A new $1.7 million grant from the Institute of Education Sciences will fund a study aimed at reducing educational disparities by refining and testing a tool intended to optimize teacher-student classroom assignments.
The tool aims to address a phenomenon that project researchers call “disproportional assignment.” In disproportional assignment, some groups of students are systematically more likely to be assigned to less-effective teachers, contributing to persistent gaps in achievement.
The project is led by Matthew Springer, the UNC School of Education’s Robena and Walter E. Hussman, Jr. Distinguished Professor of Education Reform, in collaboration with co-principal investigators Lora Cohen-Vogel, the Frank A. Daniels, Jr. Distinguished Professor; Thurston “Thad” Domina, the Robert Wendell Eaves Sr. Distinguished Professor in Educational Leadership; Associate Professor Peter Halpin; and David Stuit, managing partner of Basis Policy Research, developer of the tool.
The work is intended to improve a key element in the administration of schools: How are students assigned to teachers?
“Research has clearly demonstrated that the most important in-school factor for determining a student’s success is their teacher,” said Springer. “Research has also demonstrated that some students — more often non-White and economically disadvantaged — are more frequently assigned to less-effective teachers.”
“We also know that students who have less-effective teachers for several years in a row may never catch up to their peers, exacerbating achievement gaps and long-run outcomes like college enrollment and persistence,” Springer said.
A tool that school leaders can use to assign students more equitably to highly effective teachers should help close those achievement and opportunity gaps.
Seeking to correct inequities in teacher assignments
The project is built on studies by the research team that show the effectiveness of teachers varies within schools, certain groups of students are more frequently assigned to less-effective teachers, and more equitable resource allocation decisions can disrupt persistent achievement gaps.
A study by the research team in a forthcoming volume on recent advances in education finance and policy demonstrated that non-White, economically disadvantaged, and English language learner students are more likely to be assigned to their schools’ least effective teachers, while their White peers are more likely to be assigned to higher-performing teachers.
The work also documents the causes of disproportional assignment.
“Disproportional assignment results from a number of causes, such as unequal influence of parents or guardians, preferences among some teachers for high-performing students, lack of awareness among school leaders of the impact of repeated exposure to lower-performing teachers, and other factors,” Cohen-Vogel said.
Cohen-Vogel and others on the research team have led studies that show schools can move toward both reducing achievement gaps among students and raising the academic achievement of all students by adjusting how they allocate teachers and students across classrooms.
“Schools can be sorting machines. In this case, they tend to sort teaching in ways that perpetuate existing inequality,” he said. “We’re hopeful that use of a tool like the one we are studying in this project will eventually provide administrators with information that can help spread teaching quality in a way that helps all students.”
Applying evidence, statistical tools to make better matches
The tool being studied in the project — the Equitable Rostering Solution, or ERS — was developed by Basis Policy Research, an independent research firm with offices in Carrboro, North Carolina, and Grand Rapids, Michigan, and Hourglass Education Technology Solutions, an educational technology firm based in Scottsdale, Arizona.
ERS uses a variation of an algorithm employed in other assignment processes, including the matching of medical school students with residency positions. ERS uses data from teacher evaluation systems, student achievement scores, and other sources to develop profiles of student and teacher needs and strengths. ERS then develops recommended classroom rosters aimed at reducing in-school achievement gaps.
Peter Halpin, a UNC School of Education associate professor and collaborator on the project, has researched data-driven algorithms that can be used to discover more equitable ways of allocating educational resources, such as teachers to students.
“A main strength of ERS is that it considers how these computation procedures can be combined with other sources of information to drive decision-making within schools,” Halpin said.
The potential for a tool like ERS has been met with great interest among school leaders.
“In our conversations with principals and superintendents across the country, the support we’ve received for this tool is overwhelmingly positive,” Stuit said. “Our goal is to equip principals with information to ensure all students are assigned to a teacher who best meets their particular academic, behavioral, and socio-emotional needs. The tool has the potential to increase students’ access to effective teachers but will also make principals’ jobs easier by standardizing how rostering is conducted across schools.”
The project’s initial phase will be an examination of current student-teacher assignment practices in North Carolina, including analysis of administrative data and focus groups and interviews with teachers, education leaders and other stakeholders in the state.
The project’s second phase will include a small-scale pilot study of use of the tool with 10 elementary schools in one large urban school district in the state.
The pilot study will examine the implementation, impacts, and costs of the ERS. The ultimate objective is to produce a commercially available version of ERS, along with user guides and training materials, as well as scholarly articles documenting the design, implementation, and effectiveness of ERS in real-world situations.
“Our team is committed to identifying and implementing ways to close gaps of opportunity and achievement that have persisted for too long in so many of our schools,” Springer said. “We’re hopeful that this project will demonstrate how data-driven tools can help equip educators and school leaders to close those gaps.”
The study is expected to take three years.