Quantitative Judgment Aggregation for Evaluating Contestants

Quantitative judgment aggregation is a new research topic in (computational) social choice. In this model, agents judge the relative quality of different entities, and from this we aim to obtain an aggregate assessment of their relative qualities. As is the case for the more traditional voting-by-ranking model in social choice, there are corresponding statistical problems. For example, a race provides a “judgment” of the relative abilities of the contestants, and we may wish to aggregate the results of multiple races. In this context, we no longer need to worry about, eg, strategic misreporting by the judges, which enlarges the space of possibilities for aggregation rules. We investigate this larger space theoretically and evaluate on data from various real races.