ABSTRACT
To perform a statistical test of whether Congressional committees are representative of the chamber from which they are chosen, the measure that is used as a basis of evaluation, the statistic, and the sample space all must be specified. The committee-outlier debate has been concerned only with the committee system in the House during and after the 80th Congress. The many-dimensions view of the world fails to be at all helpful in understanding roll call voting decisions. If preference outliers are of great importance, they should be present among those dimensions that organize roll call voting. A member's one-dimensional D-NOMINATE score is the measure of preference the authors use to assess whether committees consist of preference outliers. In assessing whether a specific committee is a preference outlier, they compute the magnitude of the difference between the committee median and the median of the whole chamber.
