GOING BEYOND WITH BAYESIAN UPDATING

• Published in: Journal of applied behavior analysis 1992, Vol.25 (3), p.585-597
• Main Authors: O'Flaherty, Brendan, Komaki, Judith L.
• Format: Article
• Language: English
• Subjects: Performance Management for Business and Industry
• ISSN: 0021-8855; 1938-3703
• DOI: 10.1901/jaba.1992.25-585

It all started innocently enough. One spring‐like winter day, I happened to ask Brendan whether economists ever dealt with escalation. “With what?” he replied. “A phenomenon where people keep investing in the face of continuing losses,” I answered. Then I described how industrial/organizational psychologists had become intrigued with situations in which investors seemed to throw good money after bad, how their explanations for the phenomenon centered on individual characteristics such as commitment, how Sonia Goltz (1992) used a standard bread‐and‐butter operant procedure—fixed and variable schedules of reinforcement—to explain their persistence, and how Goltz's experiments had shown that during the extinction phase investors even increased their investments for a while when the news was all bad. Without a moment's hesitation, he exclaimed, “I bet I can predict the turning point.” Another arrogant‐economist remark, I thought to myself. “How?” “Bayesian updating.” Then we talked at length about Bayesian analysis techniques and how they could be used to predict the shape of extinction curves. I realized that these techniques might be just what psychologists needed as an enticement to study sequences of behavior over time. And that's how this commentary got started.