New Theory and New Data for the Analysis of Partisan Identification: comparing Britain and Germany with evidence from the BHPS and the GSOEP

Publication type

Conference Paper


BHPS Research Conference


Publication date

July 6, 2001


For decades, political scientists have used the concept party identification to structure analyses of political behavior. Classic work by the Michigan School places party attachments at the center of voting decisions in the United States, setting the stage for a debate that has extended over nearly four decades and that has crossed the Atlantic and the Pacific as scholars have explored the concept's utility in the established democracies of Europe and Japan. Three conceptualizations have competed: partisanship as a psychological attachment; partisanship as a summary of the balance between personal policy preferences and the record of the party power, and partisanship as a habit, a standing decision. The issue extends beyond the analysis of individual citizens. In 1969, Converse extended the concept's power, by relating the partisan attachments of citizens to the persistence of democratic regimes. Maintaining that exposure to elections strengths partisan attachments, which in turn reinforce the stability of partisan views and electoral choice, he theorized that strong partisanship is associated with successful democracies.

Survey responses measure partisan identification. Specified to the particular society, questions contained in nationally representative sample surveys provide all the data. Until recently, almost all the surveys applied to a single year (none cover more than five years). As a result, the analysis of change, the development of partisan identification and the strength of this attachment relied on data not well designed to the task. As important, the surveys provided little information on the social context of citizens' lives, treating respondents as
isolated atoms.

The BHPS and the GSOEP (German Socioeconomic Panel Survey) rectify these two serious problems. More, they offer the best avaiable data for the study of partisanship, in particular, and the dynamics of political preferences in general. GSOEP covers fifteen waves (1985-1998) and BHPS covers nine waves (1991-1999); both interview all adults in the household fifteen years of age and older. They are gold mines for the study of partisanship.

Our initial research speaks directly to the core theoretical issues in the analysis of partisanship.

(a) Converse's arguments to the contrary notwithstanding, partisanship does not strengthen with exposure to elections. In neither Germany nor Britain do the data show secular growth in the strength of party attachment. As persons age and experience elections, their partisanship does not grow stronger.

(b) The data do not lend support for the view that partisan preference changes with policy assessments.

(c) Partisan identification is relatively stable: 70 percent of the respondents report the same identication 70 percent of the time.

(d) Change almost always moves people to 'no party', not another party.

Our paper will display these results and offer a theory that makes sense of these findings. We expect the 'stickiness' of partisanship to derive from a set of psychological characteristics and the nature of the recurrent political choice. We expect as well the preferences of persons who are in a person's immediate social circle will influence the stabililty of partisanship. More generally, we expect political preferences to reflect social context as much as individual level characteristics.



Latest findings, new research

Publications search

Search all research by subject and author


Researchers discuss their findings and what they mean for society


Background and context, methods and data, aims and outputs


Conferences, seminars and workshops

Survey methodology

Specialist research, practice and study

Taking the long view

ISER's annual report


Key research themes and areas of interest