Panel: Methodological Issues with Value Added Models

Friday, November 9, 2012: 8:00 AM-9:30 AM
Hanover B (Radisson Plaza Lord Baltimore Hotel)

*Names in bold indicate Presenter

Organizers:  Cory Koedel, University of Missouri-Columbia
Moderators:  Douglas Harris, University of Wisconsin at Madison and Cory Koedel, University of Missouri-Columbia
Chairs:  James Wyckoff, University of Virginia

The papers in this panel address several methodological issues of importance to researchers, policymakers and other stakeholders as states and school districts are increasingly relying on value-added models (VAMs) for evaluation purposes. First, which controls should be included in VAMs, and what are the tradeoffs involved in choosing one set of controls over another? The choice of how to specify VAMs can have implications in terms of which schools and teachers are identified as high and low performers. This, in turn, can influence teacher incentives, particularly with respect to teaching in disadvantaged schools. The question of how to specify VAMs is highly contentious in policy discussions nationwide. Second is the issue of co-teaching. Most value-added studies in the research literature have avoided the co-teaching issue entirely – in many cases by simply dropping co-taught classrooms. However, co-teaching is a common occurrence and in policy applications it must be addressed. Third is the issue of multi-level fixed effects VAMs. In practice, the most common version of a multi-level fixed effects model is one that includes school and teacher effects. An important consideration is that different layers of fixed effects influence how teachers are compared to each other, and require different identifying assumptions. An application to the evaluation of teacher preparation programs is considered.

Does the Model Matter? Exploring the Relationship Between Different Student Achievement-Based Teacher Assessments
Dan Goldhaber1, Joseph Walch1 and Brian Gabele2, (1)University of Washington, (2)Seattle Public Schools

Sensitivity of Teacher Value-Added Estimates to Student and Peer Control Variables
Matthew Johnson, Stephen Lipscomb and Brian Gill, Mathematica Policy Research

Methods for Accounting for Co-Teaching in Value-Added Models
Heinrich Hock and Eric Isenberg, Mathematica Policy Research

Distinguishing Between School Quality and the Effectiveness of Teacher Preparation Program Graduates
Kata Mihaly1, Daniel McCaffrey2, Tim Sass3 and J.R. Lockwood2, (1)RAND Corporation, (2)RAND, (3)Georgia State University

See more of: CrossCutting
See more of: Panel