Purchase this article with an account.
Glen Gagnon, Chad Peltier, David Johnson, Mark Becker; Dangerous Ambiguity: Capturing Shooter Bias with an Eye Tracker. Journal of Vision 2017;17(10):544. doi: 10.1167/17.10.544.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
When making rapid, uncertain decisions about danger, categorical heuristics are utilized to process visual information. In particular, race and racial stereotypes influence visual perception, impacting accuracy and reaction times. This research aims to clarify what sources of information a person uses in making decisions about danger. Our study examines stereotypic vision as an extension of Correll et al.'s (2015) work on how participants use racial information to process ambiguous visual stimuli. Participants completed a first-person-shooter task in which they made rapid decisions (shoot or don't shoot) for Black and White targets holding either a gun or harmless object. While participants viewed images in quick succession, an eye tracker collected data on eye movements relative to two interest areas: the target's face and the object in the target's hand. In alignment with the previous literature, we discovered that reaction times and error rates signal stereotypic racial bias. Responses for Black targets have higher error rates, faster reaction times for gun trials, and slower reaction times for non-gun trials. Moreover, additional eye tracking measures, beyond the viewing angle proposed by Correll et al., provide insight into how participants gathered information. Overall, we found that participants relied on ambiguous information to make decisions for stereotypic targets and sought more certainty for counter-stereotypic targets. When viewed collectively, our data supports Correll's findings of stereotypic influence. Measures of dwell time and refixations on faces and guns suggest initial fixations are more ambiguous for Black targets. However, subsequent fixations become more explicit, mitigating racial bias in several contexts
Meeting abstract presented at VSS 2017
This PDF is available to Subscribers Only