Policy capturing in small-group decision making: Using linear programming to assign factor weights for the NCAA Selection Committee's ranking decisions in the 2016 through 2018 men's basketball tournaments

摘要

We present a general approach for applying policy capturing to small-group decision making and demonstrate the approach with the NCAA Selection Committee's rankings of the 68 teams in the three most recent Division I Men's Basketball Tournaments. We develop a linear programming model that minimizes the extent to which the evaluation of a lower-ranked team exceeds that of a team ranked immediately above it. The result is a set of linear weights that approximates a group decision when the decision itself is demonstrably internally inconsistent in that a group ranks some teams higher than others despite a lower implied evaluation.