¹ The first published account of the interrelationships between political party competition, socioeconomic characteristics, and state public policy outputs was Dawson, Richard E. and Robinson, James A., “Inter-Party Competition, Economic Variables, and Welfare Policies in the American States,” Journal of Politics, 25 (05, 1963), 265–289 CrossRefGoogle Scholar. Other studies which emphasize the role of economic variables in determining state spending include Hofferbert, Richard I., “The Relation Between Public Policy and Some Structural and Environmental Variables in the American States,” American Political Science Review, 60 (03, 1966), 73–82 CrossRefGoogle Scholar; and Dye, Thomas R., Politics, Economics, and the Public: Policy Outcomes in the American States (Chicago: Rand McNally, 1966)Google Scholar. Publications placing greater emphasis on the impact of political variables on state spending include Sharkansky, Ira, “Economic Development, Regionalism and State Political Systems,” Midwest Journal of Political Science, 12 (02, 1968), 41–61 CrossRefGoogle Scholar; Sharkansky, Ira, Spending in the American States (Chicago: Rand McNally, 1968)Google Scholar; and Cowart, Andrew T., “Anti-Poverty Expenditures in the American States: A Comparative Analysis,” Midwest Journal of Political Science, 13 (05, 1969), 219–236 CrossRefGoogle Scholar.

For a review of these and many other related works, see Fenton, John H. and Chamberlayne, Donald W., “The Literature Dealing With the Relationships Between Political Processes, Socioeconomic Conditions and Public Policies in the American States: A Bibliographical Essay,” Polity, 1 (Spring, 1969), 388–404 CrossRefGoogle Scholar; and Dyson, James W. and Angelo, Douglas St., “Persistent Methodological and Theoretical Problems in Analysis of Policy Outputs,” paper presented at the Annual Meeting of the Southern Political Science Association, 1969 Google Scholar.

² Jacob, Herbert and Vines, Kenneth N., “Epilogue,” in Politics in the American States, 2nd ed., ed. Jacob, and Vines, (Boston: Little, Brown, 1971), pp. 556–562 Google Scholar. The authors point to inadequate measurement of concepts, inappropriate analysis techniques, and overly restrictive definitions of policy outputs as major sources of difficulty in many of the state expenditure studies. Supportive of this critique is Cnudde, Charles F. and McCrone, Donald J., “Party Competition and Welfare Policies in the American States,” American Political Science Review, 63 (09, 1969), 858–866 CrossRefGoogle Scholar.

More recent attempts to expand the universe of relevant variables include Weber, Ronald E. and Shaffer, William R., “Public Opinion and American State Policy-Making,” Midwest Journal of Political Science, 16 (11, 1972), 683–699 CrossRefGoogle Scholar; and Morehouse, Sarah McCally, “The State Political Party and the Policy-Making Process,” American Political Science Review, 67 (03, 1973), 55–72 CrossRefGoogle Scholar.

³ Sharkansky, Ira and Hofferbert, Richard I., “Dimensions of State Politics, Economics, and Public Policy,” American Political Science Review, 63 (09, 1969), 867–879 CrossRefGoogle Scholar. See also Sharkansky, and Hofferbert, , “Dimensions of State Policy,” in Jacob, and Vines, , pp. 315–353 Google Scholar. Sharkansky and Hofferbert's research is afforded separate notation because their conclusion that different “mixes” of socioeconomic and political factors are influential in state policy making dependent upon the policy area being considered is important to our presentation. In particular, their finding that political factors are more influential with regard to the welfare and education dimension of state policy formation than with the highway and natural resources dimension is one to which we shall return.

⁴ This research is reported in Walker, Jack L., “The Diffusion of Innovations Among the American States,” American Political Science Review, 63 (09, 1969), 880–899 CrossRefGoogle Scholar, and in “Innovation in State Politics,” in Jacob and Vines, pp. 354–387. Subsequent references to Walker's research will refer to this latter work.

⁵ Walker, pp. 364–368.

⁶ Walker, p. 380.

⁷ Walker, pp. 376–377.

⁸ The development of communications networks may lead to increased decision making by emulation. This, in turn, may produce increased demand for additional channels of communication, further stimulating the process.

⁹ We are engaged in a longitudinal study of state executives. The initial research, conducted in 1970, elicited 1,870 returns of a four-page mail questionnaire which is a response rate of 80 per cent. This research is described more fully in Grupp, Fred W. Jr. and Richards, Alan R., “Partisan Political Activity among American State Executives,” paper delivered at the Annual Meeting of the American Political Science Association, 1971 Google Scholar.

The research reported here is based on a second questionnaire mailed to the original 1,870 respondents. This mailing produced 907 usable returns. The reduced response rate is due, in part, to occupational mobility; some of the executives had retired or changed jobs. Perhaps even more relevant was the inclusion of a Thematic Apperception Test in the second questionnaire. The second-wave respondents differ only minutely from the first in objective characteristics—i.e., they are very similar in terms of age, sex, race, education, and type of agency employment.

The state executives are all appointed, politically or under a merit system, and serve in a state's executive branch. Excluded are elected officials; officials of state institutions of higher learning; members of professional and occupational licensing boards; executives in agencies with less than a statewide jurisdiction, such as a port authority; executives in state institutions, such as prisons or hospitals; and uniformed police officials. The goal was to include only upperlevel appointed state executives and to include comparable executives in each of the ten states.

¹⁰ The choice of states was influenced by Luttbeg's, Norman R. study, “Classifying the American States: An Empirical Attempt to Identify Internal Variations,” Midwest Journal of Political Science, 15 (11, 1971), 703–721 CrossRefGoogle Scholar. Four political “regions” emerged from his analysis: Industrial, with 17 states; Southern, with 17 states; Sparsely Populated, with 13 states; and Frontier, with 3 states. Our sample of states includes three from the Industrial Region (California, Wisconsin, and Ohio), three from the Southern Region (Louisiana, New Mexico, and Tennessee), three from the Sparsely Populated Region (Kansas, Montana, and South Dakota), and Nevada from the Frontier Region.

¹¹ Walker at p. 379 describes the respondents to his survey as “top administrators in the departments of education, labor, mental health, welfare, the budget bureau, legislative reference service, and the governor's chief assistant for program development.”

¹² Walker, p. 381.

¹³ Walker (pp. 378–385) deals with the persistence of regionalism in cue-taking among state administrators. Regional leaders, such as Texas and Florida, emerge here as well. Overall, 20 of the 48 contiguous states are included in the two studies. The combined results suggest the existence of national and regional leaders. Because of the nature of the sampling involved, however, only the rough outlines of the influence system are revealed. No definitive mapping of the system is claimed.

¹⁴ Gini ratios are finding increased use in political science. See, among others, Russet, Bruce M., “Inequality and Instability: The Relation of Land Tenure to Politics,” World Politics, 16 (04, 1964), 442–454 CrossRefGoogle Scholar; Alker, Hayward R. Jr., Mathematics and Politics (New York: Macmillan, 1965), chap. 3Google Scholar; and Dye, Thomas R., “Income Inequality and American State Politics,” American Political Science Review, 63 (03, 1969), 157–162 CrossRefGoogle Scholar.

The formula utilized here in calculating Gini coefficients is described in Miller, Herman P., Rich Man, Poor Man (New York: Thomas Y. Crowell, 1971), pp. 274–276 Google Scholar.

¹⁵ While Gini coefficients are frequently directly compared to one another, there are biases which may lead to distortions in the coefficients if the data are aggregated on the basis of unequal numbers of cells or if a large percentage of the cases fall in one cell. See Genson, Richard A., “Gini Ratios: Some Considerations Affecting Their Interpretation,” American Journal of Agricultural Economics (08, 1970), 444–447 Google Scholar.

The Gini coefficients computed here are subject to these sources of distortion. In each instance, the coefficients were calculated on the number of states receiving citations, rather than on a base of 50 states. The rationale for this procedure is that our interest lies in the extent to which esteem is distributed unevenly among those states which are cited. When the ratios are computed on the basis of 50 states, the effect is to increase all of the coefficients to the .8 and .9 range, and the data are not as revealing. What is important is that the rank ordering displayed in Table 2 is not changed by the alternate computation method.

¹⁶ See also Gray, Virginia, “Innovation in the States: A Diffusion Study,” American Political Science Review, 67 (12, 1973), 1174–1185 CrossRefGoogle Scholar. Professor Gray makes a similar point: diffusion of innovation varies by policy area.

¹⁷ Walker, p. 366.

¹⁸ It might be argued that the inclusion of the heterogeneous auxiliary agencies artificially inflates the rho's. The respective rank-order correlations (rho) with auxiliary agencies excluded are: .555; .439; .505; .378; .714; and, .392. The critical value (p. .05) for N of seven (one-tailed test) is .714.

¹⁹ An alternate test of the hypothesis, suggested by one of the anonymous reviewers, is to compare the perceptions of agency influence at each of the three levels of policy making between those executives (type I) who named either of the two most esteemed states as having the more nearly ideal agency and those of the executives (type II) who cited a less esteemed state. This analysis was performed. The type I executives were more likely to perceive their agency as most influential in the policy-making process at each of the three levels of decision making—General policy, 24% to 19%; Specific program, 61 % to 56%; and Individual decision, 84% to 79%. Each of these results is in the direction predicted by the hypothesis, though none of the percentage differences reaches the .05 level of statistical significance.

²⁰ Cnudde and McCrone, p. 865. See also Sharkansky and Hofferbert, note 3 above. But compare Dye, Thomas R., Understanding Public Policy (Englewood Cliffs, N.J.: Prentice-Hall, 1972), p. 261 Google Scholar where he finds environmental factors (income, education, urbanization, and federal aid) are predominant.

²¹ For those interested in the origins of an hypothesis, the actual evolution began with the Sharkansky and Hofferbert findings. How might one account for the waxing and waning of political factors as policy determinants? The serendipitous juxtaposition of Walker's research in the same issue of the American Political Science Review (September, 1969) suggested a possible explanation.