students in the class was determined at random and so mirrored the diversity in the school as a whole. Research has not revealed what would happen if, for example, 17 pupils with learning or behavior problems were assigned to a small class. In such a case, positive effects are less likely without the infusion of significant additional resources.
Teacher access to adequate materials and services. STAR teachers had no change in the materials and services normally available to them. Small- and regular-class teachers had access to reading specialists, school psychologists, special education programs (although there is evidence that the need for these services was reduced), and other schoolwide services. Small classes were not intended to serve as a substitute for other programs with demonstrated efficacy (including bilingual programs).
3.What do we know about why small classes are academically beneficial?
As classes shrink, other possibilities grow. Specifically, some studies have found that small classes allow teachers to spend more time on instruction and less on classroom management. One such study from Australia also suggested that smaller classes allow more and more protracted interaction between teachers and individual students. (1) Limited observations of 52 of STARs 2nd grade classrooms showed that teachers could better monitor student reading progress and were more consistent in managing behavior. (2) Likewise, observations of small and regular classes in North Carolina discovered more "on-task events" and fewer "institutional events" (e.g., disciplinary or organizational) in the small classes.(3)
In California, which began shrinking primary-grade classes in 1996, teachers surveyed in a preliminary study report that they are using more small-group instruction and better assessment techniques. (4) They also say they can cover the curriculum faster and in greater depth. The study also supports STAR findings that students in small classes are more motivated. In a STAR follow-up questionnaire, 4th grade teachers rated students from small classes much higher than their large-class counterparts on effort (e.g., pays attention in class; completes assignments; works well with other children) and initiative taking (e.g., does more than just the assigned work; asks questions to get more information). Small-class students, they said, were far less apt to be disruptive, passive, or withdrawn. (5) Given the high stability of behaviors such as these over the years, it may be that both the immediate and long-term benefits of small classes occur because students are better engaged.
4. How small is small enough?
No one knows what the optimal class size is. Many states and districts are currently shrinking classes to 20 or 18. (6) STAR researchers continue to analyze the question, but can only say at this point that the greater the class size beyond 17, the less the likelihood that the outcomes will be as positive. Earlier research suggested that the most dramatic gains accrue when class size shrinks to 15 or below. (7) Another unknown is the academic influence of the "drop factor," i.e., the magnitude of the drop in class size. For example, a drop from 30 to 18 is clearly more dramatic (and expensive) than a drop from 22 to 18. Whether it makes a greater difference in student learning is unclear.
5. Can small-class features be identified and used in large classes to create a small-class environment?
While some suggest that using grouping strategies in large classes can help create a small-class environment, Project STAR found that the defining feature of success is smallness itself. STAR analysts conclude that only smallness reduces the number of institutional events, creates an environment in which every student becomes engaged in learning, and allows the teacher to attend to every student. Not addressed by STAR are variations on reducing class size all day in all classes for example, placing students in small classes for part of the day for subjects such as reading and math while having larger groups for PE or art. (See #9.)
6. For how long do students need to be in small classes to gain the lasting benefits?
No one knows. The STAR research as well as a smaller study done in North Carolina (8) suggest that the main benefits occur in the first year a student is in a small class and are sustained or increase slightly after that. Economist Alan Krueger says a possible explanation is that attending a small class in the lower grades may confer a one-time "school socialization effect" that permanently raises the level of student achievement. (9) Others surmise that successive years of small classes may have helped sustain the gains. But researchers as yet cannot say whether one year of small classes may be just as effective as three or four and, if so, which age or grade level should be the focus. STAR data are currently being re-analyzed to answer these questions.
7. Is CSR worth the cost?
Whether CSR is cost-effective is a matter of much debate. One problem is that although cost ingredients can be fairly reliably calculated (see #8), quantifying benefits is more complex. Consider that the effects on reading may be different from those on math, for example, or that effects may differ from one student population to another.(10)
Some argue that in the long run, potential benefits may offset costs. In addition to across-the-board academic gains, some research suggests that small classes in the primary grades begin students on a path that reduces the need for special education, grade retentions or disciplinary measures and increases the likelihood of high school graduation. (11) Such outcomes translate into real savings.
To date, there are few well-controlled studies that compare one intervention with another. That makes it difficult for policymakers to weigh an investment in CSR against spending on alternatives such as peer tutoring, professional development, or computer-assisted instruction that have been shown effective (though not in large-scale, randomized experiments such as STAR).
The issue is further complicated by questions of social priorities. If CSR is deemed important for the next generation of students, how much is too much to spend? More specifically, how much academic improvement is enough to justify a given expenditure? And CSRs popularity adds yet another dimension to the debate: political viability. As Congress bogged down in partisan arguments last year over school reform and national testing, Californias "bold stroke" of CSR had the backing of educators, Democrats, Republicans, unions, and taxpayers. (12) For policymakers convinced of CSRs benefits but concerned about its costs, this widespread appeal may hold sway.
Economic analyses of CSRs worth may have confused the issue rather than providing answers. Economists have debated whether there is a relationship between class size and student learning. (Among them, Hanushek concludes there is no relationship; others, e.g., Hedges et al. and Wenglinsky, disagree. ) But in fact, such studies have looked at pupil-teacher ratio rather than actual class size, and the two are not the same.
Pupil-teacher ratio is the number of students in a school or district compared to the number of teaching professionals. In an urban district, which may employ a number of part-time professionals such as Title 1 and special education teachers or reading specialists, that ratio may be 15:1 or lower, though each regular class with its own teacher may contain 30 or more children. In STAR, other small-class research, and in state CSR initiatives discussed in this brief, "small classes" has meant the number of pupils actually in classrooms.
8. How much does CSR typically cost?
Calculating the cost of a statewide CSR program involves considering a number of ingredients, (14) including:
As an example, costs in California have played out as follows:
9. Are there ways to contain the costs of reducing class size?
Hiring more teachers and creating more classrooms is an expensive way to gain the benefits of smaller classes. Alternative ways of funding CSR can help contain the costs. These include:
Targeting the resources. The investment can be directed to schools that need it most for example, those serving poor and/or minority students. After studying whether CSRs benefits could be gained at a lower cost, economist Allan Odden recommended reducing class size for students achieving below grade level and combining individual tutoring with classes reduced to 15 students for language artsreading instruction. (15) He also proposed coupling small classes with a "larger, comprehensive set of strategies" shown to be effective for low-income, ethnic, and language minority students.
Wisconsins Student Achievement Guarantee in Education (SAGE) chose this tack. SAGE targets primary-grade children who live in poverty. It provides up to $2000 extra dollars per low-income student in participating classrooms at 30 schools. Besides maintaining a 15:1 ratio, SAGE requires its schools to implement a rigorous academic curriculum, provide before- and after-school activities, and implement professional development and accountability plans. A first-year evaluation showed SAGE students performing significantly better than a comparison group in reading, language arts, and math. (16) African-American males, in particular, seemed to benefit.
Allowing local flexibility in attaining smaller-class goals. Policies can set CSR goals as a means of improving achievement, then encourage local creativity in reaching those goals. Schools can combine new funding with a reallocation of existing funds as well as re-think schedules to devise an array of small-class arrangements. Approaches being tried include:
Redistributing resources. A number of districts or schools have re-examined all available resources with an eye on finding ways to apply funds to CSR. Title 1 funds have become one common means of reducing class sizes in high-poverty schools. Two examples from North Carolina are Oak Hill Elementary in Guilford County, which has reduced class sizes in K2 from 23 to 15, (17) and Hillcrest Elementary in Burke County, which in 1991 began using Title 1 funds to reduce grades 13 to 15 as part of a countywide CSR initiative.
Burke County has defrayed CSRs costs by using state dollars for full-time teacher assistants to fund regular teaching positions instead. State funds given to Burke as a "low-wealth" county have been applied to the CSR initiative as well. Schools in other parts of the state, such as Draper Elementary in Rockingham County, have adopted a whole-school approach to reduce classes to approximately 13 in grades 14. The K5 school had 23 teaching positions plus two teacher assistants, two "specialty" positions (PE, Spanish, music), and a Title 1 teacher. By eliminating categorization and re-defining roles, the school "found" five extra teacher positions to use for its CSR initiative.
Creative scheduling. Some school faculties have also devised alternative schedules to reduce class size for a portion of the day. Little research has been conducted on the impact on student learning, but such changes can potentially achieve some of the benefits of smaller classes. One variation is parallel block scheduling. (18) While half the class are taught critical subject areas such as reading and math, the rest attend specialty classes such as music, art, or computer lab in larger groups. Another variation is the Oak Park Plan, (19) which requires that all teachers in a school including specialists teach 15 students in core academic areas (reading, language arts, and math) for three hours a day. For the remaining 2.5 hours, subjects are taught in regular class sizes of approximately 25 students, and specialists provide services and consultation. (See also #14.)
Drafting CSR legislation or initiatives involves weighing a range of choices, trade-offs, and as yet unanswered questions. Factors such as scale of implementation, demographics, resources, student mobility and enrollment growth rates vary markedly from place to place, making it difficult to say that strategies effective in one locale will be equally so in another. Questions to address when designing CSR policies include:
10. Is the necessary infrastructure in place to support CSR?
The two key infrastructure pieces are teaching and facilities, and each has its own set of questions.
Teaching. Will there be enough qualified teachers for the number of new classrooms created? Are existing policies on emergency or alternative credentialing consistent with your goals? Will there be enough specially-trained teachers e.g., for limited-English proficient (LEP) or special education students? Will the policy create an incentive for teacher job shifts e.g., from special to general education; from substitute to permanent status; from preschool to primary-grade teaching? If so, what are the implications?
Teachers have reported that a switch to smaller classes finally allowed them to do what they know works. But if CSR leads to the hiring of many inexperienced or unprepared teachers, those teachers will require support (e.g., mentoring; modeling) to learn and use effective classroom strategies. Moreover, in a policy climate in which CSR initiatives are accompanied by raised standards, new assessments, and/or stringent accountability measures, even veteran teachers may need more knowledge and skills, not just smaller classes, to meet increased expectations.
In California, no one knows whether the hiring of thousands of inexperienced and uncredentialed teachers will alter intended outcomes for CSR (see box). In some urban districts, schools with the highest concentrations of limited-English proficient students also have the largest numbers of teachers hired for CSR on emergency credentials. They have no special training and, often, no bilingual aides. (20)
Facilities. Will existing facilities accommodate the number of new classrooms created by the new policy? Is enrollment growth a factor? What shifts may occur due to space crunches? Will they affect other programs?
California schools have purchased portables when possible, but many have also usurped space from other programs (see box 1), reconfigured schools (e.g., moving 6th graders to middle schools), switched to year-round scheduling, changed school boundaries, remodeled schools, canceled inter- and intra-district transfers, and/or re-opened schools previously closed. (21) In Nevada, rapid growth and lack of facility funding have resulted in many large, team-taught classes (22) (see matrix and #15). In Utah, space problems have blocked schools in crowded districts from reaching their targeted reduction levels (23) (see matrix).
11. Should CSR be used in conjunction with other strategies?
Some research suggests that comprehensive planning can make a difference in the effectiveness of CSR. In a study in Austin, Texas, for example, achievement and attendance remained extremely low at 13 of 15 low-performing schools, while the other two showed dramatic gains. (24) Those two combined CSR with other changes such as new curricula and teaching methods focused on individual attention, increased parent involvement, and health services.
In Utah, which began CSR in 1990, a study found smaller classes most effective in districts that focused on improving achievement rather than just getting the numbers down. Successful schools combined CSR with teacher development, instructional improvement, and productive use of personnel and resources. (25)
In Nevada, where CSR began in 1989, new legislation prompted by school districts allows districts to opt either to expand existing primary-grade CSR to 3rd grade (at a 19:1 ratio) or after approval from the state superintendent to use their share of CSR funding to implement comprehensive programs such as Reading Recovery or Success for All, which have been shown effective in improving reading and math achievement in grades 13.
12. Will CSR be optional or mandatory?
An optional CSR program may, de facto, leave school and district leaders with little choice. In California, two realities have made rapid implementation all but mandatory: most districts are loath to turn away money after decades of successive cuts, and CSRs intense popularity and press coverage created enormous pressure to move fast and implement fully.
13. Will the funding be flat or wealth-adjusted?
It can be argued that an approach allotting the same amount of money for every student is regressive. For example, Californias program allots $800 of incentive money for every student in a 20:1 primary class.Given the states diversity, this "one size fits all policy," combined with public pressure to implement and lack of flexibility, has raised the following equity issues: (26)
One upshot is that in California, students most likely to benefit from smaller classes minority and inner-city children may be those least likely to have full opportunity to do so.
Formula-based funding can help offset inequities. Utah uses a formula that initially allotted 80 percent of the state K6 CSR funding on a per-pupil basis, with 20 percent reserved for districts with rapid growth but an insufficient tax base. (By design, the reserve pot is shrinking; eventually, all funding will be allotted per pupil.)
Decisions here have significant implications for cost, teaching quality, facilities, and other logistical considerations. Options include capping the number of students per teacher, specifying an average across a school or district, or specifying differing levels of reductions (e.g., greater reductions in high-poverty schools).
Besides urging a district average to lower program costs, Californias Legislative Analyst advocates allowing districts to hire the number of teachers they would need to implement 20:1, but then also allowing flexibility in how the district deploys those teachers. For example, teachers could be assigned to do one-on-one or small-group tutoring to supplement classroom instruction.
Utah does not stipulate a cap. District officials have discretion over allocations to schools, and school staffs have leeway in how they achieve smaller classes. For example, at some schools, half the students attend from 8:002:00, the other half from 10:004:00. The split schedule allows each group a two-hour time block in classes of 1820 for subjects such as reading and math. PE and social studies are taught in some cases team-taught in classes of 40.
Actual smaller classes are qualitatively different from mere changes in the pupil-teacher ratio. (See #7.) In Nevada, where facilities are funded entirely with local rather than state money, state CSR policy allows the 16:1 ratio in 1st and 2nd grades to be achieved by having two teachers in a classroom with 32 students. In Spring 1998, some 36 percent (down from 40 percent the prior year) of Nevadas "reduced size" 1st and 2nd grade classes contained 32 children with two teachers. (27) Though a limited evaluation study showed small gains over eight years of CSR, many worry that the team-taught classes may be limiting success.
Attention to class size is a timely and appropriate focus for education policy. Class size reduction has enormous intuitive and political appeal, bolstered by research that shows smaller classes are good for learning as well as behavior. But designing a successful CSR policy is no simple matter. As knowledge from state and local experiences continues to evolve, some lessons emerging include:
Policy Intent, Elements
Level of Implementation
Policy Intent, Elements
Policy Intent, Elements
Level of Implementation
Policy Intent, Elements
Level of Implementation
1 S. Bourke, "How Smaller Is Better: Some Relationships Between Class Size, Teaching Practices and Student Achievement," American Educational Research Journal, 23: 558571, 1986.
2 Carolyn M. Evertson and John K. Folger, Small Class, Large Class: What Do Teachers Do Differently?, paper presented at the annual meeting of the American Educational Research Association, San Francisco, March 1989.
3 Charles M. Achilles, K. Kiser-Kling, A. Aust, and J. Owen, Success Starts Small: Life in a Small Class. (Final Report, Small-Grant School-Based Research), Greensboro, NC: University of North Carolina, 1994.
4 Edward Wexler, JoAnn Izu, Lisa Carlos, Bruce Fuller and Mike Kirst, Californias Class Size Reduction: Implications for Equity Practice and Implementation, San Francisco and Berkeley: WestEd and Policy Analysis for California Education (PACE), 1998.
5 Jeremy D. Finn, John K. Folger, and Deborah Cox, "Measuring Participation among Elementary Grade Students," Educational and Psychological Measurement, 51: 393402, 1991.
6 Charles M. Achilles, If Not Before, At Least Now, paper presented at the annual meeting of the American Educational Research Association, San Diego, April 1998.
7 Gene V. Glass, Leonard S. Cahen, Mary Lee Smith, and Nikola N. Filby, School Class Size: Research and Policy, Beverly Hills, CA: Sage Publications, 1982.
8 Paula Egelson, Patrick Harman, and Charles M. Achilles, Does Class Size Make a Difference?, Greensboro, NC: SouthEastern Regional Vision for Education, 1996.
9 Alan B. Krueger, Experimental Estimates of Education Production Functions: Princeton: Princeton University Department of Economics, 1998.
10 Jeremy D. Finn, Class Size and Students at Risk: What Is Known? What Is Next?, Washington, DC: U.S. Department of Education, Office of Educational Research and Improvement, National Institute on the Education of At-Risk Students, 1998.
11 Jeremy D. Finn and Charles M. Achilles, Tennessees Class-Size Study: Questions Answered, Questions Posed paper presented at the Annual Meeting of the American Educational Research Association, San Diego, April 1998.
Finn, Class Size and Students at Risk.
12 Thomas Toch and Betsy Streisand, "Does Class Size Matter?," U.S. News & World Report, October 13, 1997.
13 L. V. Hedges, R. D. Laine, and R. Greenwald, "Does Money Matter? A Meta-analysis of Studies of the Effects of Differential School Inputs on Student Outcomes," Educational Researcher, 23: 514, 1994.
Harold Wenglinsky, When Money Matters, Princeton:Policy Information Center, Educational Testing Service, 1997.
14 This section draws substantially from Joel Schwartz, Class Size Reduction (Policy Brief), Sacramento, CA: Legislative Analysts Office, February 12, 1997.
15 Allan Odden, "Class Size and Student Achievement: Research-based Policy Alternatives," Educational Evaluation and Policy Analysis, 12 (2): 213227, 1990.
16 Peter Maier, Alex Molnar, Stephen Percy, Phillip Smith, and John Zaborik, First Year Results of the Student Achievement Guarantee in Education Program (Executive Summary), Milwaukee: University of Wisconsin, Center for Urban Initiatives and Research, December 1997.
17 Achilles et al., Success Starts Small.
18 R. L. Canady, "Parallel Block Scheduling: A Better Way to Organize a School, Principal, 69(3), 1990.
19 E. Mueller, The Oak Park Plan: The Fourth R, paper presented at the annual meeting of the American Association of School Admini- strators, Dallas, 1985.
20 Wexler et al. Californias Class Size Reduction.
21 Joan McRobbie (ed.), WestEd Class Size Reduction Updates (electronic publications), December 1996 through AprilMay 1998.
22 H. Pepper Sturm, Nevadas Class-Size Reduction Program, Carson City, NV: Nevada Legislative Counsel Bureau, 1997.
23 Karen Evans-Stout, Nancy Fleming, Bob L. Johnson, Jr., Edna Ehleringer, Joyce M. Gray, Randall Merrill, Margery Parket, Robyn Roberts, and Tracy Stewart, The Use of Class Size Reduction Funds in Five Wasatch Front Districts (A Study for the Utah Education Consortium), Salt Lake City, 1997.
24 Richard J. Murnane and Frank Levy, "Evidence from Fifteen Schools in Austin, Texas," in Does Money Matter, Washington, DC: Brookings Institution Press, 1996.
25 Evans-Stout et al., The Use of Class Size Reduction Funds in Five Wasatch County Districts.
26 Joan McRobbie, "Californias Class Size Reduction: A One-Year Status Check," Thrust for Educational Leadership, Association of California School Administrators, September 1997.
27 Letter from H. Pepper Sturm, April 20, 1998.
Joan McRobbie is California Liaison at WestEd where she analyzes education issues, especially as they evolve in California. jmcrobb@WestEd.org; 415-565-3069.
Jeremy D. Finn is a Visiting Scholar at LSS, Temple University, in the Center for Research in Human Development and Education. His regular position is Professor of Education at the State University of New York at Buffalo. He was external evaluator to Project STAR throughout its duration. firstname.lastname@example.org; 716-645-2482.
Patrick Harman, an evaluation specialist with SERVE, focuses on evaluating school reform efforts. His current area of interest is developing appropriate methodologies for investigating educational initiatives. email@example.com; 800-755-3277.
The authors gratefully acknowledge the contributions of Charles M. Achilles, Professor, Eastern Michigan University; H. Pepper Sturm, Chief Principal Research Analyst, Nevada Legislative Counsel Bureau; and Laurie Chivers, Deputy Superintendent of Public Instruction, Utah.
WestEd, LSS and SERVE are part of a network of 10 Regional Education Laboratories that work to ensure that those involved in education improvement at the local, state, and regional levels have access to the best available information from research and practice. This publication is funded in part by the U.S. Department of Education, Office of Educational Research and Improvement (OERI). Its contents do not necessarily reflect the views or policies of the Department of Education.
To learn more, contact:
730 Harrison St.
San Francisco, CA 94107-1242
Laboratory for Student Success
Temple University/Center for Research in Human Development and Education
933 Ritter Annex
13th St and Cecil B. Moore Avenue Philadelphia, PA 19122
SouthEastern Regional Vision for Education (SERVE)
P.O. Box 5367
Greensboro, NC 27435