Class Size: A Battle between Accountability and Quality Instruction
Januszka, Cynthia, Dixon-Krauss, Lisbeth, Childhood Education
A substantial amount of controversy surrounds the issue of class size in public schools (Achilles, 2003; Bracey, 1995; Kronholz, 2002). Parents and teachers are on one side, touting the benefits of smaller class sizes (e.g., increased academic achievement, greater student-teacher interaction, utilization of more innovative teaching strategies, and a decrease in discipline problems). On the other side, many legislators and policymakers doubt the benefits, claiming that smaller classes produce minimal results while incurring large costs. Those who oppose decreasing class sizes believe that funds could be more wisely spent on programs that show greater benefits. This argument is framed between two opposing perspectives: accountability versus quality instruction.
Policy analysts refer to accountability as a cost-benefit analysis of a program (Heineman, Bluhm, Peterson, & Kearny, 2002). Cost-benefit analyses are often used to examine education policy issues with a focus on school administration, finance, and leadership. Another form of analysis, experimental analysis, tests the integrity of a policy. Experimental analyses are commonly used to examine school curricular issues focusing on pedagogy and student achievement. Experimental analysis requires random assignment of subjects to experimental and control groups (Heineman et al., 2002), a treatment for the experimental group, and an evaluation to determine which group performed better. This article addresses the feasibility of reducing class size by reviewing the literature, using experimental and cost-benefit analyses. The experimental analysis includes a review of class size studies and their application on large samples. The cost-benefit analysis examines class size policies, initiatives, and costs involved for one state--Florida.
Experimental Analysis of Class Size
Project STAR. In 1985, Lamar Alexander, the governor of Tennessee, led an initiative to assess the usefulness of having small class sizes in the primary grades. He authorized funds to conduct an experimental study of this issue, known as Project Student-Teacher Achievement Ratio (Project STAR). The project used random assignment of students and teachers to three types of classes: 1) a small class of 13-17 students, 2) a regular class with 22-25 students, or 3) a regular class with an aide to assist the teacher. The program lasted four years, following students from kindergarten through 3rd grade. All groups in the study received the same curriculum and materials. The sample was large, with approximately 6,400 pupils participating. Results from the study showed that students in small classes did better than their counterparts in larger classes on subject area tests in reading, math, science, social studies, and spelling. Minority students also made higher gains in smaller classes. In addition to the academic improvements, teachers were able to give more individualized attention to students, which helped to decrease the amount of behavior problems (Achilles, 2003; Bracey, 1995; Mosteller, 1995).
Lasting Benefits Study. The Lasting Benefits Study used an experimental analysis to evaluate the long-term benefits of reduced class size on student achievement. This study was a three-year follow-up (1989-1991) that tracked the progress of more than 4,500 students from Project STAR. The progress of students in the experimental group was monitored when they returned to classes of average size. Findings from the study indicated that students who were taught in small size classes in the early grades performed better than their peers when they returned to regular size classes (Mosteller, 1995).
Burke County Initiative. In 1991, a pilot study on the effects of reduced class sizes in the elementary grades was conducted in the Burke County Schools in North Carolina. In the first year of implementation, the superintendent reduced 1st-grade class sizes in four elementary schools to 15 students per class. …