Research Strand 5: Complex measurement in self-completion surveys

Strand Leader: Professor Lisa Calderwood , Centre for Longitudinal Studies, University College London

Capturing complex phenomena is crucial for many social surveys. Existing research has shown that a key challenge for moving to online data collection is how to administer complex measures without detriment to data quality and/or comparability. RS5 will focus on three types of complex measures that are important for many surveys and which present distinct challenges. Standardised measures of industry and occupation rely on probing in interviewer-administered surveys and are consequently prone to mode effects (Peycheva et al, 2021). Consent rates to data linkages, bio-samples and re-contact tend to be substantially lower online (Jäckle et al, 2021; Sakshaug et al; 2017; Thornby et al, 2018). And event history data are more likely to be incomplete when collected online (Emery et al, 2018). Additionally, RS5 will consider a fourth type of measure, the assessment of cognitive function, as this is a good example of a complex measure that is challenging to assess comparably across modes (Ofstedal et al, 2021).

For each of these four types of measures, we will carry out a review of existing evidence on mode and measurement effects and we will investigate how best to collect these kinds of data in online surveys. In the case of event history data, the investigation will focus particularly on how to improve recall of complex histories and how to avoid break-offs. We will conduct new analysis of existing data, designed to address key knowledge gaps identified by the review. For industry and occupation, we will analyse experimental and observational data from the NatCen panel, Generations and Gender Survey (GGS), Next Steps study and the 2021 ONS Census-Link Study. For consent, histories and cognition, we may also conduct new analysis using data from appropriate surveys including GGS and Next Steps.

Outputs for each of the four areas will include practical guidance on how to effectively implement high quality measurement online and we will also produce research papers on high priority topics.