Measuring variations in party unity voting: An assessment of agenda effects

Measures of party divisiveness have been widely used in scholarly literature for a variety of different purposes. However, conventional measures of party divisiveness, such as the percentage of party votes in a particular Congress, fail to consider important changes in the agenda from one Congress to the next. We introduce a measure that controls for such changes, drawing attention to the effect that agenda change has on observed party divisiveness and providing a more accurate account of party divisiveness across time. We analyse party voting in the U.S. House of Representatives between 1953 and 2004, and find that a sizeable amount of the fluctuation in party divisiveness identified by conventional measures is mitigated using this method. While we examine party voting in the U.S. House, our theory and method is applicable to similar measures used in the study of other legislatures (e.g. Rice and Attinà indices).