Demand Forecasting, Evidence and Checklists
Introduction
Scott Armstrong seeks published or unpublished work on forecasting effectiveness
Kesten Green and I are near to the completion of an invited paper, . It reviews and integrates experimental findings and creates checklists to enable the use of effective techniques in forecasting. Historically, new discoveries take decades before they are used. For example, judgmental bootstrapping dates back to the early 1900s and it is still seldom used. And the Index method dates back to Benjamin Franklin; To our knowledge it has been used only in parole prediction, election forecasting, and predicting the effectiveness of advertisements.
Some of the conclusions were puzzling to us, and some effect sizes astonished us. This leads us to worry that we may have overlooked experimental evidence that refutes our findings, or at least better defines the conditions under which our findings apply. (Of course, it is also useful to know about supporting evidence that has been overlooked.) Might you be able to tell us about papers by you (or others) that we have overlooked?
J. Scott Armstrong, Professor
The Wharton School, JMHH 747
U. of Pennsylvania, Phila., PA 19104
Home Phone 610-622-6480