Dana Goldstein rounds up the conflicting evidence on the subject:
Common sense suggests crisis class sizes of 50 or 60 students, which were seen at the depth of the recession in Detroit and some California cities, are much too large, leading even the most skilled teachers to become overwhelmed and unable to focus on individual students’ needs. Those situations are most likely to occur in school districts with high-poverty student populations, though some research suggests poor and minority children benefit most from smaller classes.
Where the class size debate gets really complex is when you look at more typical class sizes; for example, the difference between a 25 and 30-student class.
In 1985, the Tennessee STAR project randomly assigned a few thousand elementary school students to very small classes of between 13 and 17 children. A number of studies showed students in those classes had higher academic achievement than peers in larger classes, and were more likely to graduate high school. Partly in response, states like California, Florida, and North Carolina launched efforts to reduce average class sizes, but because of budget limitations, few districts have been able to lower class sizes to the extent the research literature suggests is ideal—by at least six to eight students.
Internationally, the United States has larger than average class sizes, but a few of the nations with even bigger classes than ours, such as Korea, Japan, and Australia, clearly out-perform us academically—as do several countries, like Finland and Canada, that have made small classes a priority.
(Chart from Catherine Rampell)