Innovations in data collection methods in a household panel surveyUnderstanding Society Project Linked Studentship
- Location: the accredited SeNSS DTP
- Duration: three years, beginning in October 2019 and completing in 2022
- Supervisor: Professor Annette Jäckle
Recent technological changes and increases in the costs of survey data collection are leading to shifts in the way that National Statistical Institutes and survey organisations collect data: questionnaire based survey data are increasingly being linked with process-generated (‘big’) data and with new forms of data collected with new technologies. The aims of combining data generated in different ways are to improve cost efficiency and increase the research value of data by providing new, more detailed, or more accurate measures than can be collected with survey questions alone. Combining data generated in different ways however has implications for total survey error, in particular for selectiveness in who participates and measurement quality.
Understanding Society has an ongoing programme of research investigating innovations in data collection methods. Our research makes extensive use of opportunities for experimentation offered by the Understanding Society Innovation Panel.
This studentship will provide the opportunity for a talented student to take advantage of our programme of work. Research questions will for example include the following:
- How best to implement data collection using mobile devices, to maximise participation, minimise selection biases, and maximise accuracy of measurement? This will include experimentation with mobile methods to collect nutrition data.
- How best to implement periodic mini surveys to identify life events that happen in the interval between panel interviews, and could be used to trigger follow-up surveys. This will include experimentation with different modes (text messaging, apps, browser-based).
- Studying the extent and nature of missingness in multiple linked data sources, to examine how best to adjust for missingness and experimentally test ways of increasing consent to linkage. This will also include secondary analyses of existing data.