Skip navigation.
Home

Scientific Programme

The scientific focus of the Action is threefold. Due to the variety of national traditions of electoral research in Europe, one focus looks back in time and identifies and compares the major research strategies and findings in the different COST countries. This will help identify common ground that was covered everywhere, but also specificities and research question that have hardly been tackled in some places. This focus is imperative because the scholars of electoral behaviour participating in the Action come from different traditions, publish some or even most of their work in their own languages rather than in English, and therefore know less from one another than necessary.
The second focus is on the comparability of the wealth of data that have been collected by the National Election Studies. It involves the following activities:

Identification of national traditions in empirical electoral research. National communities of electoral researchers, in (search of) responding to different research questions resulting from the specifics of the electoral process in each country, over the years developed partly different instruments and methods to analyze them. The task here is to document these differences because they pose an important challenge for the comparative analysis of the findings of National Election Studies in a cross-national perspective. Closely supervised by the Action, this task shall be carried out by national teams of experienced scholars participating in the Action.
Identification of trend indicators, i.e. indicators that have been asked repeatedly in a plurality of national election surveys. National series of election surveys are not always directed by the same group of scholars, and changes in the composition of research teams often imply changes in the instrumentation of survey questionnaires. The task here is to identify the common ground of survey questions that can be compared within a country over time. This task shall be carried out by national teams of experienced scholars participating in the Action.
Identification of technical details of data collection, among them the sampling frames, the mode of survey administration, completion rates, etc. Comparability of findings depends not only on the questions that have been asked in the different surveys, but also on the way samples have been drawn, on the way survey have been conducted in the field (face to face, telephone, computer assisted, etc.), and not least important on the quality of the survey which is indicated by the response rate. Among the problems that the Action intends to address in this domain are the changes over time from quota samples to multi-stage random samples; from face-to-face interviews to computer assisted telephone interviews; from high to low response rates. In addition, different sorts of biases (that is: the under-representation of particular segments of the universe in the sample,
e.g. low-educated rural socialist voters) need to be identified and corrected for. This task shall be carried out by national teams of experienced scholars participating in the Action.

Definition of the communalities in the instrumentation of national election surveys. Not only do the questionnaires differ within the series of National Elections Studies, they also, and even more so, differ between the study programmes of different countries. The task here is to identify fully comparable indicators that have been asked in a plurality of COST countries and, more demanding, functionally (i.e. conceptually) equivalent indicators that use a somewhat different question wording or response categories and format. Decisions on what is or is not equivalent will be taken on the basis of construct validation techniques (two equivalent indicators should relate to an identical third in the same way). This task shall be carried out centrally by the Working Group focusing on data comparability, with appropriate input from the participating national teams of electoral researchers.
Recoding of comparable indicators into a common format. The age variable is sometimes asked as the actual age of respondents at the time of the survey, and sometimes as the birth year; the left-right self-placement is sometimes asked with the help of a 7-point-scale, sometimes using a 10¬point-sclae, and sometimes an 11-point-scale. These coding differences need to be harmonized in order to arrive at fully comparable instruments. The problem to be solved aggravates if it comes to, e.g., the occupation of respondents as an indicator of their social class because the same occupation might be associated with a different social status in the different COST countries. This task shall be carried out centrally by the Working Group focusing on data comparability, with appropriate input from the participating national teams of electoral researchers.
Integration of comparable variables into a common data-structure. A common data structure means no less than the transformation of 17 national-specific election study trend-files into one pooled data-structure in which nation becomes a variable. This is mandatory in order to be able to analyze the effect of different (country- and time-specific) context characteristics on individual political orientations and behaviors. This in turn is one of the primary research objectives of this Action. This task shall be carried out centrally by the Working Group focusing on data comparability, with appropriate input from the participating national teams of electoral researchers.
Identification, collection and integration of macro variables. Context characteristics are measured as so-called macro-variables. Potentially relevant for the explanation of individual electoral behavior are macro-variables that describe the political regime as such (federalism or not, presidential system or not, government stability, years non-democratic after World War Two, communist past), the electoral system (proportional, plurality or mixed, electoral formula, ballot structure, district size in terms of seats), the party system (fractionalization, polarization, state party funding, remuneration of Members of Parliament, party membership), the economy (GDP growth, size public sector, openness of the economy, taxation, inflation, unemployment, etc.). In countries where information is available about the electoral district in which the interview was conducted, the district-level turnout, election result and the number of seats allocated in this district will be collected. This task shall be carried out centrally by the Working Group focusing on data comparability, with appropriate input from the participating national teams of electoral researchers.
A third focus is on the development of a common theoretical framework and research agenda. It requires frequent and intense scientific exchange, collaborative analyses of the common data base, and the publication and dissemination of findings. This focus will ultimately lead to the elaboration and publication of the second collective book publication ‘The True European Voter’ that will emerge from the Action.

Scientific work plan – methods and means
This Action involves two circles of countries. The wider circle comprises the heritage of the European Voter project, in which the findings of 66 National Election Studies were harmonized. For the six North-Western European countries that participated in the European Voter project, data harmonization tasks are less severe. These countries are Norway, Sweden, Denmark, Britain, the Netherlands and Germany. The Action will update these country files by adding the election studies of the last decade, and add the context information described earlier. This shall be done in co¬operation with the partners of the former European Voter project.
The inner circle of countries includes those where data harmonization has to start more or less from scratch. There are 11 Southern and Eastern European countries with significant traditions of National Election Studies: Italy, Spain, Portugal, and Greece in the South (plus France and Belgium), and Poland, Hungary, the Czech Republic, Croatia and Serbia in the East. Experienced scholars from all of these COST countries - most of them directors of national election studies - have already agreed to participate.
Altogether, some 50 European electoral researchers will be part of the Action in different roles. They will contribute to three Working Groups: a History Group, a Data Group, and a Methods Group. These Working Groups are described in more detail under E.2.
The Action will be open to participants that for one or another reason could not take part in the preparation of it. In addition, specialists from other disciplines might be invited to participate in cases when their particular expertise will be required to achieve the research objectives. An example here is the support of a statistician in the task of harmonising datasets that have been generated by different sampling strategies.