GLES Design

GLES Design

The aim of the German Longitudinal Election Study (GLES) is to provide high-quality data for the analysis of federal elections in Germany. The GLES surveys enable the investigation of attitudes, opinions, and political behaviour of voters and candidates. In addition, the GLES collected data on media coverage for the federal elections in 2009, 2013, and 2017.

Information on the surveys for the 2021 federal election can be found here.

Linkage of GLES components: A special feature of the GLES studies is that their data can be linked with each other. In addition, all GLES survey studies are based on a common, cross-component question core, which is completed by study-specific questions in order to meet the different analysis objectives of the surveys.

International studies: Some GLES surveys are embedded in international survey programmes. The Post-election Cross-Section includes the German survey of the Comparative Study of Electoral Systems (CSES), and the GLES Candidate Study is included in the Comparative Candidates Survey (CCS).

Overview of the GLES components: In the following, we give a brief overview of the individual GLES surveys from 2018 onwards. We show which analytical purposes they are particularly suitable for and how they are linked to other GLES components. A detailed description of the individual surveys and information on access to data and documentation can be found on the GLES data page.

After the institutionalization of GLES at GESIS in 2018, the Cross-Section, the Rolling Cross-Section, the GLES Panel, the Candidate Study and the Tracking have been continued as the current GLES. Further voter surveys and media content analyses were carried out during the DFG-funding period between 2009 and 2019.

GLES Design 2009-2017
History of the GLES

Voter surveys

The assessment of political attitudes, preferences and behaviour of German voters is one of the main goals of the GLES. For this reason, the GLES regularly conducts different surveys before and after the federal elections. Thus, questions of election research can be examined in a cross-sectional and longitudinal perspective. The GLES data can be used to analyse both short-term dynamics in the course of the federal election campaign and long-term social change processes between individual elections.

What kind of data is collected? Cross-sectional surveys are an essential part of election studies worldwide. For each federal election, the GLES conducts cross-sectional surveys before and after the election. For the 2009, 2013 and 2017 elections, these were face-to-face surveys with about 2,100 interviews each. These surveys are based on a multi-stage, stratified random sample. The interviews lasted about an hour and were conducted as "Computer Assisted Personal Interviews" (CAPI). Due to the Corona pandemic, the cross-section for the 2021 federal election will be conducted as a mixed-mode survey online (CAWI) and written (PAPI).

What can be analysed? The GLES Cross-Section surveys can be used to analyse classical questions of electoral and voter research. These questions can be investigated from a longitudinal as well as an international comparative perspective. The GLES also takes up innovative and up-to-date approaches in electoral and attitudinal research through the inclusion of new questionnaire elements, such as scales on coalition preferences, emotions, and personality traits.

Can the dataset be linked with other (GLES) components? The post-election questionnaire comprises the current question module of the Comparative Study of Electoral Systems (CSES) and thus enables the analysis of internationally comparative questions of electoral research. In addition, the German CSES studies from 1998 (ZA3073), 2002 (ZA4216) and 2005 (ZA4559) can be used to conduct comparative analyses of electoral behaviour over time. Moreover, some questions are asked both in the GLES Cross-Section surveys and in the GLES Candidate Study. This allows for a comparison of the attitudes and behaviour of voters and candidates and for an investigation of the importance of local campaigns for voting behaviour.

For which elections is this study available? GLES Cross-Section surveys are available for the federal elections in 2009, 2013, and 2017. In addition, GESIS has archived at least one election study for every federal election since 1949. The data can be accessed via the GESIS-Search or via the FDZ Wahlen.

Are further surveys planned? For the 2021 federal election, a pre- and post-election cross-section will be conducted as a mixed-mode survey online (CAWI) and written (PAPI). Since 2018, the participants from the 2017 Cross-Section who are willing to be re-contacted have been surveyed within the framework of the GLES Panel.

Data Download
What kind of data is collected? Rolling cross-sectional studies form a standard instrument for recording short-term changes in the political attitudes and behaviour of the population eligible to vote. In the GLES, RCS studies are conducted approximately 60 days before the election with a total sample of approximately 7,000 respondents, of whom about 100-120 are interviewed per day. The surveys are conducted as Computer Assisted Telephone Interviews (CATI) and last about 30 minutes.
The GLES also links the pre-election RCS studies with a post-election panel survey, which takes about 20 minutes and comprises about 4,000 participants of the pre-election RCS study who agree to be interviewed again.

What can be analysed? The Rolling Cross-Section allows examining aggregated changes on a daily basis. Thus, the design enables to capture campaign-induced changes of the public opinion. Voter reactions to events during the campaign can be registered directly, and the decay or stability of these effects can be revealed. Since the RCS studies are combined with a post-election panel survey, it is also possible to investigate intra-individual changes between the two waves.

Can the dataset be linked with other (GLES) components? The GLES RCS studies can be directly linked to the Campaign Media Content Analyses (until 2017) on a daily basis.

For which elections is this study available? GLES RCS studies are available for the federal elections in 2009, 2013, and 2017. The previous study on the 2005 federal election (ZA4991) is also accessible via the GESIS search.

Are further surveys planned? The RCS survey will be conducted again for the 2021 federal election.

Data Download
What kind of data is collected? For the 2009 and 2013 federal elections, a sample of eligible voters was interviewed six times during the election campaign and once after the election. In 2009, 3,771 people were recruited from an online access panel (Respondi) and in 2013, 5,256 people were recruited. The 2017 Campaign Panel started twelve months before the federal election with a considerably larger sample size of around 18,000 persons. The interviews of the Campaign Panels were conducted online (CAWI).

Some respondents of the 2013 and 2017 Campaign Panels already participated in the Campaign Panel of the previous federal election. Since the 2013 federal election, the panellists have been invited to annual interim surveys.

What can be analysed? The Campaign Panel can be used to study intra-individual changes in political attitudes and behaviour during the campaign and between elections.

Can the dataset be linked with other (GLES) components? The Campaign Panel waves can be linked on a daily basis to the Campaign Media Content Analyses (until 2017).

For which elections is this study available? Campaign Panel surveys were conducted for the federal elections in 2009, 2013, and 2017. Since 2014, participants have been invited to annual interim surveys.

Are further surveys planned? In 2018, the Campaign Panel merged into the GLES Panel. Surveys are carried out regularly.

Data Download
What kind of data is collected? All respondents of the GLES Cross-Sections, who agreed to participate in future surveys, form the sample of the Long-term Panels. Each panel comprises a total of three federal elections. From 2009 onwards, annual surveys (by telephone, via mail, and from 2014 onwards, also online) were conducted to maintain the panel.

What can be analysed? Using the Long-term Panel, it is possible to analyse intra-individual changes in attitudes and political behaviour between successive elections. The unique feature of the Long-term Panel is the coverage of two full legislative periods and the continuity of the rolling three-wave design since 1994 (ZA4301).

For which elections is this study available? The Long-term Panel was published for the 2002-2005-2009, 2005-2009-2013, 2009-2013 and 2013-2017 federal elections.

Are further surveys planned? Since 2018, the participants from the 2017 Cross-Section who are willing to be re-contacted have been surveyed within the framework of the GLES Panel. In the GLES Panel, while the sample from the 2017 Campaign Panel is interviewed online (CAWI), the participants of the 2017 Cross-Section samples are interviewed in a mixed-mode design, online (CAWI) and via mail (PAPI).

Data Download
What kind of data is collected? Participants from the 2017 Campaign Panel and the participants of the 2017 Pre- and Post-election Cross-Section surveys who agreed to be questioned again are interviewed up to twice a year between the federal election campaigns. Using a largely identical questionnaire, the sample from the Campaign Panel continues to be interviewed online (CAWI), while the participants of the Pre- and Post-election Cross-Section samples were interviewed online (CAWI) and via mail (PAPI) in a mixed-mode design until 2019. Since spring 2020, the interviews have been conducted exclusively online (CAWI). In the fall of 2020, a further refreshment sample was drawn from the online panels of the respondi AG and GapFish GmbH for the analysis of the election campaign for the 2021 federal election. In the election year 2021, the participants of the GLES panel will be interviewed up to five times before the beginning and during the election campaign and up to two times after the federal election, depending on the sample.

What can be analysed? The GLES Panel can be used to investigate intra-individual changes in political attitudes and behaviour during the election campaign and between elections.

Can the dataset be linked with other (GLES) components? The waves of the GLES Panel that were carried out during the 2017 election campaign can be linked on a daily basis with the Campaign Media Content Analysis.

For which elections is this study available? The GLES Panel has been conducted since 2016. The data from the current survey waves are added to the existing data corpus following a modular approach. The current version of the data set can be accessed via the GESIS search.

Are further surveys planned? Outside federal election campaigns, up to two surveys are conducted annually. Before and during election campaigns, a higher number of surveys with a shorter frequency are planned, at least for a subsample of the GLES Panel that can be reached online.

Download Data
What kind of data is collected? Starting with the 2009 federal election, short cross-sectional online surveys (CAWI) have been conducted up to four times a year. Each cross-section sample consists of approximately 1,000 respondents. In addition, surveys of around 500 respondents were conducted for many state parliamentary elections between 2009 and 2017. They took place at the same time as the regular GLES Tracking, preferably before the respective state elections.

What can be analysed? The GLES Tracking aims to measure short-term aggregate changes between elections by conducting surveys at regular intervals with a constant core survey programme. In addition, the GLES Tracking contained question modules with thematic focuses until 2017. Since 2018, only the core program is queried, supplemented by some current items if necessary.

Can the dataset be linked with other (GLES) components? The GLES Tracking can be linked to the Long-term Media Content Analysis 2009-2017.

For which elections is this study available? Since 2009, three to four surveys per year have been carried out.

Are further surveys planned? Currently, three surveys are conducted per year.

Download Data

Candidate surveys

Within the framework of the GLES, candidates to the German Bundestag are interviewed. In combination with the voter surveys, this allows a comparison of political attitudes and behaviour between the voters and the candidates standing for election as well as the analysis of campaign strategies and their impact on voters.

What kind of data is collected? For each federal election, the approximately 2.500 constituency and party-list candidates of the parties represented in the Bundestag as well as of those parties with a higher probability of being elected to parliament are invited to take part in a survey. The survey is conducted in a mixed-mode design, with interviews conducted online (CAWI) and written (PAPI).

What can be analysed? The core of the study consists of questions about the background of the candidates, their recruitment and selection, their campaigns, political attitudes and their policy positions. Aggregate characteristics of the constituencies from official statistics added to the data also allow the socio-structural context to be included in analyses.

Can the dataset be linked with other (GLES) components? Due to the international cooperation with the (Comparative Candidates Survey (CCS)), the survey instrument of the Candidate Study contains the CCS core questionnaire. Thus, internationally comparative analyses of the attitudes and behaviour of candidates are possible. Moreover, some questions are asked both in the Candidate Study and in the cross-sectional voter survey. This enables a comparison of the attitudes and behaviour of voters and candidates as well as an investigation of the importance of local campaigns for electoral behaviour.

For which elections is this study available? The GLES Candidate Study is available for the federal elections in 2009, 2013, and 2017. The GESIS-Search also provides access to previous studies from 2002 (ZA4225) and 2005 (ZA4923).

Are further surveys planned? The candidate study will be conducted for the 2021 federal election.

GLES candidate brochure 2009 (in German only)
GLES candidate brochure 2013 (in German only)
GLES candidate brochure 2017 (in German only)

Download Data

Media analysis

GLES conducted quantitative media content analyses of the most important German mass media for the federal elections in 2009, 2013, and 2017. Both television coverage and print media were taken into account. As with the voter surveys, the data on the media system is available in a short-term election-campaign-centred and a long-term dimension. In addition, data on the TV debates between the CDU/CSU and SPD candidates for chancellorship and the top candidates of other parties were collected as part of experiments.

What kind of data is collected? The study is a quantitative media content analysis of political reporting in TV news programmes and print media during the election campaign. Five TV news programmes and the coverage of six German daily newspapers were collected. The topic coding was carried out using the GLES coding scheme, which is also used to code the open questions about the most important problems in Germany in the voter surveys.

What can be analysed? The media content analysis allows an analysis of the interplay between media coverage and voter perceptions during election campaigns.

Can the dataset be linked with other (GLES) components? The media content analysis can be linked to the GLES Rolling Cross-Section study and the GLES Campaign Panel on a daily basis.

For which elections is this study available? Campaign Media Content Analyses are available for the federal elections of 2009, 2013, and 2017.

Are further surveys planned? For the time being, no further Campaign Media Content Analyses are planned.

Download Data
What kind of data is collected? The Long-term Media Agenda Analysis is a quantitative media content analysis of the most important weekly political magazines as well as national and regional daily newspapers. The topic coding was carried out using the GLES coding scheme, which is also used to code the open questions on the most important problems in Germany in the voter surveys.

What can be analysed? The long-term recording of media coverage in Germany can be used in combination with the survey data of the Long-term Online Tracking to examine the influence of relevant events on political attitudes and electoral behaviour.

Can the dataset be linked with other (GLES) components? The GLES Long-term Media Agenda Analysis can be linked to the GLES Long-term Online Tracking.

For which elections is this study available? The Long-term Media Agenda Analysis is available for the period from 2009 to 2013.

Are further surveys planned? For the time being, no further Long-term Media Agenda Analyses are planned.

Download Data
What kind of data is collected? TV debates between the CDU/CSU and SPD candidates for chancellorship have established themselves as the main election campaign events in German federal elections. Within the framework of the GLES, experiments were carried out to analyse these debates. Participants were interviewed before and after the TV debate (a few days after the debate and immediately after the federal election). Using a voting machine, the experimental groups were able to express their views directly during the debate. This experimental design was supplemented by a content analysis of the TV debate. In 2009, a content analysis of the media coverage of the debate was also provided.

For the 2017 federal election, an experiment was also carried out on the television debate between the top candidates of the CSU, the FDP, the Greens, the Left and the AfD ("TV Fünfkampf").

What can be analysed? The study aims to investigate the existence and permanence of debate effects on political attitudes and voting behaviour.

Can the dataset be linked with other (GLES) components? The survey, the real-time response data and the media content analyses can be analysed jointly.

For which elections is this study available? TV Debate Analyses are available for the federal elections in 2009, 2013, and 2017.

Are further surveys planned? For the time being, no further TV Debate Analyses are planned.

Download Data