We examine the decision maker’s external environment as a predictor of her judgmental effectiveness in detecting regime changes. Relying on two experimental studies, our main task employed is one of regime-change detection over time. Our first study looks at the effect of different task factors (signal diagnosticity, transition probability, and signal length) on judgmental accuracy, particularly with regards to over- and under-reaction. We find evidence of systematic over- and under-reaction, further supporting what is known as the system-neglect hypothesis (Massey & Wu 2005). We also establish that signal length is a significant component in a regime change detection task and show that increasing amounts of information are related to increasing conservatism. Yet what has not received adequate attention in studies of regime change detection is the effect of the strength or extremeness of evidence. We hypothesize that signals that are highly representative of a change having occurred (a perception we anchor on streaks) are likely to lead to systematic over-reaction. Our second study examines the effect of signal streaks on judgmental effectiveness. We report a significant three-way interaction suggesting that streaks tend to govern over-reaction in regime change detection tasks.

Additional Metadata
Keywords judgemental forecasting, regime change, system neglect, streaks, under-reaction, over-reaction
Federspiel, F, Seifert, M., & Newman, L. (2013). Regime Change Detection – Task Factors as Drivers of Judgmental Effectiveness.