¹ Numerical equality is, of course, merely one ideal, and it is certainly an ideal which is neither very confining nor demanding. It is in no way necessarily related to other characteristics which are normally considered to be desirable in a system of representation. Gerrymandering, for example, does not necessarily require deviations from numerical equality, and, given certain characteristics of the geographical distribution of the electorate, gerrymandering itself may be a force tending to produce numerical equality.

² We realize that competition between major political parties is not the only form of political competition which is important in the United States. Clearly in many states competition within one or the other of the major political parties is much more important than that between parties. A study with more ambitious objectives than ours would have to explore the influence of this and other types of political competition. Our investigation should be viewed as an attempt merely to explore the influence of strength of competition between the two major parties upon state and local expenditure patterns.

³ For an excellent discussion of implicit theorizing in polıtical science see Lindblom's, Charles E. “In Praise of Political Science” World Politics, 24 (Jan., 1958)Google Scholar.

⁴ Trans-action, 4 (03, 1967), 41–46Google Scholar.

⁵ Ibid., p. 46.

⁶ “Malapportionment and Public Policy in the States,” Journal of Politics, 27 (08, 1965), 599Google Scholar.

⁷ “The Relation Between Public Policy and Some Structural and Environmental Variables in the American States,” this Review, 60 (03, 1966), 82Google Scholar.

⁸ We do, of course, have a priori reasons for believing apportionment patterns and party competitiveness might influence the pattern of expenditures or we would not have undertaken this study. Two examples of the type of a priori argument which led us to believe that this is so are Downs', AnthonyAn Economic Theory of Democracy (New York: Harper and Row, 1957), especially pp. 21–74 and 114–141Google Scholar, and Riker's, William, The Theory of Political Coalitions (New Haven: Yale University Press, 1962)Google Scholar.

⁹ U. S. Bureau of the Census, Census of Governments: 1962, Vol. IV, No. 4, Compendium of Government Finances (Washington, D. C.: U. S. Government Printing Office, 1964), pp. 57–58Google Scholar.

¹⁰ Fabricant, Solomon, The Trend of Government Activity in the United States Since 1900, (New York: National Bureau of Economic Research, Inc., 1952), Chapter 6Google Scholar.

¹¹ See, Fisher, Glenn W., “Interstate Variation in State and Local Government,” National Tax Journal, 17 (03, 1964), 55–74Google Scholar; Sacks, Seymour and Harris, Robert, “The Determinants of State and Local Government Expenditure and Inter-Governmental Flow of Funds,” National Tax Journal, 17 (03, 1964), 75–85Google Scholar; Bahl, Ray W. Jr., and Saunders, Robert J., “Determinants of Change in State and Local Government Expenditures,” National Tax Journal, 18 (03, 1965), 50–57Google Scholar; and Morse, Elliot R., “Some Thoughts on the Determinants of State and Local Expenditures,” National Tax Journal, 19 (03, 1966), 95–103Google Scholar. Morse's article includes an excellent discussion of the problems involved in the selection of independent variables to “explain” interstate variations in expenditure levels.

¹² The social variables, as well as the per capita personal income figure, are taken from U. S. Bureau of the Census, Statistical Abstract of the United States, 1964 (85th edition; Washington, D. C.: U. S. Government Printing Office, 1964), p. 11 and p. 329Google Scholar. Alaska and Hawaii are omitted from the analysis because data upon which the political variables are based are not available.

¹³ We do not intend to imply that our analysis is completely free from multicollinearity difficulties. As can be seen below we still have a few independent variables which are correlated with one another in the same regression equation. These simple correlations are

As we have previously indicated, the simple correlation which gives us the longest pause for thought is the .75 correlation between personal income per capita and our index of party competition. We do not feel that this level causes serious problems of interpretation although undoubtedly it does dim the clarity of our results. We do not feel that any of the other correlations are of sufficient magnitude to introduce serious multicollinearity problems in our analysis.

¹⁴ The concepts upon which this score is based are presented in Schubert, Glendon and Press', Charles article, “Measuring Malapportionment,” this Review, 57 (June, 1964), 302–327Google Scholar, but the actual score used in this article is taken from a corrected version of their numerical results in “Malapportionment Remeasured,” this Review, 57 (December, 1964), 578–580Google Scholar.

¹⁵ The measure only relates to the “ideal” of numerical equality, and as we have indicated in note 1, this “ideal” has limited theoretical appeal. We have explored the possibility of constructing a measure of “representativeness” which incorporated measurements of gerrymandering as well as deviations from numerical equality. We have, so far, been thwarted in our attempt to construct such a measure, not by difficulties in empirical measurement, but by inadequate specification by political theorists of what gerrymandering “really is.” The statistical problems in measuring gerrymandering do not seem to be too serious, but it is impossible to “measure” something which has not been adequately described.

¹⁶ Hofferbert, Richard I., “Classification of American State Party Systems,” Journal of Politics, 26 (08, 1964), 562–563CrossRefGoogle Scholar.

¹⁷ Certainly party competition in Presidential and Senatorial campaigns reflects competitiveness, but, we felt that perhaps these elections are more subject to the influence of issues unrelated to state and local spending decisions than was the gubernatorial election. Ideally we would like to use a measure which is influenced by a much wider range of elections at the state level (particularly state legislator elections) and also competition at the local level. The construction of such a measure, however, would be a major task in itself, and one much too heroic to be attempted in a paper with objectives as modest as the present ones.

¹⁸ The best description and criticısm of this work is Pfeiffer's, David G. article “The Measurement of Inter-Party Competition and Systematic Stability,” this Review, 61 (June, 1967), 457–467Google Scholar. Our simple index was fabricated and the calculations for the paper performed, unfortunately, prior to the appearance of Pfeiffer's article. Our measure does meet some of his objections to previous measures; but, more importantly—Pfeiffer's excellent analysis included—we think our assertion still stands. A theory of political or even party competition is yet to appear, and, until it does, most attempts at measurement are doomed to produce the sort of “ad hoc” results which existing measures do.