Causality In Crisis?: Statistical Methods & Search for Causal Knowledge in Social Sciences

In the past fifty years statisticians and methodologists in the social sciences have developed and refined a family of closely related statistical methods for the study of social phenomena. While the value of such methods of analysis is universally acknowledged, their use has never been wholly uncontroversial. In 1993 prominent scholars from a variety of disciplines (social sciences, statistics, philosophy of science) gathered at the University of Notre Dame to debate whether causal modeling techniques old or new can really justify the drawing of causal conclusions on the basis of correlational statistical data. The resulting volume from that groundbreaking conference is Causality in Crisis? a comprehensive and sophisticated introduction to perhaps the most important set of issues confronting social scientific researchers in the 1990s and beyond. In the essays presented here contributors critically reassess the widely accepted view that statistical methods of analysis can and do yield causal understanding of social phenomena. Although a number of technical issues receive attention, the overall emphasis is on the larger historical, philosophical, and conceptual perspectives that underlie and inform current methodological controversies. The debates in Causality in Crisis? have far-ranging implications, for on their resolution hinges the question of what sort of knowledge of social life it is possible to achieve on the basis of non-experimental social scientific research. Any scholar who makes use of causal methods, as well as all who are affected by decisions reached on the basis of such methods, will have a stake in the challenging arguments put forth in this volume.

About the Author :
Vaughn R. McKim is Emeritus Associate Professor of Philosophy and The Graduate Program in History and Philosophy of Science, and was the Acting Director of the Reilly Center for Science, Technology, and Values at the University of Notre Dame. Stephen P. Turner is a Distinguished Professor of Philosophy at the University of South Florida.

Review :
"The present book evaluates a striking new claim to provide the means for causal inference from statistical association. Readers can get a quick overview, or that plus a tutorial-like introduction to the statistical principles underlying the SGS algorithm, move on to discussions about the pros and cons of the method, and end with a deep understanding of the difficult issues that have surfaced here. And, what will prove most satisfying to the historically minded readers of JHBS, the endeavor is placed in a historical context that illuminates the nature of the issues at hand. It is refreshing to find an exception, an edited book with a consistent theme, an organization that encourages reading from beginning to end...Readers who take the time to do this will be rewarded with a new perspective on some old questions....the present book makes clear that the difficulties of inferring causation from correlational data are very much with us still. It is a pleasure to recommend this book to readers interested in opening the door to this fundamental issue in social science, whether in the form of the most recent statistically sophisticated approaches, or to the very first attempts to grapple with it." —Journal of the History of the Behavioral Sciences, "This is a collection of essays by a distinguished group of authors that is a 'must read' for those interested in causal modeling." —Philosophy in Review ". . . an exceptionally well written treatment of the current crisis in sociological methodology, with rich and lucid discussions, particularly by the editors, Vaughn McKim and Stephen Turner." —Social Forces "[A]n attempt to set out what the problems with contemporary statistical methods are, what solutions are being proposed, and to open up the debates about their effectiveness to a wider audience." —Social Studies of Science "This is a collection of essys by a distinguished group of authors that is a 'must read' for those with an interest in causal modeling." —Piers Rawling, University of Missouri-St. Louis