Potential sources of invalidity when using teacher value-added and principal observational estimates: artificial inflation, deflation, and conflation

Abstract

Contemporary teacher evaluation policies are built upon multiple-measure systems including, primarily, teacher-level value-added and observational estimates. However, researchers have not yet investigated how using these indicators to evaluate teachers might distort validity, especially when one indicator seemingly trumps, or is trusted over the other. Accordingly, in this conceptual piece, we introduce and begin to establish evidences of three conceptual terms related to the validity of the inferences derived via these two measures in the context of teacher evaluation: (1) artificial inflation, (2) artificial deflation, and (3) artificial conflation. We define these terms by illustrating how those with the power to evaluate teachers (e.g., principals) within such contemporary evaluation systems might (1) artificially inflate or (2) artificially deflate observational estimates when used alongside their value-added counterparts, or (3) artificially conflate both estimates to purposefully (albeit perhaps naively) exaggerate perceptions of validity.