Date of Degree
PhD (Doctor of Philosophy)
Frederick J. Boehmke
Since Woodrow Wilson's (1885) analysis of Congress, researchers assumed that members of Congress look to one another for information, cues, and advice on unfamiliar policy areas. The amount of time and effort that each legislator and their staffers would have to put in to make all of these voting decisions would be insurmountable. Fellow legislators are a resource to turn to for guidance or assistance. Legislators are able to influence their colleagues above and beyond each of their individual preferences. The members of Congress that are most influential will not necessarily be the same for every bill. The significant legislators may be one's co-partisans and the party leadership or they may be a group of legislators with whom they share a common interest. Spatial analysis allows researchers to look more explicitly at the relationships between legislators and their colleagues.
I use spatial probit and a spatial duration model to study these issues by examining the factors that influence voting decisions and the timing of position announcements. I look at a variety of different policy areas, including foreign policy, education, and agriculture, over an extensive time period (1933-2014) to test which relationships are most influential on their decisions. I study the interdependence between three different relationships, same party, state delegation, and ideological similarity, and hypothesize that these ties will lead legislators to behave more similarly. The use of the spatial analysis provides an opportunity to test these relationships and see if even after controlling for other influences there is dependence between legislators. In my research, I find that legislators are interdependent regardless of their individual characteristics. When I analyze voting behavior, legislators' behave similarly from one another across all three relationships above and beyond what we would expect given their personal preferences. These positive findings do not hold when I study the timing of position announcements where legislators behave dissimilarly from one another when interdependence exists. The study, overall, suggests that legislative ties are especially important in explaining voting behavior and that it is critical to account for these relationships.
Legislating is a social process. Legislators coalesce, cajole, and conspire with one another to draft and pass bills. Decades of congressional research emphasize how each legislator's decisions depend on those made by their colleagues. Yet extant empirical methods struggle to capture the resulting interdependent nature of legislators' actions. This disconnect constitutes a major failure of the literature to properly evaluate and test theories of legislative behavior such as the role of parties in binding members' fates together or whether legislators that share common interests based on their race, gender, or other characteristics achieve better outcomes when they work in concert than when they merely act independently.
My dissertation seeks to address this gap through the use of new statistical techniques that provide a more appropriate way to detect theoretically predicted forms of interdependence. Specifically, I use spatial regression techniques to explicitly model and estimate the presence of interdependence between legislators. Spatial econometrics breaks the restrictive assumption that actors' choices emerge independently of one another. Rather, it allows one legislator's choice to explicitly depend on choices made by other legislators. I analyze this interdependence by looking at voting behavior and the timing of position taking announcements. I find that the ties among legislators are critical in explaining voting behavior, leading legislators to behave more similarly to one another.
publicabstract, Congress, Interdependence, Legislative behavior, Spatial econometrics
xi, 109 pages
Includes bibliographical references (pages 91-99).
Copyright 2015 Emily Ursula Schilling