Linking survey and experimental data: behavioural experiments in health and wellbeing

The aim of this project is to assess the validity of data collected through experimental methods, by linking them with survey data. While survey data allows large scale collection of a broad range of topics, they might be unable to capture individual preferences, and in particular preferences for risk and time which play a major role in human decisions and behaviour. The different ways in which risk is perceived, and future benefits are evaluated in relation to immediate costs, in fact, can explain a range of behaviours affecting health and wellbeing, from diet, physical activity and smoking, to savings, pensions and long-term care plans. Whilst some surveys ask questions about these preferences, none have used experimental methods ensuring that responses are incentive-compatible, i.e., motivated by real monetary rewards.
At the same time, experiments need to move out of traditionally student-centred labs to involve more general samples. The time is ripe for bringing together experimental methods and general population surveys. Crossing these two different methods of gathering data is crucial for the analysis of human behaviour, as it allows combining experimental evidence on risk and time preferences, with many other data that control for individual histories and backgrounds.
The research will be carried out in the context of health risky behaviour. In particular we will combine data for the same subjects collected by the Understanding Society longitudinal study (USoc), with data collected in experiments where we will assess individuals' attitudes towards risk and time. The research will encompass data collection in several points in time in order to analyse how subjects health risky behaviour changes over time and interact with changes in risk and time preferences.
The project will have a major academic, policy, and social impact. First, it will link, for the first time, a range of experimentally-derived measures with survey data for an active representative panel of the UK population, further enhancing the leadership of Understanding Society as a unique source of high quality and impact data for academic and policy research.
Second, the project will advance inter-disciplinary research and contribute to bridge the gap between the lab and the field, ultimately pushing forward the frontiers of experimental methods in behavioural and social sciences, and exploring the synergies between different research methods and areas, including experimental analysis, health research, econometric analysis of survey and experimental data, social policy, behavioural sciences.
Third, the focus on health and wellbeing will further enhance the policy relevance and impact of the research, allowing to better relate health behaviours and key behavioural attitudes such as risk and time preferences. The project will provide valuable insights to inform the design of policies to bring about changes in behaviour and wellbeing, ultimately enhancing our understanding of some of the foundations of human behaviour.
To ensure that the project will have the highest impact and will benefit the broadest community of users, decision-makers and researchers, a
range of active dissemination strategies will be pursued, including: publication in high impact journals, dissemination seminars for academics and policy makers, presentation of results in international conferences, policy briefings and public events, engagement with the media with a professional newsletter and producing regular media briefings, building a dedicated webpage at the LSE website where podcasts and research results will be regularly published.

The project aims to combine survey data from the Innovation Panel (IP) of Understanding Society (USoc) with new experimental data on risk preferences (the desire for taking a gamble) and time preferences (the degree to which the present is valued more than the future). The project will explore the links between risk and time preferences and behaviour, with a focus on health and wellbeing. A sub-sample of the respondents within the IP will take part into two experimental sessions, taking place at an interval of 12 months, where risk and time preferences will be measured using incentive-compatible tests. Combined survey and experimental data will be used to empirically address three specific research questions related to the "validity" of experimental measures, namely: i) the cross-validation of survey and experimentally-derived measures; ii) the links between risk and time preferences and health-related behaviour; and iii) the stability over time of risk and time preferences and their response to life changes.
The project aims at benefiting a wide community of researchers - from economics to psychology, from social policy to health and behavioural sciences - in three distinct ways.
First, the project will elicit, for the first time, experimental measures for risk and time preferences for a representative sample of the UK population, and it will link them over time with current survey responses, which themselves are being linked with administrative data. USoc will become the first survey to contain a sub-sample of current panel members whose survey responses will be linked to, and broadly cross-validated by, incentive-compatible tests, which will be available to develop further applications and linkages.
This fits the ESRC priority themes in the call, encouraging research "involving secondary analysis of existing datasets and the application of innovative research methods", and will further enhance the leadership of USoc as a unique source of high impact data for academic and policy research. It will also advance the attempts to combine experimental and survey data (e.g. in the Netherlands, using Tilburg's CentERpanel), and to conduct field experiments with samples of the general population (e.g. in Germany and UK, replicating questions in the SOEP and BHPS; in Denmark).
Second, the project will bridge the gap between the lab and the field, ultimately pushing forward the frontiers of experimental methods in behavioural and social sciences, and exploring the synergies and cross-fertilisation between different research methods and inter-disciplinary areas, including experimental analysis, health research, econometric analysis of survey and experimental data, social policy, behavioural sciences.
Third, the focus on health and wellbeing will further enhance the impact of the research, allowing to better relate risk and time preferences and health behaviours, and to provide valuable insights to inform the design of policies to bring about changes in behaviour and wellbeing.
To achieve the highest impact, several dissemination strategies will be pursued, including: i) Involving world-leading experts within the advisory group; ii) Targeting high-impact peer-reviewed journals, e.g. the American Economic Review, Lancet, and Science; iii) Circulating results in working papers and policy briefings making them immediately available before publication; iv) Organising dissemination events: LSE and ISER have excellent research infrastructures for knowledge exchange and strong links with international academic and policy-making institutions; v) Presenting results at international conferences, e.g. the Royal Economic Society, and the American Economic Association; vi) Building a website for posting events, working papers, and uploading slides, videos and podcasts of presentations. vii) Engaging actively with the media, newsletters and websites (e.g. VoxEU) through regular media briefings with the help of knowledge translators.