These data were collected as part of a three-year project funded by the Nuffield Foundation. The project aimed to collect experimental evidence on labour supply responses to uncertainty about work availability and associated pay. 301 low-income UK residents took part in several experimental sessions. The experiment was conducted on a tablet/PC via a custom-built web application hosted by a cloud service platform. In the initial phase, data were collected in face-to-face sessions, on wi-fi enabled tablets provided by the research team. After the Covid-19 pandemic struck, data collection was moved online and carried out via Prolific, a recruitment platform.Insecure and unstable forms of employment such as temporary or zero-hours jobs have been increasing since the Great Recession. Yet, the consequences for labour supply are poorly understood. Traditionally, the models used by researchers to examine the impact of work incentives or job search behaviour consider job characteristics other than the wage irrelevant. Nevertheless, aspects such as job security and the predictability of earnings may influence labour supply behaviour. Psychological research shows that the majority of people try to avoid risk and uncertainty, i.e. they are risk-averse. For risk-averse individuals, one side effect of employment and earnings instability may be to reduce the desirability of paid work in comparison to receiving out of work benefits. This project will provide evidence on the impact of earnings instability and uncertainty on labour supply. It will use a combination of quantitative analysis of panel data and laboratory experiments to shed light on whether: i) earnings uncertainty is linked with the probability of leaving unemployment in the general population, and whether this link is stronger for more risk-averse individuals, ii) a causal link between earnings uncertainty and the likelihood of taking up paid work can be established using experimental methods, iii) policy interventions such as benefit sanctions and/or a reduction of benefit application costs can alter the influence of earnings uncertainty on labour supply. Findings will provide valuable insights to policymakers concerned with employment and welfare legislation.

The data has been collected as part of an experiment conducted via a custom-built web application. The data were collected in face-to-face sessions conducted in Colchester (68 participants) and via the online recruitment platform Prolific (233 participants). To be eligible to take part, a participant had to be resident in the UK, aged 18-60, have an annual family income of £20,000 or less and not be an undergraduate student. These conditions were enforced both for face to face and Prolific sessions.