The UK’s longitudinal social research studies have stepped forward during the Covid-19 crisis to offer a means of understanding the outbreak’s impact on work, family life, wellbeing and health-related behaviour. We have seen responsive, fresh data collection from Understanding Society, the national birth cohort studies, the English Longitudinal Study of Ageing and others, which will help guide policy during the pandemic and enable the identification of long-term processes and outcomes as the studies continue over the coming years.
Two features of longitudinal studies make them powerful in this context. Firstly, the rich data collected in the years prior to the outbreak provide an invaluable baseline against which to assess change at the individual level. Secondly, the existing relationships they have established with study members mean we can adapt data collection techniques to suit the situation in which we find ourselves.
The Economic and Social Research Council (ESRC) recently published an array of longitudinal methods reports by study leads and survey organisations that reflect their methodological flexibility and growing confidence implementing new approaches while optimising quality. These reports were commissioned well ahead of the outbreak and coalesced around three areas in particular: the shift to multimode approaches that are online at their core, considerations around harnessing new tech (including wearables that provide bio-measures and smartphones as powerful and flexible research tools), and linking to administrative or other data sets. These features are propelling the longitudinal studies towards more ambitious data collection in terms of the type of data they collect and its timing/frequency.
While new technology and administrative data provide exciting opportunities, in this period of unprecedented crisis the thirst has been for surveys that deliver data tailored to the specifics of new events and do this repeatedly as the situation evolves. Given the pause in face-to-face interviewing since March, this has presented the longitudinal studies with the challenge of collecting comparable measures in different modes and ensuring that the quality of the sample achieved is sufficiently reliable for analysis.
Given its existing multimode infrastructure, Understanding Society’s main fieldwork (an annual interview for sample members) was able to adapt to the new reality relatively seamlessly. Planned face-to-face interviews were switched to a web-first approach, with Kantar and NatCen setting up their experienced face-to-face interviewers to conduct telephone interviews for those sample members not completing online.
However, the real need at this time was for new questions tailored to the outbreak, and here several studies have responded. Understanding Society was able to adapt an approach it was trialling to deliver multiple short online surveys, now also backed by a telephone approach for those who need it. The English Longitudinal Study of Ageing has, for the first time, invited its cohort members aged 50 and over to complete a survey online, with a substantial telephone effort to ensure coverage of the full sample and boost response rates. The result has been very positive, with a response rate of over 70%. The cohort studies run by the Centre for Longitudinal Studies (CLS) have also conducted speedy online waves of Covid-19 data collection – and all intend to conduct further waves to track the impact of the outbreak over time.
ESRC’s suite of reports explore some of the concerns with these approaches, such as potential mode effects in multimode studies (selection effects, and the result of differences in the delivery of questions online versus with an interviewer). They point to an emerging best practice in implementation, for instance designing questionnaires and instruments for mobiles as a first principle and how incentives should be used. There are also encouraging findings from CLS around ‘measurement equivalence’ that build on the literature around mode effects.
Given the ongoing uncertainties around face-to-face fieldwork and the need for more and more frequent data, all of the longitudinal studies are now considering how they might collect data differently in the long term. But the effect of the outbreak has been to hasten thinking that was already happening, and to force them to take the plunge earlier than they might have (with the support of funding bodies like the ESRC). The efficacy of the data will become clearer in the coming months as analysis continues and policy responses are refined, but at a difficult time it’s great that the longitudinal studies have played their part.
Visit our Methodology and Innovation Hub for more information, news and methodology events.