Substantial constituency influence over the lower house of Congress is commonly thought to be both a normative principle and a factual truth of American government. From their draft constitution we may assume the Founding Fathers expected it, and many political scientists feel, regretfully, that the Framers' wish has come all too true. Nevertheless, much of the evidence of constituency control rests on inference. The fact that our House of Representatives, especially by comparison with the House of Commons, has irregular party voting does not of itself indicate that Congressmen deviate from party in response to local pressure. And even more, the fact that many Congressmen feel pressure from home does not of itself establish that the local constituency is performing any of the acts that a reasonable definition of control would imply.Control by the local constituency is at one pole of both the great normative controversies about representation that have arisen in modern times. It is generally recognized that constituency control is opposite to the conception of representation associated with Edmund Burke. Burke wanted the representative to serve the constituency's interest but not its will, and the extent to which the representative should be compelled by electoral sanctions to follow the “mandate” of his constituents has been at the heart of the ensuing controversy as it has continued for a century and a half.

The argument is presented for defining party identification by the root question, “Generally speaking, do you usually think of yourself as a Republican, a Democrat, an independent, or what?” With this definitional base, the partisan balance between Democrats and Republicans between 1952 and 1980 shows no evidence of realignment outside the South, belying the implications of the Markus-Converse and Fiorina analyses that suggest volatility in response to short-term influences. It also appears that the correlation between party identification and voter choices for president are very constant over time in the South as well as outside the South. Party line voting by party identifiers varies by region and party but did not decrease between 1952 and 1988.

NY mid-term congressional election raises pointed questions I \ about party government in America. With the personality of y m \ the President removed from the ballot by at least a coat-JL A_tail, the public is free to pass judgment on the legislative record of the parties. So the civics texts would have us believe. In fact, however, an off-year election can be regarded as an assessment of the parties' record in Congress only if the electorate possesses certain minimal information about what that record is. The fact of possession needs to be demonstrated, not assumed, and the low visibility of congressional affairs to many citizens suggests that the electorate's actual information should be examined with care.How much the people know is an important, if somewhat hidden, problem of the normative theory of representation. Implicitly at least, the information the public is thought to have is one of the points on which various classical conceptions of representation divide. Edmund Burke and the liberal philosophers, for example-to say nothing of Hamilton and Jefferson-had very different views about the information the public could get or use in assessing its government. And the periods of flood tide in American democracy, especially the Jacksonian and Progressive eras, have been marked by the most optimistic assumptions as to what the people could or did know about their government. To put the matter another way: any set of representative institutions will work very differently according to the amount and quality of information the electorate has.

The extraordinary discrepancy in the popular vote for President Eisenhower and the vote for Republican Congressmen in the 1956 election dramatized a privilege which the American electorate exercises almost uniquely in the democratic world, the right of voters to split their ballots between the candidates of opposing political parties.The fact of ballot splitting in American elections is of course a commonplace but it has not been widely studied and it is not well understood. The aggregative statistics from the 1956 election make it apparent that millions of voters must have chosen President Eisenhower and a Democratic congressman but they do not tell us how many voters split their ballots in the opposite direction or how many voted for president but not for Congressman, and they give us only the vaguest indications of what was in the voters' minds when they crossed party lines in marking their ballots.

This article presents a new way to define and measure the ideological sentiments of the mass electorate. Citizens are classified in terms of their evaluations and perceptions of liberals and conservatives. The measure is then used to assess the impact of ideology on the 1972 and 1976 presidential elections, to explore citizens' applications of ideological labels to parties, issues, and presidential candidates, and to describe the relationship between ideology and the potential for party realignment as well as meanings of issue voting.Although many Americans use ideological labels in ways that suggest only a partial understanding of the terms and their implications, those labels have political significance for their political attitudes and election-day decisions.