Teachers, parents, and district leaders are passionate about mandated class size reductions slated to take effect next school year. Reducing class sizes is usually a political win because parents like their kids to be in smaller classes and teachers like to teach fewer students. However, funding North Carolina’s legislated class size reductions is a challenge for the General Assembly.
When lawmakers passed legislation limiting class sizes to 18 in kindergarten through third grade in 2016, they essentially removed funding for elective teachers, angering many parents and teachers. The House passed legislation fixing this issue in 2017, but it went nowhere in the Senate. Instead, they ended up delaying implementation until the 2018-2019 school year.
While the debate on class size has been heated, it has largely centered on the funding issue. Noticeably absent is discussion of the research on class size reductions and student achievement.
A review of the class size reduction research and conversations with experts reveal five important findings:
1. Reducing class size is not a slam dunk
Restricting class sizes is (usually) popular because it appeals to common sense. Fewer students in a classroom should mean that teachers are able to spend more time with each student individually. Whether or not this is actually the case is still a debated issue.
The study that launched a thousand ships, so to speak, was Tennessee’s Project STAR. From 1985 to 1989, Tennessee conducted a randomized control trial, the “gold standard” for research, to investigate the impact of smaller class sizes in kindergarten through third grade. Students and teachers in Tennessee were randomly assigned to three types of class sizes – small, regular, and regular plus a teacher aide.
Alan Krueger’s analysis of the data showed students in small classes outperformed those in larger classes, and the presence of a teacher aide had little effect. Krueger found effect sizes relative to the standard deviation of the average percentile score of 0.20 in kindergarten, 0.28 in first grade, 0.22 in second grade, and 0.19 in third grade.
To put that into context, the impact of being assigned to a small class in kindergarten is roughly 64 percent as large as the white-black test score gap and 82 percent as large in third grade. These are sizable effects, especially given that it is notoriously hard to find large effects of educational interventions.
Because it is the only randomized control trial of class size, experts put a lot of weight on the results. However, several sophisticated quasi-experimental studies have found no impact of class size on student achievement in other states.
One rebuttal to Project STAR comes from Caroline Hoxby’s research on the impact of smaller classes in Connecticut. While Connecticut did not run a randomized control trial, Hoxby uses both random variation in the population and Connecticut’s maximum class size rule to estimate the impact of smaller classes on student achievement. She finds zero impact of smaller classes.
Florida passed a constitutional amendment in 2002 that capped class sizes from kindergarten to high school. Matthew Chingos took advantage of the staggered introduction of class size reductions at the district and school level between 2004 and 2009 to estimate the impact of smaller class sizes. He found no significant effects of smaller classes on test scores in third through eighth grade.
2. Tennessee’s Project STAR may not be as convincing as it looks
The results from Tennessee’s Project STAR convinced several states to introduce class size reduction legislation. According to Education Commission of the States, 25 states currently have policies restricting class sizes to 22 students or lower, and 15 states have policies specifically for kindergarten through third grade.
However, one expert believes the Project STAR results are overblown because of something called the Hawthorne effect. The Hawthorne effect refers to the phenomenon that occurs when individuals increase their productivity because they know they are being observed. Hoxby argues that teachers and administrators who were part of the study could have been motivated to show that smaller classes were beneficial so Tennessee would enact policies to reduce class sizes.
3. Class size legislation can have negative unintended consequences
As with many other policies, statewide mandates on class sizes have been found to have negative indirect effects. Researchers Christopher Jepsen and Steven Rivken analyze the impact of California’s 1996 class size reduction initiative, which reduced kindergarten through third grade class sizes from about 30 students per class to 20 students at an annual cost of over a billion dollars.
Because California’s class size initiative was so large, California hired 25,000 new teachers within the first two years. Jepsen and Rivken found that many of these teachers were uncertified and had no teaching experience. Furthermore, following the increased demand for teachers, more experienced teachers left their positions in poorer schools and moved to affluent communities, leaving the neediest students with the least qualified teachers.
Jepsen and Rivken found small positive effects of smaller class sizes on student achievement in California, but they also found that the influx of new, uncertified teachers hurt student achievement, particularly in lower-income schools. In some cases, the negative effect of unqualified teachers canceled out the positive effect of smaller classes.
Class size legislation in California also led to a greater use of mixed grade classes. David Sims found that some schools combined second and third grade classes to save money. As teachers know, the larger the variation in skill level in the classroom, the harder it is to reach every child. Unsurprisingly, Sims found lower test scores for those students placed in mixed grade classes.
4. Class size legislation can improve student achievement by attracting private school students
Last month, researchers published a new working paper on California’s class size reduction initiative that sheds new light on the debate. California’s class size reduction initiative signaled to parents that the public schools were improving. As a result, affluent parents who were going to send their children to private school ended up sending them to public school, usually for kindergarten through fifth or sixth grade.
The influx of affluent students into the public schools had two important results. First, test scores increased simply due to adding more high-ability students to the test-taking pool. However, changing school composition does not account for the total increase in test scores. Instead, the researchers found positive spillover effects of having more high-ability students. In other words, students of all levels did better when the classroom had more highability students.
The researchers found this “general equilibrium effect” of having more affluent students in the classroom was actually larger than the direct effect of smaller class sizes. This is significant for several reasons.
As one of the researchers, Hugh Macartney, said, “whether or not you think class size reduction has a positive direct effect, there is this additional benefit of public school reform, whether it is class size reduction or some other reform to improve public schools.” If parents think public schools are improving because of class size, they may improve simply as a result of having higher-ability kids in the room rather than having smaller class sizes.
5. Class size research is outdated
The last factor to consider in the class size debate is that the research is largely outdated. While there are new studies such as the one described above, these studies are using data from 10 to 30 years ago. Project STAR took place 30 years ago and California’s class size reduction initiative took place two decades ago, yet those studies are still cited as reference points.
Many schools are moving towards one-to-one devices, blended learning, and more innovative staffing models. These studies do not account for these changes, so caution is advised in extrapolating the results to present-day classrooms.