Incentives for Developing and Communicating Principles: A Reply

The commentators raised many interesting ideas in response to Armstrong and Pagell (2003), from which one general theme emerges: The commentators claim that management science lacks the incentives to encourage efforts to develop and communicate grounded principles. As a result, academics often conduct their research as an intellectual exercise with little concern as to whether their findings might eventually be of any practical use. The problem extends beyond management science. Smith (1991), an editor of the British Medical Journal, concluded from a review that only about 15 percent of medical interventions are supported by solid scientific evidence. He attributes this disconnect to an estimate that only about one percent of articles in medical journals are scientifically sound. Such results indicate problems with incentives in research. Comments Postprint version. Published in Interfaces, Volume 33, Issue 8, 2003, pages 109-111. Publisher URL: http://interfaces.journal.informs.org/ The author asserts his right to include this material in ScholarlyCommons@Penn. This journal article is available at ScholarlyCommons: http://repository.upenn.edu/marketing_papers/100 Incentives for Developing and Communicating Principles: A Reply J. Scott Armstrong The Wharton School, University of Pennsylvania, Philadelphia, PA Email: [email protected] Fax: 215-898-2534 Phone: 610-622-6480 The commentators raised many interesting ideas in response to Armstrong and Pagell (2003), from which one general theme emerges: The commentators claim that management science lacks the incentives to encourage efforts to develop and communicate grounded principles. As a result, academics often conduct their research as an intellectual exercise with little concern as to whether their findings might eventually be of any practical use. The problem extends beyond management science. Smith (1991), an editor of the British Medical Journal, concluded from a review that only about 15 percent of medical interventions are supported by solid scientific evidence. He attributes this disconnect to an estimate that only about one percent of articles in medical journals are scientifically sound. Such results indicate problems with incentives in research. Disincentives The discovery of new and useful grounded principles carries the message that the currently accepted procedures may not provide the best solutions. Some forecasting procedures, such as Box-Jenkins (Armstrong 2001) and game theory (Green 2002), have been shown to be of little value. Other procedures, such as unit roots testing in econometric forecasting, have little empirical grounding to demonstrate their value (Allen and Fildes 2001). Some researchers may be annoyed by these conclusions, especially given that they are based on empirical evidence. The history of science shows that many disincentives to scientific advances exist. Those