Information‐seeking when information doesn't matter

• Published in: Journal of behavioral decision making Vol. 35; no. 5
• Main Authors: Hilchey, Matthew D., Rondina, Renante, Soman, Dilip
• Format: Journal Article
• Language: English
• Published: Chichester Wiley Periodicals Inc 01.12.2022
• Subjects: attention; effort; finance; Financial loss; ostrich effect; Portfolios; Return on investment; uncertainty; Utility theory; visual search
• ISSN: 0894-3257; 1099-0771
• DOI: 10.1002/bdm.2280

Prior research shows that investors check their portfolios less frequently when they believe negative returns on investments are likely. This so‐called ostrich effect is accounted for by belief‐based utility theories that suggest that information demand is determined by the potential of information to evoke or maintain pleasant beliefs. An alternative is that information matters as there are more courses of action that investors would take with positive than negative portfolio returns. Across three experiments, we adapt a non‐instrumental sampling paradigm to verify whether people are more likely to seek out information when expecting small gains as compared to losses when the instrumental utility of outcome information is 0. We also explore whether the effect can be attenuated by making outcome information easier to find. Our findings suggest that people are more likely to initiate and persist in search for prospective financial gains than losses, and for unknown than known financial outcomes, even though confirming any given outcome is effectively useless. The magnitude of the ostrich effect increased with increased certainty of a financial gain and loss. Making outcome information easier to find increased the likelihood that an outcome would be discovered but did not strongly modify intent to seek it out or moderate ostrich effects. We discuss how the findings are consistent with non‐instrumental utility frameworks for information demand, inconsistent with literature showing greater attention to financial loss than gain outcome information, and propose testable hypotheses for resolving the discrepancy.