International Policy Center Home Page

"Class sizes in Michigan, the quiet crisis," new research by Education Policy Initiative

September 21, 2016

New research from the Education Policy Initiative at the University of Michigan's Ford School of Public Policy found that many Michigan K-12 students experience very large core classes—with 40 or more students—but that some students are at greater risk. The researchers summarize their findings in a new policy brief.

Core courses included homerooms for elementary school students and math, science, English and social studies courses for middle and high school students.

"Data on class size is something you think should be widely available," said lead author Brian Jacob, the Walter H. Annenberg Professor of Education Policy and professor of economics. "But it's not something that historically you could just look up on the web."

Instead, Michigan schools report student-teacher ratios, which the researchers find significantly understate the number of students in core academic courses.

Jacob's analysis found that 95 percent of first-graders and 98 percent of seventh- and ninth-graders were in core courses larger than the traditionally reported Michigan student-teacher ratio (for 2013-14) of 18:1. Using the new methodology, the median class size for first-graders was 25; for seventh- and ninth-graders, it was roughly 30.

"Student-teacher ratios combine a 45-student geometry course with a three-student pullout course and give you an average of 24 students per teacher at the school," Jacob said. "Twenty-four students might not sound bad, but 45 students is significantly more than you'd want."

Student-teacher ratios also don't allow you to analyze class size by course, he said. They assign equal weight to gym classes and more challenging academic subjects like chemistry, algebra and English.

"There is an extensive research literature on the impact of class size on student achievement," Jacob said. "Perhaps not surprisingly, students in larger classes perform worse, on average, than students in smaller classes. These effects can persist for many years, and even show up in college attendance and other important long-term outcomes."

In 2010, the Michigan State Department of Education began collecting the classroom-level data that made this analysis possible. Recently, however, the department has announced plans to discontinue this data collection.

Jacob said that this will make student-teacher ratios the only measure of class size, and would make it impossible to track trends in class size over time.

"It's very tempting for school districts that are facing budget shortfalls to make incremental class size increases to reduce costs," he said. "But if this happens year after year, class sizes can continue to rise until they're quite high—as high as some of the statistics we report in our analysis."

Some states do cap class sizes, but not Michigan.  

Using student-level administrative data for the 2014-15 school year, U-M researchers looked at important transition years—first grade for elementary school, seventh grade for middle school and ninth grade for high school.

In elementary school, they found that roughly 4 percent of Michigan first-graders were in a homeroom with 40 or more students, while 10 percent of Michigan seventh-graders and nearly 12 percent of Michigan ninth-graders had at least one core course of that size.

Race, socioeconomic status, educational performance and the location of a student's school district were associated with increased risk of being in large core courses.

Black students are disproportionately exposed to large class sizes across all grades studied. Nearly 11 percent of black first-graders are placed in large classes, which is more than three times the rate for Latino students (3 percent) and five times the rate for white students (2 percent). Eighteen percent of black seventh-graders were placed in at least one large class (versus 13 percent of Latino and 8 percent of white students) and 26 percent of black ninth-graders were placed in at least one large class (compared with 12 percent of Latino and 8 percent of white students).

Likewise, students attending high-poverty schools are much more likely than their Michigan peers to be in a core class with 40 or more students. This is most pronounced among seventh-grade suburban students: 28 percent who attend high-poverty suburban schools have at least one large core course, versus just 6 percent who attend low-poverty suburban schools.

The researchers summarized their findings in a policy brief.

Brian Jacob

Contact: Erin Spanier, 734-615-7545, [email protected]

U-M has a satellite uplink TV studio and an ISDN radio line for interviews.