Caroline Roberts, Oliver Lipps
Scientific surveys of the kind conducted by FORS are designed with the aim of maximising the accuracy of statistical estimates based on the data produced. Accuracy is threatened by multiple sources of error resulting from different survey design features, including for example, the sampling frame and selection method, the design of the questionnaire, the mode of data collection, the wording of invitations, and the incentives offered. Understanding how these design features affect errors is essential for improving data quality and managing survey costs. Our methodological research involves systematically evaluating the quality of FORS data and investigating how different survey design features affect this via impacts on measurement quality and sample representativeness.
Survey research methods are constantly evolving in response to technological developments and societal change. For this reason, much of our methodological research is focused on pre-empting and responding to these advances, to optimize the design of future surveys and mitigate threats to accuracy.
There is a long-standing history of collaboration between members of the SSP faculty specialised in survey methodology and FORS. Our joint projects include:
- Research using FORS data: Our research addresses questions relevant to the field of survey methodology and relating to the following three broad objectives of the programme:
- Evaluation of past and current survey practice at FORS and its effects on data quality;
- Optimization of data collection designs for FORS surveys;
- Preparation for future developments in survey research, including experimenting with new data collection methods and combining survey data with other data sources.
Research findings contribute to the international literature on survey methodology and also feed into developments and improvements in research practice at FORS.
- Methodological Experiments: We design and conduct methodological experiments (randomised controlled trials) alongside surveys conducted by FORS to address research questions relating to these objectives. We use data from these experiments, alongside linked survey and administrative data in our various research projects.
- Joint Seminars: We organise the “SSP-FORS Methods and Research Meetings” – a seminar series showcasing innovative methods in the social sciences and in survey methodology, which can serve as a stimulus for new research projects and collaborations. https://forscenter.ch/events-training/methods-and-research-meetings/
- FORS Guides: We publish summaries of practical knowledge gained from research experience at FORS and SSP-FORS collaborations relating to survey methodology and data management: https://forscenter.ch/publications/fors-guides/.
Ongoing Research Projects
The ongoing and recent research projects listed below are organised around FORS data collection initiatives:
FORS Opt-in Experimental Panel 2022:
- FORS Opt-in Panel Recruitment Experiment (scheduled for February 2022) – comparing the impact of different survey design features on willingness to participate
MOSAiCH/ISSP 2020 and the MOSAiCH COVID-19 Panel:
- Covid-19 impact on survey data collection methods, response rates and nonresponse bias
- Comparing the effectiveness of different incentive types and values
Selects 2019 and Selects-Civique 2019:
- Barriers to participation in app-based surveys – data privacy concerns, task characteristics, usability and onboarding
- Comparing data quality in an app and a browser-based web panel survey
- Exploring the potential of combining survey data with alternative data sources
- Comparing alternative mobile optimisation strategies
- Comparing opt-in vs. opt-out methods of asking consent to participate in a panel study
Swiss Household Panel Mixed Mode Pilot Study 2017-2018
- Mixing modes in longitudinal household panels: Effects on response rates and panel attrition
- Effects of Question Characteristics on Item Non-response in Telephone and Web Survey Modes
- Do we measure more change or more error in a longitudinal web survey compared to telephone
European Values Study 2017
- Can Push-to-Web Surveys Challenge the Position of Face-to-Face Surveys in Terms of Data Quality?
- The implications of modifying length and order of a questionnaire
- The Challenge of Matrix Designs in General Population Surveys: How Much Information Do We Need to Get Reliable Estimates? How can we use multiple imputation to fill in the gaps?
- Comparing data quality in probability-based surveys and opt-in panels
- Sequentially mixing modes in an election survey