Educational Issues Series: Class Sizes
Helen Pate-Bain (1992), in her report on the
STAR class size study, concluded, “We view education not as
a mass- production effort, but as a personal and individual
experience. The model is not the factory. The focus is on serving
clients. Class-size research is not an attempt to reduce class size;
at its best it is an effort to find appropriate casework loads,
because much of sound educational practice consists of individual
instruction, coaching, mentoring, and tutoring (p. 256)."
We’ve learned much recently about instructional practices and
how students learn. Yet, the controversial issue of class size and its
effect on educational practices and student outcomes continues to be
confusing and often contradictory.
Effective schools research has extended its research agenda by
focusing on an expanded number of variables that are presumed to be
related to student achievement. Schools are now perceived as a
cultural entity where the complex interplay of multiple variables
affect the lives of all who learn and teach in those institutions. As
the metaphor for American schools has shifted from an assembly line to
that of a caring, learning community, the class size research agenda
has also shifted to include such variables as instructional method,
teacher morale and stress, teacher work load, student behavior and
attitudes, content areas, student characteristics, and grade level.
Bennett (1987), in a review of more recent research, found broad
agreement among researchers on the following general conclusions:
- Smaller classes result in increased student-teacher contact.
- Reductions in class size to less than 20 students without changes
in instructional methods cannot guarantee improved academic
- No single class size is optimal for all grade levels and
- Smaller classes appear to result in greater achievement gains for
students with lower academic ability and for those who are
economically or socially disadvantaged.
- Classroom management improves in smaller classes (fewer
- Smaller classes result in higher teacher morale and reduced
- Individualization is more likely to occur in smaller classes.
- Class size reductions alone do not necessarily lead to adoption
of dramatically different instructional methods.
- Class size appears to have more influence on student attitudes,
attention, interest, and motivation than on academic achievement.
- Smaller classes are beneficial for children at the primary level,
particularly in math and reading.
- Very small classes of five or fewer students produce considerably
higher achievement (Bennett, p. 3).
As class size comes to be understood as only one variable in the
complex culture of school life and as research methodologies
appropriate to study this level of complexity are perfected, it will
be possible to move beyond the search for the “correct”
number of students to be assigned to any particular classroom.
Defining Class Size
Defining class size is tricky business. For example, a district may
calculate the student-teacher ratio by including all licensed
professional personnel and emerge with a ratio of 15.6 students per
teacher. If all administrative and support personnel are omitted, the
ratio may become 20 or more students per teacher. Neither ratio may
reveal with much accuracy the conditions under which the average
student is taught and learns in that district. The differences that
occur among grade levels and subject areas are obscured in most class
Allen (1992), in a Wisconsin class size study, found that the median
number of students evaluated by elementary teachers was 25 while the
mean number of students evaluated was 88. The larger mean number
reflects the significant number of teachers of art, music and physical
education who meet with hundreds of different students each school
In the same study, Allen found that the median number of students
evaluated each year by middle, junior, or senior high teachers was 136
while the mean was 164. Over one quarter of the upper level teachers
reported evaluating 200 or more students and 2.7% of those teachers
evaluated 400 or more students each year! These data indicate that
both mean and median numbers can be misleading if there are
significant numbers of teachers who meet with very large or very small
numbers of students each day.
As generic class size research has diminished in popularity, other
related issues have received more attention. Teacher load, defined as
the total number of students taught each day, has emerged as an
indicator of school adequacy (Bennett, 1987). Teacher load has become
a key variable in schools where the quality and duration of
student-teacher interaction is valued.
Although numbers of students instructed each day or evaluated each
year are illuminating, such numbers do not reflect the total range of
responsibilities confronting the classroom teacher. Research on
teacher stress and morale has begun to examine teacher workload as a
meaningful indicator of quality (Bennett, 1987). The number of
classes, the number of preparations, subjects taught outside of one’s
specialty, student diversity, non-teaching responsibilities (such as
supervision of extracurricular activities) and school facilities,
inservice obligations, and other professional responsiblities are
included in workload research. The workload issue becomes more complex
when specific teaching responsibilities are considered, as in the case
of a secondary English teacher responsible for 150 composition
The understanding that schools are complex cultural entities and
that school quality is dependent on the quality of student-teacher
interaction has produced a new generation of class size research which
explores relationships among a wide range of variables. That research
is reviewed below.
The Research Base
The emergence of increasingly sophisticated research methodologies
has resulted in the reexamination of existing studies and has produced
a number of new research reports on the impact of class size on a wide
range of variables. In 1986, the NEA published a review of major class
size research reports. Several of these reports were of original class
size research projects while others presented the results of the
application of new analytical techniques to existing data (CLASS SIZE,
1986). For example, Glass and Smith (1978), reread, coded, and
analyzed 77 empirical research studies that incorporated achievement
test data on over 900,000 students. That meta-analysis yielded what
Glass and Smith called “general trends” in the effect of
class size on achievement.
The NEA report (CLASS SIZE, 1986, p.11), concludes: “The
central theme which runs through the current research literature is
that academic achievement does not necessarily improve with the
reduction of student-teacher ratios unless appropriate learning styles
and effective teaching styles are utilized. It is also evident that no
single class size is best for all grade levels and all subject areas.”
The central themes which emerged from the NEA study were:
- Smaller class size seems to result in higher achievement among
students who are economically disadvantaged.
- Students with lower academic ability seem to do better in smaller
classes than in larger ones.
- It may be that class size affects student attitudes more
significantly than it affects achievement.
- A direct effect of large class size is to lower the morale and
increase the stress of teachers.
- There is typically little to be gained from reductions in class
size that do not bring class size below 30.
The most recent comprehensive review of existing research was
completed by Robinson (1990), and used a cluster analysis approach.
Studies were “clustered” into categories considered
important for class size decisions such as grade levels, subject
areas, student characteristics, student achievement, student behavior,
and teaching practices. The results of part of Robinson’s
analysis appear below.
The most promising effects of small classes on pupil learning
occurred in grades K-3 in reading and mathematics, particularly in
classes of 22 or fewer students (Robinson, p.82). Two recent studies
found that these gains could not be sustained in subsequent years
In grades 4-8 the cluster of 21 studies indicated that smaller
classes had a slight positive effect on student achievement, but not
as great as in grades K-3. The gains were evident in reading and
mathematics but not in other subject areas. The data for grades 9-12
did not indicate that smaller classes had a positive effect on student
achievement; however, the number of studies done at the secondary
level were few and even "small" classes ranged from as few
as five to as many as 40 students. Economically disadvantaged and
minority students performed better academically in smaller classes at
all grade levels (Robinson, pp. 84-86).
Student Behavior and Attitudes
Studies in the K-3 cluster showed the most favorable relationship
was between smaller classes and positive student attitudes and
behavior. A minority of studies at other grade levels found positive
results as well.
Although a minority of the studies reviewed reported more favorable
teaching practices in smaller classes, there is no evidence that the
opportunities presented by smaller class size are routinely exploited
For example, in a two-year Toronto study, a substantial majority of
teachers reported that they had given more individual attention to
students and had made changes in classroom management, pupil
evaluation, and classroom layout when working with smaller classes.
However, these self-reported changes were not found in classroom
observations by the researchers. In a 1987 New York City study that
reduced class size from 26 to 16 pupils, researchers concluded, “Although
it was expected that the reduced-class size would permit teachers to
provide more small group and individualized instruction in all
curriculum areas, no meaningful differences were observed . . . .
(Robinson, p. 87).”
Robinson offers the following observations as a result of his
- The most positive effects of small classes on pupil learning
occur in grades K-3 in reading and mathematics.
- Studies examining student attitudes and behavior found the most
favorable effects of smaller classes in the primary grades.
- Smaller classes can positively affect the academic achievement of
economically disadvantaged and ethnic minority students.
- Within the midrange of 23 to 30 pupils, class size has little
impact on the academic achievement of most pupils in most subjects
above the primary grades.
- The positive effects of class size on student achievement
decrease as grade levels increase; however, the available studies in
specific subject areas in the upper grades are limited in both
number and quality.
- Little, if any, increase in pupil achievement can be expected
from reducing class size if teachers continue to use the same
instructional methods and procedures in the smaller classes that
they have used in larger classes (Robinson, p. 82).
Recent Research Data
Several recent studies of outstanding quality have emerged to
strengthen the data base on class size and effective instruction. They
include the Prime Time project in Indiana, Project STAR in Tennessee,
and the WEAC Class Size and Teaching Assignment study in Wisconsin.
These studies are described in some detail below.
In 1984, the state of Indiana allocated money to reduce first grade
classes to 18 students; in 1985, money was budgeted to reduce second
grade classes to 22 students; and in 1986, money was budgeted for a
reduction in either third grade classes or kindergarten classes.
Districts were reimbursed for their efforts to reduce class size on
the basis of average enrollments. Funds were prorated for districts
with classes averaging more than the targeted numbers; thus, classes
were smaller than targeted in some districts and much larger in others
(Mueller, Chase, and Walden, 1988). As an alternative to reducing
class size, districts were permitted to hire teacher aides; each aid
counted as .33 FTE (full time equivalent) in the student-teacher
The achievement effects of this class size project were mixed.
Project analysis, done by Indiana University faculty, found that
student achievement, as measured by traditional standardized tests
(Iowa Test of Basic Skills and Stanford Achievement), showed strong
gains in first grade reading levels with smaller gains in mathematics.
Second grade gains were much more limited. Robinson’s (1990)
analysis of the Prime Time project concluded that while there were
small but positive gains during the first year of the project,
extending the effort to third grade classes did not have any
significant effect on either reading or mathematics test scores.
Prime Time teachers who were interviewed believed that the project
had a variety of positive benefits. They indicated that students
received more individual attention, that students received more
immediate feedback, that both below-average and above-average students
achieved more, that a greater variety of instructional materials were
used, that the instructional atmosphere was less hectic, that teachers
assigned more homework, and that teachers were happier and more
enthusiastic about their teaching (Mueller, Chase and Walden, p.50).
Tennessee’s four-year study of class size effect has, to date,
produced the most compelling data on class size, effective teaching
strategies, and the sustainability of early student achievement gains
that is currently available. Project STAR, a longitudinal study,
followed students from kindergarten in 1985-86 through the third grade
in 1988-89. A selected number of fourth grade students (4,230 students
in 216 classes) were part of a Lasting Benefits Study done in 1989-90.
Project STAR analyzed student achievement and development in three
types of classes: small classes (13-17 students per teacher), regular
classes (22-25 students per teacher), and regular classes (22-25
students) with a teacher and a full-time teacher aide (Word, et al.,
1990). The project included 17 inner-city, 16 suburban, 8 urban, and
39 rural schools in order to assess the effects of class size in
different school locations. A major strength of the study was the “within-school”
design which required each school to have at least 57 students at the
appropriate grade level so that each school would contain at least one
of each class type (small, regular, and regular/aide). Students
(6,000) and teachers were randomly assigned to class types.
Student achievement was measured with three instruments: the
Stanford Achievement Test, STAR’s Basic Skills First Criterion
Tests, and Tennessee’s Basic Skills Criterion Tests. Kindergarten
results showed a definite advantage for small classes but no
significant advantage for the regular/aide classes. By the end of the
first grade, Project STAR students in small classes were outperforming
students in regular and in regular/aide classes by substantial margins
(statistically and educationally significant).
The presence of a teacher aide in a regular class benefited students
but not as much as the small class condition. Students in second grade
small classes continued to outperform students in regular and
regular/aide classes on all tests. Differences between students in
regular and regular/aide classes were not significant. By grade three,
the pattern of results established in kindergarten had become firmly
fixed. A strong class-size effect was evident in all school locations
(urban, rural, inner-city, and suburban) and for all students. The
absence of a statistically significant teacher aide effect was
consistent (Word, et al., pp. 10-13).
Several other findings were relevant:
- The only consistent positive effect in regular classes with a
full-time aide occurred in first grade.
- Teachers reported that they preferred small classes in order to
identify student needs and to provide more individual attention, as
well as to cover more materials.
- The importance of the economic background of students was
underscored by the finding that, in every situation, those students
who were not economically eligible for the free lunch program always
outperformed those students who were in the free lunch program
(Pate-Bain, et al., p.254).
The achievement benefits of small class size were cumulative. The
top achieving STAR classes in the study were categorized by class type
each year.That is, in the first year 55% of the "top classes"
were small classes (13-17 students). By the time these students
reached third grade, 78% of the "top classes" consisted of
small classes. In other words, over time, the benefits of small class
size on student achievement became increasingly apparent.
A Lasting Benefits Study was conducted with a sample of fourth grade
students in 1989-90. The LBS study yielded clear and consistent
“Students who had previously been in small STAR classes
demonstrated significant advantages on every achievement measure over
students who had attended regular classes. Further, these results
favoring small classes were found to be consistent across all school
locations'' (Pate-Bain, et al., pp. 254-255).
Teacher aides were less effective than small classes in enhancing
student performance at each grade level. Although aides performed a
variety of clerical, custodial, and instructional tasks, the pattern
of aide activities was not related to student achievement. While
teachers liked teacher aides, the teacher aide did not have much
effect on student learning in Project STAR.
If educators are to make the best use of opportunities provided by
small class size, they must be aware of what constitutes effective
teaching in such settings. STAR researchers studied the classroom
strategies used by the 49 first grade teachers whose small classes had
made the greatest gains in achievement. These effective teachers
displayed similar affective behaviors and characteristics:
Their enthusiasm was obvious as they engaged in “acting,”
demonstrating, and role-playing activities. The teachers frequently
expressed positive attitudes toward children, emphasized positive
behavior, praised success, and used humor to promote learning and to
motivate students. A love of children seemed to permeate their
professional repertoires. The most effective teachers engaged their
students through the use of creative writing, hands-on experiences,
learning centers, and math manipulatives'' (Pate-Bain, et al., p.
Project STAR recognized at the outset that small class size does not
make a difference if teachers do not change the way they teach. A
subgroup of randomly selected STAR teachers were selected for three
days of inservice training before the project began. The training was
designed to help them teach more effectively in whatever class type
they were assigned to teach. Subsequent data revealed that there were
no significant differences in student achievement in reading or
mathematics between classes where the teachers were trained and all
other classes where the teachers had not been trained. Several
- Teachers with small classes must not only be trained to be
effective in such settings, but they must also be committed to try
new skills and procedures.
- Effective inservice training must include time to visit other
teachers who have had success in working with small classes.
- Inservice work must include training in effective communication
with the home, in involving the family with the education of their
children, and in making home visits as part of regular instructional
Project STAR has set high standards for future class size research
and has provided clear guidelines for schools committed to improving
student achievement in grades K-3.
Class Assignment and Teaching Assignments Study (WEAC)
This WEAC study (Allen, 1992) extended the parameters of traditional
class size research by examining teacher load and workload, the
allocation of the teacher workday, the impact of special needs
students on classroom instruction, and the consequences of classroom
conditions for teacher stress and morale. Allen surveyed a random
sample of 1,150 elementary and 1,500 middle, junior and senior high
teachers in Wisconsin in order to ascertain the qualitative conditions
of classroom life. Student achievement data was not reviewed.
The Wisconsin study revealed that there was little difference in the
length of the school day reported by teachers at the various grade
levels; the median length of the school day was 467 minutes for
elementary teachers and 468 minutes for teachers in the upper grades.
However, some variation was reported in “direct contact”
time, with elementary teachers reporting 325 minutes each day in
direct contact with students and teachers at the middle, junior and
senior high grades reporting 283 minutes of direct contact time
(Allen, p. 8). These data do not include supervisory time.
The complexity of teacher workload was addressed in a number of ways
in this research. It was determined that for teachers meeting with the
same group of students each day (60% of elementary teachers and 4% of
middle, junior and senior high school teachers) the median number of
students in class was 24. However, evaluation responsibilities were
determined by Allen to be a more accurate indicator of workload. These
data revealed that the median number of students evaluated by
elementary teachers each year was 25 while the mean number was 88. The
median number of students evaluated each year by middle, junior and
senior high school teachers was 136, while the mean was 164. Over one
quarter of these teachers reported evaluating 200 or more students
each year (Allen, pp. 10-11). As noted earlier, teachers of art,
music, and physical education confront heavy workload and evaluation
The impact of student diversity, specifically the numbers of
students with Exceptional Educational Needs, was also addressed. The
number of students identified as having special needs in Wisconsin
schools has increased from 5% to more than 9% of the student
population over the past fifteen years. These students require extra
attention and more preparation time if they are to be successful in
the classroom. The WEAC study found that the median number of EEN
students taught by elementary teachers was between 4 and 5 while the
median number taught by teachers at the upper grade levels was nearly
8. To assess the impact of these numbers on classroom instruction,
Allen created an "EEN Index Number." The number reflects the
total number of minutes of instructional time during which the teacher
is in contact with EEN students. It is a measure of the increased
responsibility the teacher confronts in such circumstances. The index
number is calculated by multiplying the number of EEN students present
by the number of minutes the teacher meets with each student on a
daily basis. For example, if Teacher A meets with three EEN students
for ten minutes each day and Teacher B meets with three EEN students
during the entire school day (assume 325 minutes), the Index Numbers
would be 30 and 975 minutes, respectively.
Elementary teachers reported a median index number of 600 minutes
spent each day with students who have special needs, while teachers at
the middle, junior and senior high reported an index number of 400
minutes per day (Allen, pp. 13-14). The amount of total contact time,
as an indicator of professional responsibility, is more important than
is the total number of EEN students enrolled.
Between two-thirds and three-fourths of the teachers responding to
the WEAC study indicated that they would make changes in how they
teach if they had significantly fewer students in their classes.
Changes mentioned included:
- more individual attention
- more written assignments
- more creative activities
- more opportunity to monitor the progress of each student
- more problem-solving activities, more projects, more essay tests
- more attention to gifted children, and
- more field trips (Allen, p. 17).
Between 8 and 14% of the teachers indicated they would not make
changes in how they would teach and about 18% were not certain.
Implications of the WEAC Study
In a previous study of teacher mobility in Wisconsin, Allen and
Helming (1991) found that a significant number of teachers expressed
high levels of job dissatisfaction. Thirty-one percent of Wisconsin
teachers had given “quite a lot” or “a great deal”
of thought to changing teaching jobs, leaving teaching, or both. Large
class sizes and excessive responsibilities contribute to high stress
levels and high levels of job dissatisfaction among teachers. This
research offers new insights into the complexities of teacher workload
and , by implication, suggests avenues by which student achievement
may be enhanced.
Summary and Recommendations
Major findings presented in the research on class size include:
- There is probably no optimum class size for all types of
students, in all subject areas, and at all grade levels.
- Smaller classes produce the necessary, though not sufficient,
conditions for successful teaching and learning.
- Reduced class sizes in grades K-3, in the range of 13-17 students
per class, significantly enhance student achievement.
- Reductions in class sizes to less than 20 students without
changes in instructional methods cannot guarantee enhanced student
- Small class size, in the range of 13-17 students, benefits all
students in all contexts at the K-3 level.
- Regular-size classes with a teacher aide are less effective than
small classes in enhancing student achievement at the K-3 level.
- There are identifiable teaching behaviors which will enhance the
benefits of a small class design; these include individual
instruction, coaching, mentoring, and tutoring.
- The evidence favoring small class size at the upper grade levels
is weak because teaching behaviors appear to be more rigid and
research methodologies have been inadequate.
- Teacher inservice opportunities must accompany reduced class
sizes so that appropriate teaching can be developed and reinforced.
- Class size definitions vary, depending on whether mean or median
numbers are used.
- Class size averages may obscure the fact that some students,
such as EEN, require extra attention and care.
- Both class size and teacher workload (the number of students
evaluated during the year) are important indicators of school
This review of existing research, with all of its incumbent
limitations, suggests that increasing student learning by reducing
class size is a complex matter. It suggests that class size reductions
should be targeted to specific student groups for specific purposes,
that teachers must receive the training needed to make the most of the
new learning opportunities available in smaller classes, and that
little is known about how to effectively approach the problem of large
student numbers in the upper grade levels without confronting the
basic structures of those institutions.
Allen, Russ. CLASS SIZE AND TEACHING ASSIGNMENTS OF WEAC’S
TEACHER MEMBERS. Madison, Wisconsin: Wisconsin Education Association
Allen Russ and Helming, Margaret. TEACHER MOBILITY IN WISCONSIN.
Madison, Wisconsin: Wisconsin Education Association Council, 1991.
Bennett, Susan. NEW DIMENSIONS IN RESEARCH ON CLASS SIZE AND
ACADEMIC ACHIEVEMENT. Madison, Wisconsin: University of Wisconsin,
Cotton, Kathleen. SUMMARY OF RESEARCH ON CLASS SIZE. Portland,
Oregon: Northwest Regional Educational Laboratory, 1988.
Glass, Gene V. and Smith, Mary L. META-ANALYSIS OF RESEARCH ON THE
RELATIONSHIP OF CLASS SIZE AND ACHIEVEMENT. San Francisco: Far West
Laboratory for Educational Research and Development, 1978.
Hawley, Willis D. et al., “Good Schools: What Research Says
About Improving Student Achievement.” PEABODY JOURNAL OF
EDUCATION (Summer 1984).
Mueller, Daniel J.; Chase, Clinton I.; and Walden, James D. “Effects
of Reduced Class Size in Primary Classes.” EDUCATIONAL LEADERSHIP
(February 1988): 48-50.
Pate-Bain, Helen; Achilles, C.M.; Boyd-Zaharas, Jayne; and McKenna,
Bernard. “Class Size Does Make a Difference.” EDUCATIONAL
LEADERSHIP (November 1992): 253-256.
Robinson, Glen E. “Synthesis of Research on the Effects of
Class Size.” EDUCATIONAL LEADERSHIP (April 1990): 80-90.
WHAT RESEARCH SAYS ABOUT CLASS SIZE. Washington, D.C.: National
Education Association, February, 1986.
Word, Elizabeth, et al., STUDENT/TEACHER ACHIEVEMENT RATIO (Star):
TENNESSEE’S K-3 CLASS SIZE STUDY. Nashville, Tennessee: Tennessee
State Department of Education, 1990.
This document was prepared by Ken Kickbusch of the WEAC
Professional Development Division
Posted May 30, 1996