Empirical research fundamentally depends on the theoretical concepts which it is based on. Hence, to foster empirical discoveries and to better understand and explain the world we live in, these concepts need to be reassessed regularly. Yet often well-established concepts continue to be applied based on common practice rather than fundamental justification.
Party identification is one of these often used and heavily relied-on concepts in the social sciences. In current research, party identification is mostly used as an independent variable to explain political behavior. In my paper I will focus on a new specification of the concept of party identification and show that party identification can be understood in a number of ways. My work will evolve around the hypothesis that the concept of party identification cannot be used indiscriminately and without further differentiation.
The decline in party identifiers points to the question whether people are just not interested in political parties and politics at all, or whether the understanding of party identification has changed over time. Current research suggests that party identification can on the one hand be understood as a deep psychological tie and hence as group belonging (Green et al. 2002, and Greene 1999, 2004). On the other hand, Fiorina (1981) has a much more limited concept about the evaluation of parties which will be updated as citizens get new political information. As researchers have these different understandings, it can be expected that citizens will have them as well.
I will argue that there is not 'the one' party identifier as the vast majority of literature assumes. Rather, people mean different things when they are asked about their party identification and hence should not be lumped together in empirical analyses. I will show that a new specification of the concept of party identification is required and new terms need to be introduced for different kinds of party identifiers, which in some cases cannot even be called ‘identifiers’ anymore.
Further, the affective and cognitive characteristics of party identification (Lodge and Taber 2013) need to be taken into account. I will consider the balanced identity theory by Greenwald et al. (2002) in this context.
Most of the discussion and evidence around party identification has so far focused on the United States. In my paper I will analyze the possible differentiation of party identifiers in multi-party settings. In addition, my paper aims firstly at developing a new specification of the concept of party identification and secondly at getting a better understanding of its implications. This will put the conventional research approach of considering party identification as one amorphous independent variable in the analysis of e.g. voting behavior into perspective.
