First-Year Findings from the Middle School Mathematics Professional Development Impact Study

| Michael S. Garet, Andrew J. Wayne, Fran Stancavage, James Taylor, Kirk Walters, Mengli Song, Seth Brown, Steven Hurlburt, Pei Zhu, Susan Sepanik, Fred Doolittle

This report presents interim results from the Middle School Mathematics Professional Development Impact Study, which is sponsored by the Institute of Education Sciences (IES). The report presents results immediately following one year of the study’s professional development. A future report will present results following two years of professional development.

Student achievement in mathematics has been a focal concern in the United States for many years. The National Research Council’s 2001 report and the recent report of the National Mathematics Advisory Panel (2008) both called attention to student achievement in mathematics, and both called for all students to learn algebra by the end of eighth grade. Reports have argued, further, that achieving this goal requires that students first successfully learn several topics in rational numbers — fractions, decimals, ratio, rate, proportion, and percent. These topics are typically covered in grades 4 through 7, yet many students continue to struggle with them beyond the seventh grade. The National Mathematics Advisory Panel wrote that “difficulty with fractions (including decimals and percent) is pervasive and is a major obstacle to further progress in mathematics, including algebra” (p. xix). The panel also specified that by the end of seventh grade, “students should be able to solve problems involving percent, ratio, and rate, and extend this work to proportionality” (p. 20).

The U.S. Department of Education’s National Center for Educational Evaluation and Regional Assistance (NCEE) — within the Institute of Education Sciences — initiated the Middle School Mathematics Professional Development Impact Study to test the impact of a professional development (PD) program for teachers that was designed to address the problem of low student achievement in topics in rational numbers. The study focuses on seventh grade, the culminating year for teaching those topics. The study is being conducted by the American Institutes for Research (AIR) and MDRC together with their evaluation partners REDA International and Westat.

Currently, through the Elementary and Secondary Education Act, the federal government provides significant resources for PD, but little rigorous evidence is available on the impact of PD on teacher and student outcomes. Hundreds of studies have addressed the topic of teacher learning and PD (for reviews, see Borko 2004; Clewell, Campbell, and Perlman 2004; Kennedy 1998; Richardson and Placier 2001; Supovitz 2001; Yoon, Duncan, Lee, Scarloss, and Shapley 2007). The most recent review of studies of the impact of teacher PD on student achievement revealed a total of nine studies that have rigorous designs — randomized control trials (RCTs) or certain quasi-experimental designs (QEDs) — that allow causal inferences to be made (Yoon et al. 2007). Four of the nine studies focused on the effect of a PD program on mathematics achievement, and none focused on mathematics at the middle school level.

The Middle School Mathematics PD Impact Study is the first rigorous test of the impact of a PD program focused on teachers of middle school mathematics. Within 12 participating school districts, the study randomly assigned 77 mid- and high-poverty schools to treatment and control conditions and collected outcome data on teachers and students. The PD was delivered by two provider organizations, each of which served the treatment schools in six of the 12 participating districts. Seventh-grade teachers in the treatment schools had the opportunity to receive the PD program offered by the study and could also continue to participate in the PD activities that they would have received in the absence of the study. Seventh-grade teachers in the control schools received only the PD that they would have received in the absence of the study.

The study has three central research questions:

  1. What impact did the PD program provided in this study have on teacher knowledge of rational number topics?
  2. What impact did the PD program provided in this study have on teacher instructional practices?
  3. What impact did the PD program provided in this study have on student achievement in rational number topics?

The study produced the following results:

  • The study’s PD program was implemented as intended. The PD providers delivered an average of 67.6 hours of PD per site, compared to 68 hours intended, and the treatment group teachers attended an average of 83 percent of the PD that was delivered. In surveys given to treatment and control group teachers, treatment group teachers reported participating in 55.4 hours more mathematics-related PD than the control group teachers.
  • The PD program did not produce a statistically significant impact on teacher knowledge of rational numbers (effect size = 0.19, p-value = 0.15). On average, 54.7 percent of teachers in the treatment group answered test items of average difficulty correctly, compared with 50.1 percent for teachers in the control group.
  • The PD program had a statistically significant impact on the frequency with which teachers engaged in activities that elicited student thinking, one of the three measures of instructional practice used in the study (effect size = 0.48). This measure encompasses such behaviors as asking other students whether they agree or disagree with a particular student’s response and also includes behaviors elicited from the students such as offering additional justifications or strategies. Treatment teachers on average engaged in 1.03 more activities per hour that elicited student thinking. The PD program did not produce a statistically significant impact on the other two measures of instructional practice: Teacher uses representations (effect size = 0.30; p-value = 0.0539) and Teacher focuses on mathematical reasoning (effect size = 0.19; p-value = 0.32).
  • The PD program did not produce a statistically significant impact on student achievement (effect size = 0.04, p-value = 0.37).