Referendums are rare events in most parliamentary democracies, and when they do occur they present an analytical puzzle. Are they such unusual events that they fall outside of the theoretical frameworks familiar to students of elections? Or, even though they enter political life infrequently, can they be understood as something not entirely foreign to our thinking about electoral politics? Here, we argue that voting in referendums such as the constitutional referendum of October 26, 1992 is driven by many of the same factors that are present in elections—parties, leaders, issues, a campaign timetable, the interplay between long- and short-term forces and the dynamic of the campaign itself. In spite of their unique features, referendums can be understood in terms of models of voting behaviour familiar to students of elections in Canada and elsewhere. But, devoid of some of the long-term partisan and social anchors which play a role in elections, their outcome is even more dependent on the short-term elements of the campaign. As such, referendums are subject to greater volatility and uncertainty than that typically found in ordinary parliamentary elections.

This article compares various indicators in order to establish the weight of linguistic considerations in the constitutional choices of francophone Quebeckers. This comparison is made within a model which also takes into account the respondents' feelings of attachment to Quebec and Canada and their expectations of sovereignty's economic consequences. The results show that an adequate measure of the expected linguistic gains associated with sovereignty must take into account both the comparative and prospective dimensions of these expectations. The use of one such measure confirms the often-stated but previously unproven thesis that linguistic considerations play a crucial role in Quebeckers' constitutional choices.

Multiple regression analysis is used to test the existence of a link between the programmes of the Progressive Conservative, Liberal and New Democratic parties and subsequent government spending priorities. The analysis shows that the programme of the governing party is a poor predictor of government policies. Instead, public expenditures have been sensitive to changes in the programmes of opposition parties. The analysis also shows that government spending priorities in some important policy areas have been more sensitive to opposition party programmes when the popularity of these parties was rising. While public support has favoured the Liberals for most of the period of analysis, the fragile nature of this support has left the governing Liberals uncertain about their prospects of subsequent victory at the polls. This uncertainty has led the governing Liberals to compromise with the Progressive Conservatives on some issues and to mirror the proposals of the New Democratic party on other issues.

Theories of hegemonic leadership often begin with a constraining assumption: that international structure can explain both a state's capability to provide leadership and its interest in doing so. By conflating these explanations, traditional theories, even those from quite different approaches, share common problems. These problems are illustrated by examining three well-known models, and comparing their applications to the eighteenth century. This period provides difficult cases for all three, since countries with power did not provide political or economic leadership, and those which were attempting to provide leadership were not powerful. An alternative theory of leadership is then offered, using structural factors to explain capabilities, but domestic political economic factors to explain interests. The ability of this model to handle the historical cases with greater accuracy suggests that domestic factors could offer fresh insight into theories of leadership.