Despite the voluminous documentation available for each of the American National Election Studies, the recounting of the intellectual history of the studies is not a simple matter. For better or worse, none of the original team associated with the Michigan studies was a diarist. We were, moreover, remarkably devoid of self-conscious attending to, or speculating about what we were doing. In somewhat later years, as we would overhear colleagues speculating about our work, we would occasionally wryly note that the sociology of knowledge might have been better served, where our work was concerned, had there been a professional observer present to chronicle our planning discussions.

One of the most important tests of the success of the survey based study of electoral behaviour is its ability to explain electoral change. Great Britain has provided as severe an examination on that score as any western democracy in the post-war period. In 1950, the country exhibited all of the characteristics of the classic two-party system. The electorate divided its support almost equally between the two main parties, Conservative and Labour, while third parties secured no more than 10% of the vote - and by 1951 as little as 3%.

The Norwegian programme of electoral research was introduced at the Storting Election of 1957 by Stein Rokkan and Henry Valen. Except for the election of 1961, the programme also covers all subsequent Storting elections, which are held regularly every fourth year. The programme has at all times been located at the Institute for Social Research in Oslo. But in 1958, when Rokkan moved to Bergen, the Chr. Michelsen Institute and later the University of Bergen became partners in the programme (until 1966).

Opinion polls came early to Sweden. The Gallup method, as it was called then, was introduced in 1941. A Danish entrepreneur, Haagen Wahl Asmussen had been in contact with George Gallup in 1939, when Gallup tried to start a German Institute in Berlin (it failed because of the invasion of Poland). Asmussen subsequently formed a Danish Gallup Institute and acted as a kind of founding father by disseminating ideas and material from Gallup to Sven 0. Blomquist, the Swedish businessman who created the Swedish Gallup Institute in October of 1941.

Electoral research in the Netherlands is well developed. Sizeable election studies are conducted at every parliamentary election and a considerable number of scholarly publications have appeared in print. The field has always been characterized by diversity in theoretical approaches. This pluralism has however, long been accompanied by collaborative efforts to maintain the common infrastructure of national election studies.

Interest in empirical social research in Germany dates back at least to the turn of the century (Kaase et al. 1983; Scheuch 1990a). However, while in the late 1930s Americans and European emigrants like Paul F. Lazarsfeld were at work developing and applying modern social science research techniques in the United States, the Nazi regime had put a deadly lid on the free flow of ideas and information in Germany.

André Siegfried is well known as the pioneer of electoral studies in France. His Tableau politique de la France de l’Ouest sous la Troisième République (1913) is the beginning of a French ecological school which analyzes the electorate’s behaviour using aggregate data. Maps at very different levels, which were basically the major traditional technique, have been supplemented in recent years by more sophisticated techniques made possible by the coming of computers, techniques which also provide new facilities for comparing electoral and other data.