Methodology

The German Internet Panel is based on a random probability sample of the general population in Germany aged 16 to 75. The study started in 2012, and was supplemented with additional participants in 2014 and 2018. The panel participants were recruited offline using strict statistical procedures.

Every other month, panel participants are invited to take part in a voluntary online survey. The GIP questionnaires cover a variety of topics covering current events (for more details, please click here).

The online panel methodology in combination with the random probability sample of the general population enables the GIP to conduct research on both cross-sectional and longitudinal issues, to run randomized experiments, and to react quickly to current events in politics, the economy, and society.

Sampling and recruitment
The sampling and recruitment of participants in 2012 and 2014 differs from the sampling and recruitment of new participants in 2018.

In 2012 and 2014, the GIP sampled with a so-called ADM design. In a first stage, 250 and 299 regions were randomly selected from a database of 52,947 regions, with approximately the same number of households. The regions were stratified in advance according to federal state and urbanity to ensure a proportionate distribution across the whole of Germany. A random starting address was then drawn in each region. Following a strict random route procedure, the first 200 households (identified via names on mailboxes or doorbells) along the pre-defined random route were listed in a database. At the head office, this household list was cleaned for non-existent households. Subsequently, we randomly drew households with a fixed spatial distance of exactly 5 households. Finally, selected households received an advance letter announcing the visit of an interviewer. The final gross sample consisted of 4,878 households in 2012 and 9,316 in 2014.

Interviewers conducted 15-minute face-to-face interviews in the households to establish contact, to introduce the study, to determine which household members were part of the eligible sample (target population: persons aged 16-75), and to identify households in need for computer and internet equipment to enable their participation in the study.

Eligible household members who agreed to be contacted were subsequently invited by postal mail to participate in the online study. Households that did not have the necessary computer and internet equipment were provided with user-friendly devices, internet and the appropriate technical support.

In 2012, the recruitment process yielded 1,578 registered online panelists with a cumulative AAPOR response rate of 18.5%; in 2014, this were 3,386 registered online panelists with a cumulative response rate of 21.0%.

In 2018, the GIP recruited the so-called experimental panel. The aim of this extremely successful experiment was finding ways to significantly reduce costs, while maintaining the same high data quality of previous GIP recruitments. For this purpose, at the first stage, a random probability sample of 180 municipalities from all municipalities in Germany. The list of municipalities was stratified in advance according to federal state and population density. At the second stage, all the municipalities were contacted and asked to draw a random probability sample of persons aged 16-75 from their local population registers. The number of persons drawn per municipality was proportionate to the total number of persons registered in the municipality as their primary home. A total of 50,214 persons were drawn, from which the head office randomly sub-selected 13,050 persons for the GIP gross sample.

All persons randomly selected in this way were invited to the online panel by postal mail. Six different procedures were tested experimentally. These procedures are called “paper-first”, “concurrent”, “push-to-web”, “online-only”, “early-bird-incentive 20” and “early-bird-incentive 50”.

In 2018, the recruitment process yielded approx. 3,000 registered online panelists with a cumulative AAPOR response rate of 24.1%.

Regular surveys
All samples are regularly surveyed to collect individual data on socio-economic characteristics, behaviors, and attitudes with bi-monthly 20-25 minutes questionnaires. For each survey, an email-invitation is sent out to all panelists on the 1st day of the odd months. The survey remains open for one month. At the beginning of each questionnaire, the identity of the respondents is verified through self-assessment. The participants receive a conditional incentive of 4 Euros and plus a yearly bonus: 10 Euros for all who took part in all surveys of year, or 5 Euros for all who took part in all but one surveys of year. Study participants may choose whether the incentive is transferred into their bank account, paid out as an Amazon voucher, or donated to charity.

Open-Science
The GIP survey data are made available to researchers as Scientific Use Files in the GIP data archive at the GESIS-Leibniz Institute for the Social Sciences six months after the completion of fieldwork and after data preparation and pseudonymization. As part of the Open Science movement, the GIP makes all survey data available to the scientific community. Researchers worldwide can apply for data access free of charge (except for a processing fee).

The GIP collects a wealth of detailed information as well as paradata that cannot be provided in Scientific Use Files for reasons of data protection. In order to enable scientists to work on these data, the GIP offers On-Site Data Access (ODA) in our Data Center in Mannheim. On request, we can provide sensitive data such as small-area information, the year of birth, verbatim data, ECSP collected paradata, as well as detailed response categories that were aggregated in the Scientific Use Files. This sensitive information may be used in the controlled environment of the ODA.

Further information
The GIP is a methodological research project in the field of data science and aims to develop new methods and improve existing methods. To this end, we regularly conduct experiments in the GIP operations and utilize existing data to investigate methodological effects

The results of our methodological research and detailed descriptions of the GIP methodology have been published in various scientific articles. For a complete list of publications please consult the section publications on our website.

Below you will find a selection of articles that specifically address the GIP methodology:

  • Blom, A. G., Gathmann, C., & Krieger, U. (2015). Setting Up an Online Panel Representative of the General Population: The German Internet Panel. Field Methods, 27(4), 391–408. https://doi.org/10.1177/1525822X15574494
  • Blom, A. G., Bosnjak, M., Cornilleau, A., Cousteaux, A. S., Das, M., Douhou, S. & Krieger, U. (2016). A Comparison of Four Probability-Based Online and Mixed-Mode Panels in Europe. Social Science Computer Review, 34(1), 8-25. doi: 10.1177/0894439315574825
  • Blom, A. G., Herzing, J. M. E., Cornesse, C., Sakshaug, J. W., Krieger, U., & Bossert, D. (2017). Does the Recruitment of Offline Households Increase the Sample Representativeness of Probability-Based Online Panels? Evidence From the German Internet Panel. Social Science Computer Review, 35(4), 498–520. https://doi.org/10.1177/0894439316651584
  • Herzing, J. M. E., & Blom, A. G. (2019). The Influence of a Person’s Digital Affinity on Unit Nonresponse and Attrition in an Online Panel. Social Science Computer Review, 37(3), 404–424. doi.org/10.1177/0894439318774758
  • Felderer, B., & Blom, A. G. (2019). Acceptance of the automated online collection of geographical information. Sociological Methods & Research, 1-21. https://doi.org/10.1177/0049124119882480
  • Cornesse, C., Blom, A., Dutwin, D., Krosnick, J., De Leeuw, E., & Legleye, S. et al. (2020). A Review of Conceptual Approaches and Empirical Evidence on Probability and Nonprobability Sample Survey Research. Journal Of Survey Statistics And Methodology, 8(1), 4-36. doi.org/10.1093/jssam/smz041