Three models of interest groups, power and political process in America are contrasted: (1) the Truman-Dahl-Lindblom pluralism of the 1960s; (2) the unfinished plural elitism of the 1970s, a theory emphasizing special-interest capture of policy systems whose most influential exponent is Lowi; (3) the ‘triadic’ model of process set forth by Wilson in The Politics of Regulation. The triadic model assumes the normality in policy systems of organized economic producers being challenged by the countervailing power of other organized interests, while state agencies act autonomously. It is argued that the triadic model is the most advanced of the three, although it still needs development. Eighteen illustrative propositions are presented in terms of triadic power. These include relationships among interest groups and state autonomy, ‘high polities’ and routine politics, and types of coalitions in policy systems. Other propositions describe links to possible cycles between triadic power and plural elitism, to corporatist decision-making, and to the ‘resource mobilization’ theory of social movements.

This article extends existing political-economic models to deal more rigorously with politics in countries with trade-dependent economies, and in particular with the policy consequences of oil-exporting in industrial countries. Models drawn from economics and finance show how much of Britain's recent unemployment results from North Sea oil, at first through speculation in sterling in rapidly-growing international currency markets and more recently through the balance of payments. In Norway, by contrast, speculation was deterred by a variety of policies on fixing exchange rates, and the unemployment problem contained by better-planned and executed employment subsidy programmes. These policy variations are explained by differences in available ideas, institutions and, ultimately, structural characteristics.

The authors examine the countervailing effects of two forces on external news coverage of Southern Rhodesia (now Zimbabwe) and South Africa during the 1970s. The first is purposeful government efforts at news management and information control undertaken by each of the two regimes. The second is the civil unrest which was present in the region during that period. They conclude that these effects and the policy consequences that flow from them are functions of the pre-existing image environment of each country in the foreign (US) press and of the character of its domestic unrest.

The surveys for the 1983 general election marked the first twenty years' systematic study of the British electorate. The British Election series – originating with the 1963 pre-election Nuffield survey by Butler and Stokes, followed by the Essex series in the 1970s conducted by Särlvik and Crewe, and latterly in the hands of Heath, Jowell and Curtice – yield exceptionally rich portraits of voters and their responses to politics. The growing crop of secondary analyses, tracing changes in the electorate over those years, reflects the continuities of the series. More recently, a number of studies have challenged the ‘Michigan conventions’ central to the first generation of British electoral studies. It is a good time, then, to take stock of our understanding of voting behaviour in Britain. Let us start with what is generally agreed.

There have been many comparative studies of British and American parties and elections, but little attention has been paid to comparisons of special elections (or by-elections) to fill vacancies in the two lower houses of the central legislatures, the House of Commons and the House of Representatives. Although British by-elections have been closely studied for a long time, even the most basic facts about American special elections were unknown until recently. Despite the lack of direct examination of how by-elections compare in the two countries, American academics have often dismissed American special elections with such passing comments as ‘because of their unsystematic nature and unsynchronized timing, American special elections, unlike British by-elections, have negligible predictive value’.