Biased Judgment in Censored Environments

• 成果类型: Article
• 署名作者: Feiler, Daniel C.; Tong, Jordan D.; Larrick, Richard P.
• 署名单位: Dartmouth College; University of Wisconsin System; University of Wisconsin Madison; Duke University
• 刊物名称: MANAGEMENT SCIENCE
• ISSN/ISSBN: 0025-1909
• DOI: 10.1287/mnsc.1120.1612
• 发表日期: 2013
• 页码: 573-591
• 关键词: inference; heuristics and biases; demand estimation; inventory decisions; risky choice; learning

摘要

Some environments constrain the information that managers and decision makers can observe. We examine judgment in censored environments where a constraint, the censorship point, systematically distorts the observed sample. Random instances beyond the censorship point are observed at the censorship point, whereas uncensored instances are observed at their true value. Many important managerial decisions occur in censored environments, such as inventory, risk taking, and employee evaluation decisions. In this research, we demonstrate a censorship bias-individuals tend to rely too heavily on the observed censored sample, biasing their belief about the underlying population. We further show that the censorship bias is exacerbated for higher degrees of censorship, higher variance in the population, and higher variability in the censorship points. In four studies, we find evidence of the censorship bias across the domains of demand estimation and sequential risk taking. The bias causes individuals to make costly decisions and behave in an overly risk-averse manner.