Hostname: page-component-8448b6f56d-jr42d Total loading time: 0 Render date: 2024-04-25T00:38:46.293Z Has data issue: false hasContentIssue false
  • English
  • Français

Measurement Identity in the Longitudinal Analysis of Legislative Voting*

Published online by Cambridge University Press:  01 August 2014

Aage R. Clausen
Aage R. Clausen*
Affiliation:
University of Wisconsin
Get access Rights & Permissions [Opens in a new window]

Extract

Recent advances in data processing technology have made it possible for the political scientist to extend the coordinates of his research space across political systems as well as through time. Without the present-day capacity to retrieve and process speedily the large banks of data accumulated through comparative analyses, such studies would be prohibitively time-consuming, and probably not done at all. As it is the technical hurdles are diminishing in importance, only to be replaced in our attention by the methodological barriers to comparative analysis. In this paper the focus is on one of the basic problems of comparative analysis: the achievement and validation of measurement identity.

Measurement identity refers to the content equivalence of two or more measures and is a key consideration in comparative studies whether the comparison is cross-cultural or historical. Unless there are clear indications of the identity of the measures on which the comparisons are based, such comparisons are meaningless. For the political scientist engaged in cross-cultural research, the problem of measurement identity virtually thrusts itself upon him, since he is already sensitive to differences between culturally different political systems. In contrast, the historical researcher who is working within a single cultural context, and is attuned to the continuity of historical themes, may neglect the measurement identity requirements of his research. The purpose of this paper is to give visibility to this measurement issue as it confronts historical research in the field of legislative behavior. My specific referent is the longitudinal study of legislative voting behavior.

Type
Research Article
Information
American Political Science Review , Volume 61 , Issue 4 , December 1967 , pp. 1020 - 1035
Copyright
Copyright © American Political Science Association 1967

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Footnotes

*

I am deeply indebted to Professors Warren E. Miller and Philip E. Converse for their critical evaluations and judicious suggestions. My thanks go to the Survey Research Center, The University of Michigan, for the use of its facilities.

References

1 For a review of the methodology and literature of roll call voting research see Anderson, Leeet al., Legislative Roll-Call Analysis (Evanston: Northwestern University Press, 1966).Google Scholar

2 Lingoes, James C., “Multiple Scalogram Analysis,” Educational and Psychological Measurement, 23 (Autumn 1963), 501–524.CrossRefGoogle Scholar

3 Blalock, Hubert M., Social Statistics (New York: McGraw-Hill Book Company, Inc., 1960), 290–291.Google Scholar

4 The expectation of minor reversals in ordering on two measures of the same dimension is based, in the present case, on the error characteristics of Guttman scales. In a Guttman scale there is a steeply inverse relationship between the frequency of error response patterns and the number of errors in the patterns. Thus if the Guttman scale is accepted as a reliable representation of legislator ordering on a given dimension, when two such scales tapping the same dimension are correlated, the minor reversals in the orderings will be predominant.

5 Suchman, Edward S., “The Intensity Component in Attitude and Opinion Research” in Studies in Social Psychology in World War II, Vol. IV (Measurement and Prediction), Stouffer, S. A. (Ed.), (Princeton, N. J.: Princeton University Press, 1950).Google Scholar

6 Guilford, J. P., Psychometric Methods, 2nd ed. (New York: McGraw-Hill Book Company, 1954), p. 360.Google Scholar

7 Clausen, Aage R., “Policy Dimensions in Congressional Roll Calls: A Longitudinal Analysis,” (Ph.D. dissertation, University of Michigan, 1964), pp. 108–119.Google Scholar

8 I want to emphasize that this procedure for clarifying statistical relationships is not being presented as a measurement innovation of widespread utility. It is useful in the present case; it may be useful in other analyses of roll call voting. Perhaps the strongest feature of this correlation reduction technique is that it does not simply reduce all correlations by eliminating the significant variance on each measure. As we shall presently observe, the differential pattern of correlations involve substantial variations in the magnitude of the correlations.

9 This definition of the analytic subset might have involved one, two, or four Congresses as well as three. In the present case, the selection criterion involved three Congresses because this was the minimum number necessary to provide the required differentiation in the pattern of inter-scale correlations.

10 Miller, Warren E. and Stokes, Donald D., Representation in Congress (forthcoming book to be published by Prentice-Hall).Google Scholar