Keith A. Markus, The City University of New York (CUNY)

Four Takes on Construct Validity: From Logical Empiricism to Varieties of Scientific Realism

Invited Speaker

The idea of construct validity or construct validation has been with us for at least 70 years (Cronbach & Meehl, 1955) and has received different formulations.  The original formulation emerged in the context of logical empiricist philosophy of science.  Samuel Messick (1989) and others recast construct validity (as the whole of validity) in terms of realist philosophy of science.  Kathlene Slaney (2017) has proposed shifting the focus of construct validity away from the attribute being measured and onto the linguistic term used to refer to it.  Brian Haig (2023) proposed placing more emphasis on theory evaluation.  Each in their own way, Slaney and Haig seek to separate constructs from psychometric validation practices.  Nonetheless, while presentations of various views tend to emphasize difference and discontinuity, there are many common threads binding these different views together as parts of the same research program.

about the speaker

Keith A. Markus is professor of psychology at John Jay College of Criminal Justice of The City University of New York and affiliated with the Graduate Center of The City University of New York.  He runs the The Lady Mary Shepherd Laboratory for Causation, Methodology and Measurement, named in honor of the Scottish philosopher’s contributions to causation.  His primary research interests involve test validity theory and causal inference/explanation in the behavioral sciences.  He has explored alternative accounts of how test validity can come in degrees.  He has also proposed a tripartite distinction between three kinds of equivalent models.  His work on causal explanation focuses on integrating theories of causation with methodological approaches to causal inference.  He has also proposed an incremental approach to causal inference in which a research program moves from weaker causal claims to stronger causal claims over time.  He also has an interest in the role of normative claims in science including applications in test validity and program evaluation.  He teaches courses in psychometrics, data analysis with R and Python, structural equation modeling, program evaluation research, and other topics.  He is coauthor with Denny Borsboom of Frontiers of Test Validity Theory: Measurement, Causation and Meaning.

Log in