Skip to Content

Department of Cognitive Science

Seminar Abstract

Belief bias in syllogistic reasoning: A meta-analysis of ROC data.

Speaker : Dr Dries Trippas, Center for Adaptive Rationality, Max Planck Institute for Human Development.
Date : 22nd of February 2017, 11:00AM until 12:00PM
Location : Australian Hearing Hub, Level 3, 3.610, Macquarie University.

    The belief bias effect is one of the most-studied biases in reasoning. A recent analysis of the effect using the signal detection theory (SDT) model (Dube, Rotello, & Heit, 2010) has called into question previous theoretical accounts of belief bias by demonstrating that presumed differences in reasoning ability could be an artifact stemming from use of an inappropriate linear measurement model. Specifically, this linear measurement model tacitly assumes linear and symmetric receiving operator characteristic (ROC) curves, which are not in line with the curved and asymmetric ROCs obtained with confidence-rating data. When described by the SDT model, these ROCs indicate that belief bias can be interpreted as a simple response-bias effect. Moreover, this controversial result has also raised questions on the informativeness of much of the research on belief-bias conducted in the last four decades, implying extensive replications including ratings data may be necessary. The present work argues that the latter conclusion may not be valid, and raises questions regarding prior reliance on data aggregated across participants and stimuli. Using a hierarchical-Bayesian meta-analysis that does not rely on aggregation, we reanalyzed a corpus of twenty-two studies with almost 1000 participants. Results corroborated the notion that belief bias can be described as a response-bias effect, but also indicated that data from previous studies can be safely reanalyzed using appropriate methods instead of being discarded. On a more general level, our results suggest a new standard for analyzing data and evaluating theories in this domain as well as important methodological and theoretical considerations for future work on belief bias.

    Further Information

    Who is Visiting

    Contact Details

    Telephone: (02) 9850 9599
    Fax : (02) 9850 6059
    Email : cogsci@mq.edu.au
    Web : www.cogsci.mq.edu.au