The Effects of Information Choice and Information Use in Analysts' Predictions of Municipal Bond Rating Changes.
研究分析师在预测城市一般义务债券评级变化时，信息选择和信息使用对预测质量的影响。分析师从菜单中自选信息，或使用统计模型选定的信息进行预测，结果显示人类表现与统计模型相当。
Abstract The research reported here investigates the effects of information choice and information use on the quality of municipal financial analysts' predictions concerning changes in the general obligation bond ratings of cities. Analysts were given a menu from which they chose information; a statistical model of bond rating changes was constructed using the same information. Analysts predicted bond rating changes using either their self-selected information or information selected by the statistical model. Although previous research concluded that information choice was the main factor in humans' sub-optimal prediction achievement, the results in this study show that humans performed about as well as statistical models in a validation sample. Previous research was reinterpreted in light of the current results.