Among the most important elements in assessing the cross-national utility of the concept of party identification is its stability over time, in particular its stability relative to that of voting choice. This article utilizes three-wave panel studies in Great Britain, Canada, the Netherlands and the United States to measure the nature and extent of changes in individual partisanship over a period of three to six years. It is argued that there is substantial instability in party identification when all types of change are taken into account, and that the United States appears unique among the cases examined in its ability to combine stable partisanship with instability of voting behaviour for reasons relating to the particular nature of the American electoral system. In Britain and Canada, party identification exhibits a greater tendency to travel with vote, while in the Netherlands it is less stable than the vote itself.
|Journal||European journal of political research|
|Publication status||Published - 1981|