Reducing class size to increase student achievement is an approach that has been tried, debated, and analyzed for several decades. The premise seems logical: with fewer students to teach, teachers can coax better performance from each of them. But what does the research show?
Some researchers have not found a connection between smaller classes and higher student achievement, but most of the research shows that when class size reduction programs are well-designed and implemented in the primary grades (K-3), student achievement rises as class size drops.
We identified 19 studies that met our standards. Most of these addressed reduced class size programs in grades K-3. Indeed, most programs in the past 20 years have targeted these early grades, in part because earlier research (see, for example, Glass and Smith 1978), suggested that these are the optimal years for such programs, and in part because of the more recent and comprehensive evidence from Tennessee's influential Project STAR (Student/Teacher Achievement Ratio). (For more information on Project STAR, visit http://www.heros-inc.org/star.htm.)
From this review of the research, we can scientifically document several important findings about reduced class size, which local school districts may find useful:
- Smaller classes in the early grades (K-3) can boost student academic achievement;
- A class size of no more than 18 students per teacher is required to produce the greatest benefits;
- A program spanning grades K-3 will produce more benefits than a program that reaches students in only one or two of the primary grades;
- Minority and low-income students show even greater gains when placed in small classes in the primary grades;
- The experience and preparation of teachers is a critical factor in the success or failure of class size reduction programs;
- Reducing class size will have little effect without enough classrooms and well-qualified teachers; and
- Supports, such as professional development for teachers and a rigorous curriculum, enhance the effect of reduced class size on academic achievement.
The following sections describe several reduced class size programs and examine the evidence.
Tennessee's Project STAR—commissioned by the state legislature in the mid-1980s—may be the most influential class reduction program in recent years. Project STAR found substantial evidence that reducing class size improved student academic achievement. Project STAR:
- Was designed as a large-scale, longitudinal (four-year) study;
- Involved nearly 80 schools from 42 school districts;
- Included about 7,000 K-3 students from families ranging from very poor to very affluent;
- Required schools to commit to the four-year time frame as a condition of participation; and
- Required schools to agree to random assignment of teachers and students to small (13 to 17 students) or large (22 to 26 students) classes as a condition of participation.
Class size was verified through site visits. Project STAR's impact on academic achievement was measured using a scientifically valid research design.
A decade later, the Wisconsin State Department of Public Instruction initiated SAGE (Student Achievement Guarantee in Education), a program intended to increase student achievement by reducing K-3 class size to no more than 15 students per teacher. Phased in over five years, beginning in 1996-97, SAGE targeted high-poverty schools and districts. (Visit http://dpi.wi.gov/sage/ for more information.)
California initiated its Class Size Reduction (CSR) program. Phased in over four years, its purpose was to decrease the population of K-3 classrooms to no more than 20 students per the states initiating reduced size programs, the dimensions of California's program are unequalled in terms of the numbers of students, teachers, and schools affected. (Visit http://www.cde.ca.gov/ls/cs/ for more information.)
Implementation of programs
Class-size reduction programs in grades K-3 have been implemented in different ways. Some programs focus on student-teacher ratio, while others restructure classes in some way. Still others emphasize professional development and related policies.
The student-teacher ratio in the programs studied varied widely. Some set a ratio of 13:1 as their benchmark for "small" classes, which were compared to "large" classes with a ratio of 20:1 (or higher). Others considered classes with a ratio of 20:1 to be "small." In other words, one program's "small" class could be another program's "large" class, making comparisons even more difficult.
- California, Tennessee, and Wisconsin set different thresholds for student-teacher ratios:
- California's CSR program defined a small class as one with a ratio of 20:1
- Tennessee's Project STAR defined a small class as one with a ratio in the 13:1-17:1 range
- Wisconsin's SAGE defined a small class as one with a ratio of 15:1.
Small classes may vary in other ways, in addition to student-teacher ratio, depending upon school leadership, facilities, and teaching staff, as well as official policy. The typical model is one teacher in one classroom teaching an assigned number of students, but other models have been implemented.
Project STAR defined two categories of reduced size classes:
- Classes of 13 to 17 students taught by one teacher; and
- Classes of 22-26 taught by a teacher and a teacher's aide.
Project STAR defined large classes as 22-26 students taught by one teacher.
Wisconsin's SAGE had four categories of classrooms, all intended to achieve the mandated 15:1 ratio (Molnar, 1999):
- Regular classroom (15 or fewer students with one teacher);
- Team-teachers classroom (30 or fewer students in a classroom with two teachers);
- Shared-space classroom (a regular classroom divided by a temporary wall with one teacher and 15 students on either side of the divider); and
- Floating teacher classroom (30 students with one teacher plus an additional teacher during reading, language arts, and mathematics).
Professional development and other related policies
Wisconsin's SAGE (Molnar, 1999) was unusual in that it incorporated into its class size intervention other initiatives designed to promote student academic achievement:
- A rigorous academic curriculum
- A staff development program
- Accountability measures.
In contrast, Project STAR focused strictly on class size reduction, and did not provide special training or professional development to teachers or aides (Finn, 2002). Studies of other interventions concluded that more attention should have been paid to the training and experience of teachers, and to the professional and curriculum development that might have helped them make the most of smaller classes (see, for example, Kurecka and Claus, 2000 and Munoz, 2001).
Gains in student achievement
The landmark 1978 study by Glass and Smith strongly endorsed reduced class size as a reform likely to produce improvements in academic achievement. The researchers reviewed 80 research reports on the relationship between class size and achievement, obtaining more than 100 comparisons from "well-documented" studies of smaller and larger classes using rigorous statistical analyses. The meta-analysis showed:
- As class size decreases, achievement increases
- Benefits begin to emerge as class size falls below 20 students.
The most influential contemporary evidence that smaller classes lead to improved achievement is Tennessee's Project STAR. Because this program set up randomly selected control and experimental groups of students, researchers could compare students who had four years of small class participation to students who had none. This meant that researchers could more reliably evaluate the impact of the class size reform. Project STAR (Finn, 2002) found:
- Students in smaller classes did better than those in larger classes throughout the K-3 grades;
- Minority and inner-city children gained the most from smaller classes; and
- The more years spent in reduced classes, the longer lasting the benefits.
Evaluation of Wisconsin's SAGE found similar evidence to support the policy of reducing class size in the early grades. Molnar et al (1999) found that participation in any one of the four types of SAGE classrooms was a predictor for student achievement. Controlling for pretest scores, family income, school attendance, and race/ethnicity, they found that:
- Students in grades 1-3 in SAGE classrooms scored significantly higher on the Comprehensive Test of Basic Skills in reading, language arts, mathematics, and in total scores, than students in traditional classrooms.
- Although first grade SAGE students had lower pretest scores than students in larger classrooms, the SAGE students had significantly higher post-test scores, indicating that SAGE students made greater gains than those in larger classes—and subsequent analyses showed they maintained these gains through second and third grades.
- Gains for African American students were even greater than those for white students.
Mitchell and Mitchell (1999) assessed the impact of California's CSR program after two years of implementation. They analyzed test data from the Stanford Achievement Tests 9th edition reading, language, and mathematics subtests as well as more than 30 variables related to student demographics, schools, classrooms, and teachers from more than 80 schools across eight school districts in Southern California.
In the Mitchell and Mitchell study, many demographic variables—such as gender, student poverty, ethnicity, and language spoken at home—were related to lower student gains. Nonetheless, when controlling for these variables, students in reduced size classes made small, positive gains in student achievement. But, the researchers acknowledged, other factors could have contributed to these gains. Major changes in California's public education system took place in tandem with the first phase of the CSR program, including new requirements for teacher preparation, revisions to bilingual education, new curriculum frameworks and materials, new statewide tests, and a new performance accountability system—none of which were taken into account by this study and any of which could have contributed to the achievement gains.
Gains over time
Nye et al (2001a) explored the relationship between the number of years that students participated in Project STAR small classes and their level of achievement. After one year, the students in smaller classes had significantly higher achievement scores on the Stanford Achievement Test reading and mathematics subtests than students in larger classes. The gap in scores widened after two years, indicating that the effects of small classes are cumulative.
Nye et al (2001b) also conducted a follow-up study of Project STAR students, which showed the positive effects of small classes maintained over time. The researchers compared the math achievement of 9th grade students who had been in small classes for at least one year during grades K-3 with that of 9th grade students who had been in larger classes in the third grade (and in earlier grades, depending on the year in which they were assigned to the study). In general, students who participated in small classes for at least one year continued to show higher scores on standardized mathematics tests at grade 9. Minority students in small classes had greater gains in achievement than white students in small classes, and girls in small classes had larger gains than boys in small classes.
Fidler (2001) looked at the impact of smaller classes over time within the California CSR program. The study examined the Stanford Achievement Test reading, language, and mathematics subtest scores of students in grades 4 to 6 who had completed at least one year of the CSR program in the Los Angeles Unified School District. The analysis found that the longer students participated in CSR, the greater their achievement gains. Specifically, students who participated in CSR for three years or longer had greater gains than students who participated for just one year. These gains were modest, but statistically significant.
Impact on minority populations
Molnar et al (1999), in their evaluation of Wisconsin's SAGE, also assessed the achievement gains of African American students. Results for grades 2-3 showed that African American students in SAGE classrooms performed higher on all subtests of the Comprehensive Test of Basic Skills (CTBS) (i.e., reading, language arts, and mathematics) than African American students in comparison schools. African American SAGE students in grades 1-3 also achieved greater gains than white SAGE students on the total score from the pre-test to post-test, closing the achievement gap.
Nye (2000) analyzed Project STAR data to determine if certain subgroups of students had greater gains in achievement when placed in small classes. The researchers found that minority students participating in small classes had larger gains in achievement than white students in small classes for both reading and mathematics in grades K-3. (The gains were slightly more modest for math.) In both reading and math, minority boys had greater gains than white boys, and minority girls had greater gains than white girls.
In another look at Project STAR data, Nye et al (2004) explored the long-term effects of reading and mathematics achievement for minority students who had participated in small classes. When the original experiment concluded, minority students in small classes had showed greater gains in reading and mathematics achievement than white students in small classes. In this study, the researchers found that students maintained these gains, to some extent, for up to five years—through grade 8. Both white and minority students who took part in small classes had statistically significant higher scores in reading and mathematics than students in large classes. Minority students who participated in small classes for four years had higher reading achievement scores than white students who were in small classes for the same amount of time. It also appears that girls maintained their gains in mathematics achievement for up to five years.
Lack of gains in student achievement
Other studies identified through our literature search found no convincing evidence linking reduced class size and increased student achievement. These studies often involved programs that
- Set reduced class size at 20 students or more;
- Were short-term and/or unevenly implemented; and
- Encountered unanticipated consequences that obviously affected outcomes.
In some instances, the researchers pointed out inadequacies in the research methodology that might contribute to these findings.
A state-mandated, four-year evaluation of California's CSR program ultimately described its assessment of the program's relationship to improved academic achievement as "inconclusive" (Bohrnstedt and Stecher, 2002). Statewide, average test scores in California improved as CSR reached more students, but the researchers could not determine whether the gains in test scores were attributable to CSR or to one or more of the other major initiatives under way in California at the same time. The state's rapid implementation of CSR also left researchers without a control group of students enrolled in larger classes.
Perhaps most importantly, the scale of California's initiative, and the speed with which it was implemented, created unanticipated consequences whose impact the researchers were not prepared to measure. The state's goal of reducing K-3 class size to no more than 20 students per teacher (from an average of 29 students) required an enormous increase in the number of teachers in these grades; between 1995-96 and 1998-99, an increase of 46 percent. Many districts hired teachers without full credentials to meet this demand. Most of these hires were made by schools serving the most disadvantaged, because they were slower to implement CSR and more of the teachers with full credentials had already been hired elsewhere (Bohrnstedt and Stecher, 2002). In 2000-01, more than one in five K-3 teachers in schools with high percentages of low-income, ELL, minority, or Hispanic students lacked teaching credentials.
Researchers also have examined several district-level class size reduction programs. Michigan's Saginaw City School District reduced class size for first and second grades, to 18:1 and 21:1, respectively. The control groups consisted of first grade classes of 21 students or more, and second grade classes of 28 students or more. An analysis by Kurecka and Claus (2000) found no significant statistical differences between the control groups and experimental groups in special education placement, attendance, or promotion rate for either first or second grade. For first grade, they found no significant differences in either math or reading achievement between control and experimental groups over the course of the year. No pretest was administered for the second grade. As the researchers noted, an array of issues could have contributed to their finding of no impact: students were not randomly assigned, reading comprehension measures were inadequate and poorly administered, and the intervention was too brief.
In 2001-2002, North Carolina's Wake County Public School System launched a class size reduction (CSR) program, through which the system allocated 40 new teacher positions to 23 schools. Most schools created one entirely new class with the new teacher position, with the new class about equal in size to the others in that grade, enabling the school to somewhat reduce the student-teacher ratio for that grade, but rarely to achieve the goal of 18:1 set by the enabling legislation—even three years into implementation (Speas, 2003). More schools placed the new teachers in first grade than grades K, 2, or 3. Schools that received two teacher positions usually established CSR classes in two separate grades.
Speas (2003) found mixed results in an evaluation of the Wake County CSR program's impact on achievement in its third year. At grade 3, no significant differences were found in either reading or math achievement between students in reduced size classes and those in regular size classes. The researcher noted, however, that Wake County's pattern of implementation may have contributed to this finding: many classrooms had a student-teacher ratio of more than 18:1, and implementation across grades K-3 was uneven.
Likewise, Munoz (2001) found no impact on student learning as measured by standardized mathematics and reading tests after a one-year class reduction in the third grade of the Jefferson County Public Schools in Louisville, Kentucky. The researcher observed, "A one-year intervention does not produce immediate results in student learning."
Johnson (2000) examined the data from the 1998 National Assessment of Educational Progress (NAEP) reading examination to analyze the effect of class size on academic achievement, and found little or no impact. Small classes were defined as those with 20 or fewer students per teacher, and large classes, as 31 or more students. The study looked at academic achievement by analyzing six factors: class size, race and ethnicity, educational attainment of parents, number of reading materials in the home, free or reduced price lunch participation, and gender. The researcher examined a nationwide sample of public school children in grades 4, 8, and 12.
The study found that being in a small class did not increase the likelihood of a student's attaining a higher score on the NAEP reading tests. Looking at the other variables, Johnson also found no evidence of any impact of class size on the reading ability of any of the students in specific populations (e.g., free or reduced lunch, race/ethnicity, educational attainment of parents, etc.). Johnson concludes that "it is quite likely, in fact, that class size as a variable pales in comparison with the effects of many factors not included in the NAEP data, such as teacher quality and teaching methods."
Conclusion and implications
Even in light of findings that suggest no relationship between class size and student achievement, the preponderance of the evidence supports positive effects and academic gains when class size reduction programs in the primary grades are well-designed and properly implemented.
Student achievement, however, is not the only factor in play. The possible benefits of smaller classes must be weighed against the costs (see Hoxby, 2002 for one analysis of the costs of class size reduction). To reduce class size in a meaningful way, school districts might need to hire more teachers, add more classes, purchase more supplies—or all of the above. Questions of class size can figure in decisions from teacher contracts to school construction.
Hiring more teachers can be especially difficult. Public schools already are straining to fill positions as an aging workforce edges closer to retirement and fewer young people enter the profession. New standards for teacher quality established by federal legislation could further complicate the supply and demand problem.
Some researchers (e.g., West & Woessmann, 2003) believe that school districts would do better to hire fewer teachers with better credentials than to hire more teachers without regard to the level of credentials and experience. They argue that the quality of the teacher, rather than the size of the class, drives student achievement.
In short, the stakes are high when undertaking these initiatives since debates continue about the ability of reduced class size to fuel student achievement, making it critical to approach the issue armed with credible research that helps inform decision-making.
 California's elementary classes at the time averaged 29 students, the largest in the country. Referenced in Bohrnstedt and Stecher 2002. See http://nces.ed.gov/pubs2002/digest2001/tables/PDF/table069.pdf.
 The No Child Left Behind Act requires local school districts to ensure that all teachers hired to teach core academic subjects in Title 1 programs are "highly qualified." In general, a highly qualified teacher is fully certified, holds a bachelor's degree, and demonstrates competence in subject knowledge and teaching. (Core subjects included English, reading or language arts, mathematics, science, foreign languages, civics and government, economics, arts, history, and geography.) The Act also calls for all teachers of the core academic subjects (teaching in the Title 1 programs or elsewhere) to be highly qualified by the end of the school year 2005-06.