Constituency-level Data Research Articles

Voter turnout at the british general election of 1992: rational choice, social standing or political efficacy?

Abstract. Analyses of turnout in British general elections fall broadly into two camps: those based on constituency–level data, and those based on survey data. The former stress the importance of local context, while the latter stress personal characteristics and viewpoints. Underlying both are a range of theories purporting to explain turnout. However, to date, there has been little systematic attempt to analyse turnout in the round. In this paper, we combine survey and constituency data to study the individual and contextual correlates of turnout at the British general election and 1992. Constituency level analyses seem to confirm the importance of local context, though it declined during the 1980s. However, and contrary to analyses which employ constituency data only, while individual electors’ decisions on whether or not to turn out are influenced by their personal circumstances, they are not influenced by local context.

Where to count parties

To evaluate theories about the impact of electoral law on party systems, it is useful to have an indicator of exactly how many parties are competitive in a given election. Several formulae convert the vote shares won by multiple parties into single number-of-parties indices. While these formulae are typically applied to national vote shares, it is often more revealing to examine constituency-level data. For some purposes, the calculated value for the number-of-parties-in-the-nation is a poor substitute for a fuller description of how many parties or candidates exist, constituency by constituency.

Family Leave Legislation in the U. S. House: Voting on the Family and Medical Leave Act of 1990

Family and medical leave legislation has been one of the most hotly debated topics in family and social policy in recent years. Since 1985, when Representative Patricia Schroeder (D-Colorado) and her colleagues introduced the first family leave bill in the U.S. Congress, supporters of such policies experienced notable gains and setbacks. In the mid-1980s, few (if any) states had established family leave policies. Now, as many as half the states have some form of family leave law, and nearly all the states have introduced family leave proposals into their legislative processes. In Congress, there has been a steady march toward passage of family leave legislation. Congress passed the Family and Medical Leave Act (FMLA) of 1990 in the early summer of that year, only to have the legislation vetoed by President Bush on June 29, 1990. In late 1991, a similar measure was held in conference committee so that supporters of the bill could force the President to veto the bill in an election year (see Schroeder, 1991). Finally, in early 1993 both houses of Congress passed the Family and Medical Leave Act, and in February the bill became the first piece of legislation signed into law by President Clinton. Although Clinton's signature closed the book on national family leave legislation, at least in the short run, there could be no doubt that family and medical leave had become an issue of high political visibility. As the family leave policy neared its implementation date in August 1993, debate over various provisions of the law made it clear that the topic is far from settled. The topic of family leave has not been ignored by the scientific community. The need for programs that help employees balance work and family demands has been discussed widely (e.g., Hopper & Zigler, 1988; Zigler & Frank, 1988). The political environment and process likely to result in passage of family leave laws at the state level have been documented (Garand & Monroe, 1994b; Wisensale, 1990). Arguments for and against family leave laws have been analyzed (Haase, 1988; Lenhoff & Becker, 198, and the costs of failing to adopt such laws have been reported (Spalter-Roth, Hartmann, Gibbs, Lenhoff, & Stoneback, 1989). In this article, we utilize a theoretical model developed in previous research to explain and predict roll call votes in the U.S. House of Representatives on the FMLA of 1990. We previously examined the votes of U.S. Senators on debate surrounding the Parental and Medical Leave Act of 1988 (Monroe & Garand, 1991). That model had excellent predictive capability, but the vote examined was a cloture vote (i.e., a vote to end a filibuster, rather than a vote on final passage of the legislation). While there was sound logic for examining a cloture vote as a proxy for a substantive roll call vote on family leave legislation, the lack of a recorded ballot on final passage for that legislation was a limitation. Next, we tested a similar model to explain adoption of family leave laws at the state level (Monroe & Garand, 1989; Garand & Monroe, 1994b). Again the model provided an excellent fit to the data, but the test of the model was limited because so few states had adopted such a law and because the content of the laws varied considerably from state to state. The present article is not bound by the above limitations. The FMLA was a very public vote on a widely publicized piece of legislation. Supporters and opponents of the legislation were operating in full force. President Bush made his veto intentions clear. Moreover, in this study we are unconstrained by the limited number of cases inherent in any study of the U.S. Senate or the American states. The size of the House membership (N = 435) provides for better estimates of the effects of relevant independent variables and sufficient degrees of freedom for the number of variables in our model. Finally, measures of the various independent variables can be finely developed because constituency level data are available at the House district level. …

Constituency variations in aggregate voter behavour across the 1965–1983 German federal elections

Two important assumptions underlying the Riker—McKelvey—Ordeshook (RMO), expected-utility-based, multicandidate calculus of voting are that citizens decide whether or not and for whom to vote in a single-member-district (SMD). plurality-winner electoral system and that they ignore aggregate considerations (for example, outcomes in other districts). Aranson modifies this calculus by investigating citizen behavior in proportional-representation (PR) systems and situations in which aggregate outcomes require parties to form coalitions in order to govern. The RMO calculus and Aranson augmentations are both concerned with two basic questions: What stimulates citizens to vote? If they vote, for whom do they vote? In this study, aggregate voter behavior is modeled as a set of equations with two endogenous variables related to these questions. Equation parameters are estimated with pooled, cross-sectional, time-series, constituencylevel data on six German elections, which combine features of SMD and PR systems and are characterized by coalition politics. After appropriate controls are included and within the constraints imposed by this data, evidence is found that supports the relevance of the RMO calculus and Aranson augmentations, especially the latter, in accounting for observed variations in the endogenous variables.