American Statistical Association
Behavioral science methodology tends to pass over questions about the nature of causation and focus directly on procedural issues related to causal inference. In a separate literature, a number of distinct approaches to understanding causation have arisen including regularity, counterfactual, process, and non-reductive approaches. The fact that none has clearly distanced itself from the others has also led to serious consideration of causal pluralism: The idea that more than one type of causation may be at work in what we study. These developments reflect back on behavioral science methodology. Some methods seek to learn about causal facts that exist in the world, whereas others assume that causation only exists as part of the way we represent the world, not in the world itself. Some conceptions of causation are more clearly transitive than others, and this has ramifications for causal mediation. In contrast to causal explanations that explain current states in terms of past states, one can also construct causal explanations that explain the current state of a whole in terms of the current states of its parts. Some forms of causation require that causal effects are the same in within-subject and between-subject data whereas at least one form of causation is consistent with differences between these. Laboratory experiments need not always investigate the same sort of causation as field studies. Developing a richer vocabulary for specifying what one seeks to obtain from a causal inference can help behavioral science methodology develop more fine-grained advice for study design.
Keith A. Markus is Professor of Psychology at John Jay College of Criminal Justice of The City University of New York. His work on causal inference focuses on fine tuning methodology to account for different notions of causation. His work on test validity includes a forthcoming book with Denny Borsboom entitled Frontiers in Test Validity Theory: Measurement, Causation, and Meaning (Routledge) and focuses on strengthening test validation practice with richer theoretical accounts of test validity itself. He has served as Associate Editor of Structural Equation Modeling: A Multidisciplinary Journal and is currently finishing a second term as Associate Editor of Psychological Methods. He has also done research in job satisfaction, organizational culture, student admissions, deception detection, partner violence, severe mental illness, sex offenders, and multi-dimensional scaling for crime scene investigation. He offers courses in psychometrics, data analysis, structural equation modeling, program evaluation, and other areas.
|Date:||Tuesday, October 2, 2012|
|Time:||3:30 - 4:30 P.M.|
New York State Psychiatric Institute
1051 Riverside Drive
Pardes Building Multipurpose Room 6602
New York, New York