Abstract
Influence of group learning in search processes. Our decisions about the world are often based on a sequential analysis of various uncertain cues. To achieve high accuracy, the order in which these cues are searched is crucial. Previous research, however, has shown that people demonstrate slow progress in learning cue ordering by validity when they could update such cue ordering through feedback. A possible clue to how people solve this problem is group learning: By exchanging information with other individuals, people can learn which cues are relevant for making decisions and in which order the cues should be considered. In a computer simulation study, we showed that exchanging information about which cues are good for making decisions can help to overcome the limitations of individual learning. Thus, the dilemma of individual learning can be collectively solved. In line with the spirit of the fast and frugal heuristics approach, we found that several simple group learning rules performed better than computationally demanding ones. We relate our results to previous findings on bounded rationality in the social context.
