Understanding the role of computer-mediated counter-argument in countering confirmation bias

摘要

Confirmation bias has long been discussed in the behavioral decision-making research stream. Although decision support systems were designed to counter cognitive biases and speed up information processing, confirmation bias still can be observed during the decision-making process and causes some unwanted behaviors, such as selective reading. An experimental design was conducted to examine the impact of confirmation bias in a computer-supported decision-making context. In addition, we attempted to explore whether the providing of computer-mediated counter-argument can effectively eliminate the impact caused by selective reading. The experiment results show that confirmation bias can be observed when decision makers possess strong preconceptions and selective reading behaviors, caused by confirmation bias, resulting in skewed adjustment and high confidence. This means that computer-mediated counter-arguments can effectively reduce the effects caused by confirmation bias as well as lead to higher satisfaction with the decision outcome. Lastly, the research results were discussed and implications of this finding for academics and practitioners were provided.