Class Sizes and School Choice: Parsing Yet Another Mixed Bag of Research

We are proud to share this guest post by Mary Field, a board member at the International School of San Antonio, discussing the research about class sizes and education, and offering advice for parents doing a school search.

In 2023 it is completely uncontroversial to sit on a park bench with other parents you’ve just met and extol the virtues of small class sizes. (In case you need a thought experiment for comparison, would you sing the praises of contact sports or social-emotional learning or universal free school lunches to someone you didn’t know?) Considering how popular small class sizes are with parents, one could be forgiven for thinking that they are scarce commodity. In fact, elementary class sizes have been steadily declining in America since the 1960s from an average of 29 pupils per class in 1961 to 20.9 in 2018. In Texas, as in many other states, state law dictates that prekindergarten to fourth grade classes in public schools have no more than 22 students.

The preference for smaller class sizes is widespread amongst constituencies who almost never completely agree with one another: teachers like smaller class sizes for the ease of management and reduction in workload, parents like smaller class sizes because their children will get more attention, and administrators like smaller class sizes because they make both parents and teachers happy. It is precisely because small class sizes are so popular that stakeholders owe it to themselves to take a look at what the data actually say. Further, we must look at what tradeoffs may be lurking before we pull the levers of policy or sign our children up for the smallest classes we can find.

Research Showing Benefits of Small Class Sizes

In the mid-1980s this preference was evaluated in the state of Tennessee in a large experiment called Project STAR (Student-Teacher Achievement Ratio). There were three phases to the study. Students were randomly assigned to either a small class (13–17 students), a regular-sized class (22–25 students) or a regularly-sized class with a paid aide in addition to the teacher. This was a huge study that involved thousands of students and over 80 schools. After the initial four-year study, the second phase looked into whether the effects of smaller class sizes faded over time. A third phase of the study looked at whether small class sizes help close the achievement gap.

Researchers in the Tennessee experiment looked at both standards-based tests (reading and math) and curriculum-based tests to evaluated the benefits of reducing class sizes. Test scores in reading and math for students in the smaller classes were on average both one-fourth of a standard deviation higher. On the curriculum-based tests, scores were an average of one-fifth of a standard deviation higher for reading and one-twelfth of a standard deviation higher for math. In the third phase of the project, it was found that smaller class sizes had particular benefits for economically-disadvantaged students. Later analysis of the Tennessee experiment has upheld the finding that smaller class sizes benefit students, at least in terms of raising test scores. Many subsequent studies have found benefits from reducing class sizes.

There are, however, competing studies that claim to show that class sizes have no positive effects on test scores. The economist Eric Hanushek has argued extensively that reducing class sizes does not systemically lead to better outcomes for students.

The Tennessee study was widely lauded and cited for policy changes because it was designed as a randomized control. While randomized controlled studies are indeed the gold standard for scientific experiments, it is difficult for experiments in education to actually adhere to these parameters. Participants in these experiments are always aware they are part of the experiment and may (consciously or not) alter their behavior. Further, true randomization in the Tennessee experiment was never achieved. Predictably, parents of children in the regular-sized classes complained and the design of the experiment had to be changed to that children in the regular-sized classes could be randomly reassigned the following year to either a class with an aide or without an aide. About 10 percent of students were also reassigned due to parental complaints or behavioral problems. It was also noted that some students were reassigned “to achieve racial and sexual balance” for the classes belying the idea that the students were randomly assigned to begin with.

The classic conundrum of studies in education is to control for the endogenous variables which inevitably have a strong influence on student outcomes, like parental education levels. The only way to eliminate the influence of endogenous variables (and parents who complain about their child’s placement in a randomized control trial is the ultimate endogenous variable) is to look for a natural experiment in class size reduction.

Research Showing No Effect of Small Class Sizes

One such natural experiment was found in the state of Connecticut, where most school districts are essentially towns. The vast majority have rules about the maximum class size and the majority have rules about minimum class sizes. This study looked at test scores when unusually large cohort of students triggered a rule on maximum class size creating two small classes rather than one large class, or if there was a variation in population that created an unusually large or small cohort. For example, if a small town typically only has one class per grade level with 22–26 students and a maximum class size rule of 29, a year in which there are 30 kids registered for first grade would trigger creating two classes of 14 and 15 students each, without any parents choosing the school for its smaller class sizes of or perhaps creating an experiment in which almost all stakeholders desire a positive finding. This study found no effect on student achievement from smaller class sizes.

The results of the Connecticut study imply that what was found in Tennessee was really more of a “Hawthorne effect.” That is, the participants knew they were part of a study about class size reduction and altered their behavior accordingly.

Class Size Data in San Antonio

Considering the thoroughly mixed-bag of studies on whether reducing class sizes has any benefit for student achievement, one could be forgiven for thinking that parents need to vigilant lest their children end up in a class with 30 other kids. Looking at the data on schools in Texas and San Antonio, this just is not the case. A sampling of reports on cases from reveals that many schools and districts have relatively small class sizes: 15.2–19.6 students (NEISD), 18.5–28.6 (NISD), 17.6–24.6 (Alamo Heights ISD), 9.4–18 (Brooks Lone Star) and 19.8–28.2 (KIPP Esperanza). Some charter schools have larger class sizes since they do not have to follow the cap of 22 students for prekindergarten to fourth grade. Since these are choice schools, any parent simply has the option of not applying and not sending their child to a charter school with large class sizes.

Despite the general trend of ever-smaller classes, the concern about large class sizes does not come out of nowhere. Many of the grandparents and great-grandparents of children going to elementary school today had large class sizes. My own mother still has a picture of her first grade class. It was taken in 1965 and there are 42 other children in the picture. Thirty-five years earlier, the average number of pupils per elementary school class was 38. There was a great deal of variation in class sizes as well: 17 percent of elementary classes 45 or more students. At the very beginning of the 20th century, my great-great grandmother was a kindergarten teacher in Chester, New Jersey. Records show that in 1904, she had 46 students in her class! With such large classes in living memory for educational stakeholders in the 20th century and later, perhaps it no wonder there has been such a persistent push for smaller class sizes.

School Choice and Class Sizes

If the body of research does not clearly support the benefits of small class sizes, but the overwhelming preference for smaller classes is still there, how should parents think about the issue? As with many other decisions in life, choosing a school is not always one great option vs a few terrible ones. It is a question of preferences and tradeoffs. Parents know their children best and may genuinely be right that their child will do better in a smaller class.

This invites the question of what they are willing to compromise on to ensure their child attends a school with small class sizes. It is possible that a school that just embarked on a hiring spree to decrease class sizes may have teacher with much less experience. Encouraged by the findings of Project STAR, California reduced its maximum class size to 20 students per class in the 1990s. In classic case of good intentions and the road the hell, the sudden demand for more elementary teachers led to an average decrease in the years of experience for teachers in the classroom. A teacher’s years of experience in the classroom is positively correlated with student achievement.

Smaller classes may have ancillary benefits that have nothing to do with test scores, but still make it worthwhile to focus a school search around small class sizes. A teacher with 16 students may have a lot more time to respond to parent emails than a teacher with 30. An introverted student may find a smaller class less overwhelming. Similarly, a school with large class sizes may also be worth going against the grain and pursuing. Large classes may mean more opportunities for socialization. Large expenditures for technology and facilities may not pencil for a boutique school with small classes, but would be financially feasible for a large school with large classes.

If your heart is set on smaller classes for your child, it is entirely possible that the district school down the street has 18 kids per class. This is only two more kids than the cap on the smaller classes that were found to be beneficial in the Tennessee study. Likewise, if you love almost everything about a charter school, except the 25–30 kids per class, there is plenty of research to suggest it is not such a hindrance to learning after all.

Charter Moms Chats

Watch Mary Field speak with Inga Cotton on Charter Moms Chats on February 16, 2023 at 4:00 PM Central live on Facebook and YouTube.

Mary Field is a member of the board at the International School of San Antonio. She has taught Mandarin Chinese in San Antonio since 2015.

Read More About Teaching and Learning

Share with friends:

Mary Field