Roll calls are votes taken by a legislative chamber in which the names of legislators are recorded along with their votes on a question. Rules and practices surrounding roll call votes vary across legislatures because of differences in constitutions, chambers rules, actions of legislative leaders, political cultures, legislature sizes, and the technologies used to record votes. Because they are highly visible, roll call votes are useful to social scientists who research legislative behavior. Techniques used to study roll call votes range from simple descriptions to sophisticated scaling techniques adapted from psychometrics and econometrics.
The most thoroughly studied roll call record is that of the U.S. Congress. Article I, section 5 of the U.S. Constitution requires that “the Yeas and Nays of the members of either House on any question” shall be entered in the journal upon the demand of one-fifth of those present. Through the end of the 109th Congress (2006), the U.S. House of Representatives and Senate together had recorded nearly 94,000 roll call votes.
The first social scientific use of roll call votes is attributed to A. Lawrence Lowell’s 1902 essay “The Influence of Party upon Legislation in England and America.” Lowell introduced the idea of the “party vote,” defined as a vote in which at least 90 percent of one party in a legislature voted against at least 90 percent of the other party. An important extension of Lowell’s idea was suggested by Stuart Rice in 1928 through his “index of cohesion” and “index of likeness.”
Efforts by the liberal interest group Americans for Democratic Action (ADA) in the 1940s encouraged the use of roll call votes to summarize the voting tendencies of individual legislators. The ADA began publishing an annual list of “key votes” in Congress, reporting how often each member of Congress (MC) supported the ADA position. These reports allowed one to array MCs along a continuum, ranging from the most liberal, with a “liberal quotient” of 100 percent, to the most conservative, at a quotient of 0 percent. Ever since the ADA pioneered this technique, dozens of groups have developed “support scores” that reflect their own agendas.
Interest-group support scores are easy to calculate and intuitive to understand, but they have serious shortcomings. Interest-group ratings tend to be based on a small number of roll call votes, thus discarding the rich information that the omitted roll call votes can provide. The roll call votes chosen are typically unusually divisive, which causes the resulting ratings to be too extreme. These techniques also require a priori agreement about which votes best reveal the underlying ideological dimension being described.
Beginning in the 1980s social scientists started exploring ways to overcome these deficiencies, taking advantage of advances in measurement theory and computational capabilities. A frequently used example of this is a class of techniques pioneered by Keith Poole and Howard Rosenthal, called NOMINATE (for Nominal T hree-step E stimation). NOMINATE scores can be constructed from a large number of roll call votes analyzed simultaneously across a long period of time. Ratings from such techniques as NOMINATE show that a single left-right ideological dimension underlies most roll call behavior across American history, that a second dimension of choice sometimes emerges that corresponds to race-related issues, and that individual legislative voting behavior is highly stable over time.
SEE ALSO Congress, U.S.; Conservatism; Constitutions; Interest Groups and Interests; Left and Right; Liberalism; Political Science; Voting
Lowell, A. Lawrence. The Influence of Party upon Legislation in England and America. Annual Report of the American Historical Association for 1901 1 (1902): 321–542.
Oleszek, Walter J. 2007. Congressional Procedures and the Policy Process. 7th ed. Washington, DC: CQ.
Charles Stewart III