In our lead up to #ICA25, we are providing information about papers that received Top Paper awards from the Communication Science and Biology (CSaB) Interest Group. Each paper received exceptionally high scores from reviewers. These papers reflect outstanding scholarship in CSaB. Today’s Top Paper features Overbye-Thompson, Hamilton, & Fisher’s paper: “Expectation of Algorithm Bias Increases Caution: Implications for Human-in-the-Loop Decisions”. Be sure to check out the paper at #ICA25
In a few short sentences, what is your study about? Algorithms increasingly mediate our interactions with the world. Growing evidence indicates that these algorithms are prone to bias, favoring certain outcomes over others. However, little is known about whether humans overseeing these algorithms (known as a Human-In-the-Loop) can detect algorithmic bias or what factors may influence this ability. This study investigates how individuals perceive and respond to potential algorithm bias in a simulated loan approval process, employing drift-diffusion modeling (DDM) to analyze decision-making.
How did you come up with the idea for this line of research?: I (Hannah) had already been exploring the impacts of algorithmic bias on technology use and adoption when my advisor, Kristy Hamilton, encouraged me to approach the issue through the lens of decision-making literature. She introduced me to Jacob Fisher, who taught me about drift diffusion modeling during the 2023 ICA Hackathon. The timing was ideal, as recent legislation is increasingly mandating human oversight in high-risk algorithmic decisions, which led me to become particularly interested in whether people can detect errors in algorithmic outputs, an area where the drift diffusion model provides an especially effective framework.
Tell us more about the team:
Hannah Overbye-Thompson: Fourth-year PhD candidate at the University of California, Santa Barbara, loves computational methods, decision making, and thinking about algorithmic bias
Kristy Hamilton: Assistant professor at the University of California, Santa Barbara, cognitive science expert, advisor extraordinaire
Jacob Fisher: Assistant professor at Michigan State University, expert on all things computational methods, and has a near-magical ability to make Google Cloud and Docker run seamlessly