Home | Business News | Browse by Publication | J | Journal of the American Statistical Association

Causality: Models, Reasoning, and Inference.

Article Excerpt
Causality: Models, Reasoning, and Inference.

Judea PEARL. New York: Cambridge University Press, 2000. ISBN 0-5217-7362-8. xvi + 384 pp. $39.95.

Those readers familiar with causality will already know the earlier work of Pearl (1988), a text on probabilistic reasoning in the framework of artificial intelligence that helped define a broad research agenda for more than a decade. Pearl's pursuit of many aspects of that agenda has led to numerous new developments, most of which have already been presented in various journals and forums. This book synthesizes and unifies the developments that treat causality, and links them with related progress by statisticians, philosophers, computer scientists, psychologists, and others. We now have a single volume that brings much of this corpus together, to reflect on where we are at and to make it easier to bring this material into classes and seminars. Of course, Pearl has his own particular approach, and others have somewhat different methods; however, the debates among them are exactly what the field needs as it matures and sorts itself out. Some of these can be followed on a web forum set up by the author at http://bayes.cs.ucla.edu/BOOK-2K/discussion.html.

This review is directed more toward a second and broader audience, namely those readers of this journal who were raised hearing the constant admonition "don't deduce causality" and who have been passing it on mantra-like to the next generation in many statistical contexts. What we usually want in investigating the connections between physical, social, political, or economic phenomena is precisely an understanding of...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!