Inattention in Multi-Attribute Search : An Experiment

Humans do not have infinite attention. Contrary to what traditional economic models would predict, only a subset of all available alternatives is considered for most decisions one makes in a lifetime. On top of this, only a limited number of attributes of these alternatives are taken into account. We study how choices between alternatives and the associated search behavior among their attributes change when an optimal search strategy is communicated. We do so by implementing a treatment manipulation targeting the amount of search and studying inattention as a moderator. Our experimental study expands the recent work on inattention. We design and implement an online search experiment with a representative UK sample, where we proxy attention levels. We study the relationship between these levels and the amount of search and investigate if we can adjust the amount of search toward a communicated optimal strategy. We find that search behavior is better described with partial attention. Additionally, the estimated levels of inattention depend on the attribute’s importance—-the more important the attribute, the higher the level of attention allocated. Secondly, our intervention has a converging effect on search behavior, where so-called "over-searchers" diminish their search while "under-searchers" increase their search levels towards the communicated optimum. Lastly, we find that estimated inattention levels have a moderating role in the shift of search behavior for the group of over-searchers