No CrossRef data available.
Article contents
Contested Elections and Voter Turnout in a Local Community: A Problem in Spurious Correlation
Published online by Cambridge University Press: 02 September 2013
Extract
This paper presents an attempt, in a case study, to infer causal relationships between contests for local elective office and voter turnout—an attempt which required, in turn, the use of a recently developed mathematical technique for identifying and eliminating spurious correlations. The method no doubt has wider applications.
The case study involves a small New England town and its annual February town elections over the period 1947–1963. The town was chosen as one of the six affected by the newly established Cape Cod National Seashore. The time period was chosen to cover the interval during which the town was faced with the rapid increase in tourism which transformed it from a declining agricultural town to a service-based town. The policies and methods by which the town dealt with this basic change in its economic structure were of immediate interest, especially in the latter years when the National Park Service became involved in the problems the Cape has faced in this postwar period.
- Type
- Research Notes
- Information
- Copyright
- Copyright © American Political Science Association 1965
References
1 The data for this study were obtained in conjunction with the author's unpublished dissertation, “The Cape Cod National Seashore: A Study in Conflict” (University of Illinois, 1963)Google Scholar. The author wishes to acknowledge special indebtedness to Professors Phillip Monypenny and Bernard Lazerwitz of the University of Illinois and Professor Hubert M. Blalock, Jr. of the University of North Carolina.
2 Another possible quantitative check involves the warrant articles which form the agenda for the town meeting. Ascertaining who backed the articles and how these articles fared over time should give some indication as to whose will tends to prevail. This approach would focus directly on the Board of Selectmen since most articles funnel through their office as they make up the final warrant. Thus it would tend to be a test of the ability of the Board of Selectmen to get their way and not an equal test of all town institutions, official roles, groups or individuals. A further question arises, which seemed unanswerable: what about those articles that did not make the final warrant? Incomplete records as to article sponsorship and non-warrant articles allowed too great a chance for error in the findings by this route of inquiry.
3 This assertion is based on examination of nomination papers for candidature for local office, newspaper accounts during the period, and the overwhelming dominance of the Republican Party in the Town. To demonstrate the last point two sorts of evidence are at hand. First, between 1947 and 1963 there were 100 candidates for local public office, defining candidature arbitrarily as anyone who received more than five votes for a particular office. The party registration of these candidates was: Republican 72; Democrat 7; Independent 18; Unknown 3. Of the seven Democrats who ran, only two were successful. Second, the town's per cent Republican vote for the gubernatorial candidate during the period was: 1946— 92.7%; 1948—84.8%; 1950—83.7%; 1952—89.7%; 1954—88.9%; 1956—85.2%; 1958—83.1%; 1960—81.7%; 1962—78.5%. While there is a decline in the Republican percentage over the period, it remains high enough to classify the town as one-party, Republican.
4 Boskoff, Alvin and Zeigler, Harmon, Voting Patterns in a Local Election (Philadelphia and New York, J. B. Lippincott, 1964), p. 18Google Scholar and Lane, Robert E., Political Life (Glencoe, 1959), pp. 308–11Google Scholar.
5 Lazarsfeld, Paul F., “Interpretation of Statistical Relation as a Research Operation,” Lazarsfeld, and Rosenberg, Morris, eds., The Language of Social Research (Glencoe, 1955), pp. 115–125Google Scholar.
6 Simon, Herbert A., “Spurious Correlation: A Causal Interpretation,” Journal of the American Statistical Association, Vol. 49 (09, 1954 pp. 467–479Google Scholar.
7 Blalock, Hubert M. Jr., Social Statistics (New York, 1960)Google Scholar and Causal Inferences in Nonexperimental Research (Chapel Hill, The University of North Carolina Press, 1964)Google Scholar.
8 For the purposes of simplicity and clarity, the following symbols are used in this portion of the presentation: Board of Selectmen contests (t); voter turnout (y); Moderator contests (x1); School Committee contests (x2); Planning Board contests (x3); and, “Other office” contests (x4).
9 If it becomes apparent that a specific “outside” or “extraneous” variable is not random in its effects, this variable should be brought into the system of variables to aid in explaining the relationships under investigation. Simon, op. cit., pp. 471–473 and Blalock, , Social Statistics, pp. 339–340Google Scholar.
Comments
No Comments have been published for this article.