UIUC Quant Brownbag

Updates from the Frontiers of Test Validity Theory

Keith A. Markus - CUNY City University of New York

The first edition of Frontiers of Test Validity Theory (Markus & Borsboom, 2013) divided open questions in test validity theory into three main topics: measurement, causation and meaning. Measurement is important because many tests seek to measure some quantitative attribute. Causation is important because validation supports explanations of test scores. Meaning is important because test score use rests on interpretation of test scores. This talk elaborates on several updates from the second edition. The first update involves the impact of a turn toward consequentialism underlying the forthcoming validity and validation chapter in the next edition of Educational Measurement. The second update involves a shift in focus regarding theories of causation. Different kinds of causal claims about scores require different validity evidence. The third update focuses on contributions toward a grammar of validity arguments. Test developers use validity arguments to organize their validation efforts and developing a grammar for them holds the potential to provide guidance that is both concrete and generalizable while also suggesting some blind spots in current practice.