Get Updates via Email Get Updates Get our RSS Feed
  Follow Mathematica on Twitter  Share/Save/Bookmark

Using Value-Added Growth Models to Track Teacher and School Performance

photo of teacher with studentValue-added methods (sometimes described as student growth models) measure school and teacher effectiveness as the contribution of a school or teacher to students’ academic growth. The methods account for students’ prior achievement levels and other background characteristics. Mathematica works with school districts, charter management organizations, state education agencies, and policymakers to develop value-added methods. In addition to implementing these methods on the ground in schools and districts, we also conduct value-added research to inform issues such as teacher quality distribution as well as teacher incentives and performance.

Education stakeholders also use value-added measures to identify promising approaches for improving student achievement. The measures help detect high-performing schools and teachers whose students make strong gains during a school year, for students with low or high initial achievement. For example, using value-added data provided by Mathematica, New Leaders for New Schools studied high-performing teachers within high-performing schools to document the instructional practices these teachers use.

Because value-added measures are used to judge the relative effectiveness of schools and teachers, their statistical foundations must be solid, clearly documented, and hold up to public scrutiny. Further, they must be communicated clearly to educators so that the findings are perceived as accurate and fair. Our value-added measures are designed to be technically sound, objective, transparent, and responsive to stakeholders’ concerns. We are known for our ability to address methodological challenges inherent in modeling student achievement and to communicate the results with language and graphics that resonate with educators.

Recognizing the trade-offs inherent in developing value-added measures, we work with local stakeholders and do not insist on a single “correct” approach. At the same time, our experience can help inform educators about the consequences of modeling decisions. We use existing data on student test scores and demographics as much as possible, with additional district-level data incorporated as needed.

Partnering with School Districts and States

Our researchers have assessed the effectiveness of schools and teachers in public and charter school settings. In the Pittsburgh Public Schools, we have developed school performance measures and helped the district use them to identify ways to improve performance. For the District of Columbia Public Schools, we work within a tight timeline to help measure the effectiveness of schools and teachers as part of the IMPACT evaluation system. We have also provided value-added measures to the Achievement First network of charter schools that they use to assess teacher effectiveness as new data become available.

Improving Value-Added Methods

Mathematica researchers have made significant contributions to methodological advances in value-added methods. For instance, a Mathematica report to the U.S. Department of Education analyzed how the amount of data used in value-added models affects the accuracy with which these models can identify high- and low-performing educators. This analysis provides a clear framework by which policymakers can determine the number of years of data on which teachers’ and schools’ evaluations should be based. In a brief for the U.S. Department of Education, we estimated teachers’ value-added and measured access to highest-performing teachers in high- and low-income schools.

Use of Value-Added Methods for National Evaluations

In our work with districts and in national studies, we have developed value-added methods to estimate school and teacher performance for large randomized controlled trials. For example, our Evaluation of the Talent Transfer Initiative used value-added methods to identify high-performing teachers, who then receive incentives to transfer to low-performing schools. We measure students’ performance in the schools to which the teacher has transferred and in a set of similar schools to evaluate whether teachers recruited under this program are more effective than new recruits who fill vacancies under a traditional recruiting process. In another study, we used value-added methods to evaluate the effectiveness of intensive mentoring programs for teachers in the first two years of their careers. We also evaluated the relative effectiveness of teachers trained through different routes to certification.

As national evaluator for the 2010 Teacher Incentive Fund, Mathematica’s value-added experts are advising school districts and states in the design and implementation of performance measurement systems for use in policies at aimed at supporting and rewarding effective teaching in sites around the country.

We are also using value-added methods to document the distribution of teacher quality in a set of large, urban school districts. The study is measuring differences in teacher quality for  disadvantaged students compared to students of higher socioeconomic status. We are also looking at the level of equity in the distribution of effective teachers in light of district-level policies.


"Sensitivity of Teacher Value-Added Estimates to Student and Peer Control Variables" (November 2013)
"Accounting for Co-Teaching: A Guide for Policymakers and Developers of Value-Added Models" (October 2013)
"Elementary School Data Issues: Implications for Research Using Value-Added Models" (October 2013)
"Measuring Teacher Value Added in DC, 2012-2013 School Year" (September 2013)
"Classroom Observations from Phase 2 of the Pennsylvania Teacher Evaluation Pilot: Assessing Internal Consistency, Score Variation, and Relationships with Value Added" (May 2013)
"Shrinkage of Value-Added Estimates and Characteristics of Students with Hard-to-Predict Achievement Levels" (April 2013)
"Does Tracking of Students Bias Value-Added Estimates for Teachers?" (March 2013)
"Rewarding Physicians for Their Patients' Health Outcomes: What Can Medicare Learn from Education's Value-Added Models?" (June 2012)
"Value-Added Estimates for Phase 1 of the Pennsylvania Teacher and Principal Evaluation Pilot" (April 2012)
"Value-Added Models for the Pittsburgh Public Schools" (February 2012)
"Estimating Teacher and School Effectiveness in Pittsburgh: Value-Added Modeling and Results" (November 2010)
"Teacher and Principal Value-Added: Research Findings and Implementation Practices" (September 2010)
"Error Rates in Measuring Teacher and School Performance Based on Student Test Score Gains" (July 2010)
"Measuring School and Teacher Effectiveness in the EPIC Charter School Consortium: Year 2" (October 2009)

Other Publications

"Value-Added: It's Not Perfect But It Makes Sense" (December 2010)