Panel: Innovations in Value-Added Models for Education

Friday, November 8, 2019: 1:30 PM-3:00 PM
Plaza Building: Concourse Level, Governor's Square 14 (Sheraton Denver Downtown)

*Names in bold indicate Presenter

Organizer:  Christopher A. Candelaria, Vanderbilt University
Panel Chair:  Jane Best, Education Commission of the States
Discussants:  Jesse Rothstein, University of California, Berkeley and Susanna Loeb, Brown University

It is well established that highly effective teachers are important for student achievement gains. In response to this substantial body of evidence, policymakers and school leaders have leveraged statistical models to estimate the gains teachers contribute to test scores, known as value-added, as a way to rank teachers from least to most effective. Although value-added has a prominent role in assessing teachers’ impacts on short-term learning gains, there has been a recent shift in education policy to include students’ longer-term test score gains, adult outcomes, and non-cognitive outcomes. This broader policy focus requires recognition of other essential school staff, like school counselors, as well as novel approaches to examine the ways school staff impact students. This panel provides ways for re-thinking value-added for use in education research and policy. The papers provide insights into how school staff—teachers and counselors—uniquely contribute to longer-term student outcomes and success in a way that traditional value-added approaches fail to recognize.

Blazar and Pollard’s paper highlights the problem of value-added capturing a limited aspect of a teacher’s role: increasing student test score gains. In their paper, the authors recognize that teachers are also responsible for maintaining student engagement, especially as it is associated with persistence in school and a desire to pursue higher education. Leveraging data in which teachers are randomly assigned to rosters within schools, the authors identify which classroom practices are associated with student-learning gains and which are associated with higher student engagement. This paper broadens the scope of teacher effectiveness and suggests that teaching practices that contribute to test score gains differ from those contributing to social and emotional development.

Whether the short-term impacts teachers have on student achievement gains are different from the impacts that teachers have on gains in future grades is addressed by Candelaria and Bartanen’s paper. The authors refer to impacts that teachers can have on future-grade gains as medium-term effects. For example, they are interested in the impact that a 4th-grade teacher has on 5th-grade and 6th-grade test scores. The authors explore the variability in medium-term impacts and assess whether teachers that are good for short-term learning gains are also good for medium-term learning gains. Results suggest that teachers are differentially effective, and the authors attempt to identify which teacher characteristics explain the variability in medium-term learning gains.

Mulhern’s paper shifts the focus from teachers to school counselors, a staffing group that is hypothesized to play an important role in student outcomes. The author is the first to provide causal evidence of school counselors’ impacts on student achievement and education attainment. The paper estimates the variability of counselor impacts and compares them with prior studies on teachers. Importantly, Mulhern finds that improvements in counselor quality are associated with high-school outcomes as well as impacts on four-year college enrollment and degree completion, with larger effects for underrepresented minority students. 

Collectively, these papers use rich administrative data to inform policy discussions about the extent to which key personnel, teachers and counselors, contribute to student learning.

Rethinking Value-Added: Medium-Term Teacher Effects on Student Achievement
Christopher A. Candelaria and Brendan Bartanen, Vanderbilt University

See more of: Education
See more of: Panel