Project ranking using partial ranks

A competition was recently held for new elder care technologies and a method was needed for selecting an audience favorite project. Since each of the eight projects received a half hour presentation time slot spread out over a full day, attendance varied significantly with morning projects having a much lower attendance. An algorithm was desired that was robust with respect to varying evaluator. A new algorithm for aggregating ranks from a large number of incomplete judgment was developed and applied to the projects to select a winner. This paper presents the new algorithm, tests the new approach against others in the application, and discusses relative tradeoffs.