Adapting to information search costs in sample-based decisions

When searching for information before making a decision, one faces a fundamental trade-off. On the one hand, more information will almost always improve the decision outcome. On the other hand, spending more time, money, or effort on further information search may be more costly than that improvement is beneficial. If decision makers are adaptive, they should use their resources in information searches efficiently. Using a sampling-based decision paradigm with financial costs for each observation, we conducted five experiments to investigate whether and how participants adapt their information amount to relevant aspects of the environment. Experiment 1a (N = 67) and 1b (N = 79) focused on changes in response to different ratios of information cost and decision payoffs, revealing that participants are capable of adapting, but seem to do so mainly based on a priori planning. Experiment 2 (N = 131) assessed the role of a second process, online adjustment, by varying whether participants received feedback. Surprisingly, participants changed their behavior to the same extent irrespective of whether they received feedback. To further improve the opportunities for metacognitive monitoring and control, Experiments 3a and 3b (N = 160 and N = 318) implemented different feedback types across more decision trials for three different ratios of information cost and decision outcomes, which stayed constant per participant. Whether participants changed their sample size and efficiency across trials depended on that ratio, but not on the type of feedback. This provides crucial insights into how participants adapt and what challenges might befall the necessary metacognitive monitoring, serving as a valuable prompt for future research.