Premium
Learning to Decide with and without Reasoning: How Task Experience Affects Attribute Weighting and Preference Stability
Author(s) -
Leisti Tuomas,
Häkkinen Jukka
Publication year - 2018
Publication title -
journal of behavioral decision making
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.136
H-Index - 76
eISSN - 1099-0771
pISSN - 0894-3257
DOI - 10.1002/bdm.2063
Subject(s) - task (project management) , consistency (knowledge bases) , weighting , stability (learning theory) , preference , cognitive psychology , perception , contrast (vision) , psychology , preference learning , artificial intelligence , social psychology , computer science , machine learning , statistics , mathematics , neuroscience , medicine , management , economics , radiology
Certain experiments have shown that reasoning may weaken the stability of people's preferences, especially with regard to well‐learned perceptual judgment and decision‐making tasks, while learning has an opposite, consistency‐enhancing effect on preferences. We examined the effects of these factors in a visual multi‐attribute decision‐making task where reasoning, in contrast, has been found to benefit judgments by making them more stable. The initial assumption in this study was that this benefit would be typical for novel tasks, like the one employed here, and that it would decrease when the task is thoroughly learned. This assumption was examined in three experiments by contrasting it with an alternative assumption that this previously obtained beneficial effect is caused solely by learning, not by reasoning. It was found that learning indeed makes preferences more stable by consolidating the weights of the attributes. Reasoning, however, does not benefit this task when it is completely novel but facilitates learning and stability of the preferences long run, therefore increasing the consistency of the participants in the macrolevel. Copyright © 2017 John Wiley & Sons, Ltd.