Achievement

IGERT Trainee Margot Kimura analyzed data from a psychophysics experiment in which groups of human observers perform a fixed-sample signal detection task. She found that the groups used non-optimal weights but still generally out-performed their most capable individual, and favored equality-based rules for determining the group response. These results are relevant to applications involving a visual task, such as cancer detection in mammograms or threat detection in baggage screening.

Margot also analyzed different group decision making schemes for a two-alternative forced choice task and derived explicit analytical solutions for the group performance (error rates and decision time distributions) for three group rules: Race, Majority Total, and Majority First. She compared the rules in terms of performance over a range of group sizes and elucidated the strengths and weaknesses of each scheme.