At the Crossroad of Research: The Impact of Projection Bias on the Decision to Explore or Exploit Ideas

Focusing on projection bias and overconfidence/underconfidence, we demonstrate the effects of projection bias on the decision of a researcher to delve deeper into the current research idea or explore a new one in infinite periods. We consider a researcher who has a belief regarding whether other researchers have already come up with the same idea. Although beliefs dominated by overconfidence/underconfidence are consistent over time, beliefs dominated by projection bias are inconsistent, depending on whether the researcher has obtained a new idea. We show that a researcher with projection bias changes ideas more often than overconfident/underconfident researchers. Moreover, successful researchers are more likely to have projection bias when the required value of the research outcome for success is sufficiently high. Our results have implications for the file drawer problem, which is the tendency of researchers to shelve negative or insignificant results.