Measuring party affiliation


FORS Guide Nº 12

How to cite

Lutz, G., & Lauener, L. (2020). Measuring party affiliation. FORS Guide No. 12, Version 1.0. Lausanne: Swiss Centre of Expertise in the Social Sciences (FORS). doi:10.24449/FG-2020-00012


Party affiliation is a key concept used in many social science surveys. To measure such affiliations, one can either ask about previous, current or future voting behavior, about party identification or about party proximity. Moreover, there are questions that ask about an evaluation of different parties simultaneously. This guide provides an overview of the different concepts to measure party affiliation, shows examples of their use in Swiss surveys and outlines the implications for researchers who have to choose among them.


  • The question about actual voting behavior in previous elections is the most obvious question to use in surveys related to elections, because it relates to a real individual behavior in the past. However, question formats need to be adapted to the actual electoral system that is in place in the country of interest. Questions about previous voting behavior become more problematic the further in the past an election took place, because previous behavior may no longer reflect current party preferences.
  • Commercial pre-election polls typically ask about the current state of mind regarding party preferences, as to say, they use the question about party choice if elections were to take place today (or on next Sunday). The answers to this question allow to model where parties actually stand at a given moment in time, but they should not be mistaken as predictions about the outcome of an upcoming election. For academic surveys, current state of mind questions can be used in long-term panel waves to approximate the Election Day that is still far away in the future. In a panel design that covers a shorter time span (typically the election campaign period) and wishes to draw conclusions about the short-term dynamics and changes in public opinion, asking about actual vote intentions might be more purposeful.
  • Party identification questions should be used if researchers want to measure the traditional party identification concept, as to say the sense of belonging to a specific political group over (mostly) a long period of time. When being confronted with suchlike questions, many voters implicitly think of group (or party) membership even if it is not a necessary condition for the theoretical concept. Furthermore, party identification questions often serve as a strong predictor of (past and/or future) voting behavior.
  • If you want to measure party proximity in a more general sense and, especially, if you want to capture the representation aspect of political parties, it may be useful to use a question that asks about what party represents a citizen’s view best. Hence, party proximity questions are more appropriate in this case than party identification questions.
  • On a general note: For all questions that ask about vote intentions, party identification or party proximity, it is worth adding a second (follow-up) question for those who do not reveal a party preference on the first question. Research has shown that a lot of respondents hesitate to give an answer straightaway, either because they consider it a sensitive question, or they are unsure and opt for a “don’t know” answer.
  • If respondents shall evaluate not only one party but several parties simultaneously, voting propensity (PTV) questions do not only relate more strongly to the actual vote choice but also include strategic components of the individual party choice. Party sympathy questions, on the other hand, relate mainly to the affective component of party affiliations.


Copyright: © the authors 2020. This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0)