| There exists a large body of literature regarding the organization of the United States Congress and the importance of political parties in shaping legislative outcomes. Ideological scores developed in the last two decades have become increasingly popular in this debate. However I show here that the scores themselves are affected, and most likely biased, by the strength of the majority party. First, I show that the distribution of DW-NOMINATE scores for majority members is significantly different than for the minority party. In Chapter 2, I show that the probability DW-NOMINATE incorrectly predicts a majority party member's vote is significantly lower than for a minority party member's vote. I argue that this is because fewer bills dividing the majority party are likely to reach the floor than those dividing the minority party. Finally, in Chapter 3, I demonstrate one method of using DW-NOMINATE scores that accounts for the party strength bias of the scores. Using this method, I show that majority-party membership is more important for shaping legislative outcomes than is committee membership or ideological centrism. |
