¹ See Broder, David S., The Party's Over (New York: Harper and Row, 1972)Google Scholar as well as the classic report of the APSA, Toward a More Responsible Two-Party System, Report of the Committee on Political Parties (New York: Rinehart, 1950)Google Scholar.

² See Turner, Julius, Party and Constituency (Baltimore: The Johns Hopkins Press, 1951)Google Scholar; Turner, Julius and Schneier, Edward V. Jr., Party and Constituency, revised edition (Baltimore: The Johns Hopkins Press, 1970)Google Scholar; Truman, David B., The Congressional Party (New York: Wiley, 1959)Google Scholar; Matthews, Donald R., U.S. Senators and their World (New York: Random House, Vintage Books, 1960)Google Scholar; Marwell, Gerald, “Party, Region and Dimensions of Conflict in the House of Representatives, 1949–1954,” American Political Science Review, 61 (June, 1967), 390–399CrossRefGoogle Scholar; Sapira, Michael J., “The House and the Federal Role: A Computer Simulation of Roll Call Voting,” American Political Science Review, 62 (June, 1968), 494–517Google Scholar; and Cherryholmes, Cleo and Shapiro, Michael J., Representatives and Roll Calls (New York: Bobbs Merrill, 1969)Google Scholar. The work by Cherryholmes and Shapiro comes closest to an assessment of party within the full context of legislative behavior. For an extensive review of the literature on party voting, see Shannon, W. Wayne, Constituency and Congressional Voting (Baton Rouge: Louisiana State University Press, 1968), pp. 3–63Google Scholar.

³ In purely descriptive terms, the role of party in roll call voting has gradually declined. In 1921, nearly one of every three roll calls showed a split in which more than 90 per cent of the members of one party opposed more than 90 per cent of the members of the other party. In 1967, barely three of every hundred roll calls met that standard. See Turner, and Schneier, , Party and Constituency, p. 17Google Scholar. Even while this secular decline was under way, it was common for a congressional session up until the mid 1960s to witness a majority of Democrats oppose a majority of Republicans on most roll calls. During every session from 1966 until 1970, however, only a third of the roll-call votes fitted that description. See Sorauf, Frank J., Party Politics in America, 2nd ed. (Boston: Little, Brown and Co., 1972), p. 351Google Scholar. However intriguing these symptoms may be, hinting at both a long-term as well as an accelerated recent decline of the party factor in roll-call voting, they lie outside the scope of this analysis.

⁴ See Truman, , The Congressional Party, pp. 203–226Google Scholar; Matthews, , U.S. Senators, pp. 123–129Google Scholar; and Jackson, John E., “Statistical Models of Senate Roll-Call Voting,” American Political Science Review, 65 (June, 1971), 451–470CrossRefGoogle Scholar. Some of the conditions under which party leadership may affect the voting decisions of representatives are examined by Froman, Lewis A. Jr. and Ripley, Randall B., “Conditions for Party Leadership: The Case of House Democrats,” American Political Science Review, 59 (March, 1965), 52–63CrossRefGoogle Scholar. For a more extended discussion of the activities, organization, and influence of party leadership, see Ripley, Randall B., Party Leaders in the House of Representatives (Washington, D.C.: The Brookings Institution, 1967)Google Scholar. For a look at the Senate, see Huitt, Ralph K., “Democratic Party Leadership in the Senate,” American Political Science Review, 55 (March, 1961), 331–344CrossRefGoogle Scholar. Unlike the U.S. Congress, state legislatures have been found to be more heavily dominated by the party leadership. See, for example, Welch, Susan and Carlson, Eric H., “The Impact of Party on Voting Behavior in a Non-partisan Legislature,” American Political Science Review, 67 (September, 1973), 854–867CrossRefGoogle Scholar. Also see Morehouse, Sarah McCally, “The State Political Party and the Policy-Making Process,” American Political Science Review, 67 (March, 1973), 55–72CrossRefGoogle Scholar. Morehouse argues that the governor, through the control of the party's nominating organization, brings about party cohesion.

⁵ This view is best expressed by Froman, Lewis A. Jr., Congressmen and Their Constituencies (Chicago: Rand McNally 1963), p. 89Google Scholar. Like his study, most of the other studies devoted to this relationship rely on the demography of the congressional district in their attempt to measure the presumed influence of the local constituency. Besides Froman, pp. 92–96, see Turner, and Schneier, , Party and Constituency, pp. 222–225Google Scholar; also Flinn, Thomas A., “Party Responsibility in the States: Some Causal Factors,” American Political Science Review, 58 (March, 1964), 60–72CrossRefGoogle Scholar; MacRae, Duncan Jr., “The Relationship Between Roll Call Votes and the Constituencies in the Massachusetts House of Representatives,” American Political Science Review, 46 (December, 1952), 1046–1055CrossRefGoogle Scholar; and Mayhew, David R., Party Loyalty Among Congressmen (Cambridge: Harvard University Press, 1966)CrossRefGoogle Scholar.

⁶ For the development of complex linkages between the local constituency and legislative decision making, see, above all, the pioneering study by Miller, Warren E. and Stokes, Donald E., “Constituency Influence in Congress,” American Political Science Review, 57 (March, 1963), 45–56CrossRefGoogle Scholar; also Bauer, Raymond A., Pool, Ithiel de Sola, and Dexter, Lewis A., American Business and Public Policy (New York: Atherton Press, 1963)Google Scholar; and Kingdon, John W., Candidates for Office (New York: Random House, 1968)Google Scholar.

⁷ See the lucid theoretical delineation of the two rival models of party government in Warren E. Miller and Donald E. Stokes, Representation in the American Congress, Part I, chap. 2, mimeo.

⁸ See Matthews, , U.S. Senators, p. 133Google Scholar.

⁹ See Epstein, Leon D., “Cohesion of British Parliamentary Parties,” American Political Science Review, 50 (June, 1956), 360–377CrossRefGoogle Scholar; also Kornberg, Allan, “Caucus and Cohesion in Canadian Parliamentary Parties,” American Political Science Review, 60 (March, 1966), 89CrossRefGoogle Scholar. For a recent probe of party differences in policy attitudes in the British House of Commons, see Kornberg, Allan and Frasure, Robert C., “Policy Differences in British Parliamentary Parties,” American Political Science Review, 65 (September, 1971), 694–703CrossRefGoogle Scholar.

¹⁰ See the APSA report, Toward a More Responsible Two-Party System.

¹¹ The studies of McClosky et al. cover a somewhat wider spectrum of party activists. See McClosky, Herbert, Hoffman, Paul and O'Hara, Rosemary, “Issue Conflict and Consensus Among Party Leaders and Followers,” American Political Science Review, 54 (June, 1960), 406–427CrossRefGoogle Scholar.

¹² See Kingdon, , Candidates for Office, pp. 122–123Google Scholar.

¹³ Ibid., p. 144.

¹⁴ Ibid., pp. 55–66; Kingdon strongly argues that these ties do not imply control. The representative is not “owned” by the coalition supporting him nor does he feel he owes it any favors. Instead, he shares from the outset the policy goals of those groups.

¹⁵ The literature on the notion, definition, and impact of attitudes is vast. See, for example, Katz, Daniel and Stotland, Ezra, “A Preliminary Statement to a Theory of Attitude Structure and Change,” in Psychology: A Study of Science, vol. 3, ed. Koch, Sigmund (New York: McGraw-Hill, 1959), pp. 424–465Google Scholar. For a discussion of the role of attitudes in the realm of political behavior, see Greenstein, Fred I., Personality and Politics (Chicago: Markham Publishing Company, 1969)Google Scholar.

¹⁶ See Miller and Stokes, “Constituency Influence;” also Sullivan, John L. and O'Connor, Robert E., “Electoral Choice and Popular Control of Public Policy: The Case of the 1966 House Elections,” American Political Science Review, 66 (December, 1972), 1256–1268CrossRefGoogle Scholar.

¹⁷ That a representative's predisposition cannot have a strong effect on his voting decisions is strongly argued by Miller, Arthur H. in his doctoral dissertation, “The Impact of Committees on the Structure of Issues and Voting Coalitions: The U.S. House of Representatives, 1955–1962” (The University of Michigan, Ann Arbor, 1971), p. 6Google Scholar.

¹⁸ See the vivid testimony cited by Stevens, Arthur G. in his dissertation, “Informal Groups and Decision-making in the U.S. House of Representatives” (The University of Michigan, Ann Arbor, 1970), pp. 54–59Google Scholar; also see Clapp, Charles L., The Congressman: His Work and as He Sees It (Washington, D.C., The Brookings Institution, 1963), pp. 149, 292Google Scholar. A more rigorous assessment of this relationship is presented by Miller, , “Impact of Committees,” pp. 158–208Google Scholar.

¹⁹ See Clapp, , The Congressman, pp. 14–17Google Scholar. Also consult Kingdon, John W., Congressmen's Voting Decisions (New York: Harper and Row, 1973), p. 78Google Scholar.

²⁰ Two files, made available by the Inter-University Consortium for Political Research (I.C.P.R.), contain the data used for this analysis: The Representation Study: The Candidate File, 1970, and Candidates and Constituents, 1971, Ann Arbor, MichiganGoogle Scholar.

²¹ Within the domain of “social welfare,” the issues include public housing, full employment, aid to education, and the federal development of atomic energy. The “civil rights” items refer to Negro voting rights, the establishment of a Fair Employment Practices Commission, and school desegregation. The “foreign policy” issues include foreign economic aid, foreign military aid, aid to neutral countries, and the general theme of isolationism vs. involvement abroad. For a wording of the questions, see I.C.P.R., The Candidate File, pp. 28–43Google Scholar. The items within each of the three policy domains have been combined into Guttman scales, and these scales are used whenever policy attitudes are examined in this analysis. For details of the scale construction, see I.C.P.R., The Candidate File, pp. 219–222Google Scholar.

²² The South is considered to include the eleven states of the former Confederacy plus the four border states of Maryland, Kentucky, Oklahoma, and West Virginia.

²³ For such evidence, see Turner, and Schneier, , Party and Constituency, pp. 171–179Google Scholar; also Truman, , The Congressional Party, p. 149Google Scholar; Price, H. Douglas, “Are Southern Democrats Different? An Application of Scale Analysis to Senate Voting Patterns,” in Politics and Social Life, ed. Polsby, Nelson W., Dentler, R. A., and Smith, P. A. (Boston: Houghton, Miflin Co., 1963), p. 755Google Scholar; and Shannon, , Constituency, pp. 81–94Google Scholar. The early development of the Republican-Dixiecrat coalition is traced by Key, V. O. Jr., Southern Politics (New York: Vintage Books, 1949)Google Scholar.

²⁴ Since the data are weighted, the significance of the differences was tested by relying on the unweighted N for finding the degrees of freedom in the F-table. For information on the assignment of weights, consult I.C.P.R., The Representation Study: The Candidate File. With a random sample of about 120 observations and three groups, any value of eta² exceeding .06 (eta = .25) is significant at the .05 level. For a detailed presentation of eta², see Freeman, Linton C., Elementary Applied Statistics (New York: John Wiley & Sons, Inc., 1965), pp. 120–130, 199–209Google Scholar.

²⁵ For each of the three policy domains, the data set of the Representation Study contains Guttman scales of roll-call votes, one scale for the 85th Congress and another one for the 86th. The social welfare scale for the 85th Congress is made up of items referring to public housing, aid to education, area development, and appropriations for various agencies of the Departments of Labor and Health, Education and Welfare. The items of the foreign policy scale (85th Congress) include a variety of foreign aid votes. And the civil rights scale (85th Congress) consists of items dealing with the Civil Rights Act of 1957. For a more detailed reference to the titles of the bills and a description of the scale construction, see I.C.P.R., The Candidate File, pp. 226–229Google Scholar. For the roll-call scales pertaining to the 86th Congress, see I.C.P.R., Candidates and Constituents, pp. 619–624Google Scholar.

²⁶ Three objects can be ranked six different ways, and each ranking can form a pair with any other one of the six, including itself. This results in 36 pairs of such rankings, six of which are in agreement. Thus the probability of independently ranking three objects the same way twice is 6/36 or .167. The probability of obtaining such a result in three different domains independently, therefore, is (.167)³ or .0046.

²⁷ The importance of a congressman's perception of constituency opinion in the civil rights domain is emphasized by Miller and Stokes, “Constituency Influence.”

²⁸ The zero variance displayed by nonsouthern Democrats in roll-call voting on civil rights bills is rather exceptional and is not duplicated in the subsequent Congress (86th). But the lack of any variance shown by the data for the 85th Congress is not the result of sampling oddities. The sample mean of 6.0 lies well within the neighborhood of the population mean of all nonsouthern Democrats, amounting to 5.9 with a standard deviation of 0.62.

²⁹ The basic ideas for this model have emerged from a great deal of communication betwen the author, Gudmund Iversen, and Lawrence Boyd. See Iversen, Gudmund R., “Recovering Individual Data in the Presence of Group and Individual Effects,” American Journal of Sociology, 79 (September, 1973), 420–434CrossRefGoogle Scholar, as well as Boyd's, Lawrence doctoral dissertation, “Multiple Level Analysis With Complete and Incomplete Data” (The University of Michigan, Ann Arbor, 1971)Google Scholar. Both authors also provide an extensive bibliography of the literature on group, structural, and contextual effects.

³⁰ This finding can also be demonstrated by comparing the correlations between attitudes and roll-call voting within parties. When averaged across policy domains, these correlations stand at a value of .28 for nonsouthern Democrats, at .43 for Republicans, and at .49 for southern Democrats. Herb Weisberg suggested to me that the more “individualistic” behavior of southern Democrats was a function of the broad definition of the South, which included four border states along with the former Confederate states. One could argue that if the Democratic representatives from the border states were counted among the non-southern Democrats, it would be the latter group which would exhibit the more marked covariation between attitude and vote. While this might prove to be the case, it is worth noting, that southern Democrats, even including Border representatives, were found to be more united in their policy attitudes than were nonsouthern Democrats.

³¹ A vast literature has documented and discussed the strength of the party factor in structuring roll-call voting in this domain of policy. Among the more recent publications, see Marwell, “Party, Region and Dimension of Conflict;” Shapiro, “Computer Simulation;” also consult Clausen, Aage R. and Cheney, Richard B., “A Comparative Analysis of Senate-House Voting on Economic and Welfare Policy, 1953–1964,” American Political Science Review, 64 (March, 1970), 138–152CrossRefGoogle Scholar. Clausen and Cheney ((pp. 144–145) distinguish between an “economic policy” dimension and a “social welfare” dimension within the larger domain — not separated in this paper — of government intervention; they found that party was twice as strong a predictor of voting in the economic policy area as in the social welfare area.

³² For a construction of the consultation index and the frequencies, see the codebook for the Candidate File, variable numbers 82, 83, 86, 87, 88.

³³ For a review of these different approaches, see Arthur H. Miller, “The Impact of Committees,” chap. I; also Kingdon, Congressmen's Voting Decisions, chap. 3.

³⁴ For evidence on this point, see Kingdon, , Congressmen's Voting Decisions, pp. 72–74Google Scholar.