Lifestyle, decision-making, and behaviour: econometric analysis and experiments using wearable and online chess data

Possible project directions

You will develop an independent thesis within this research agenda. Possible thesis directions include the effect of sleep quality and circadian disruption on decision-making, how exercise and recovery shape next-day cognitive performance, environmental shocks and behavioural outcomes, heterogeneous effects of lifestyle factors, and causal machine-learning approaches for personalised behavioural recommendations. The project is likely to use panel methods, fixed effects and two-way fixed effects designs, difference-in-differences, instrumental variables, quasi-experimental methods, and experimental or RCT designs.

Required skills and experience

The project is especially suited to a candidate with strong applied microeconometrics skills, strong R or Python capability, and a genuine preference for AI-integrated research workflows. The project workflow will involve reproducible coding, local file-based analysis, version control, documentation, agent-assisted coding and research operations, and careful verification of statistical and AI-generated outputs. Experience using tools such as Claude Code, Codex, or multi-agent coding/research pipelines is highly desirable, especially where applicants can demonstrate judgement around verification, human review, model comparison, and reproducibility rather than simple prompt-and-response use.

Experience with designing and running experiments/RCTs, behavioural data collection, or large health, wearables, sleep, or platform datasets is desirable, but not required. Chess knowledge is welcome but not required. The strongest fit is likely to be a candidate with deep quantitative training, strong coding habits, and curiosity about how AI-assisted workflows can improve empirical research when used with human oversight and methodological discipline.

You will be based in the School of Economics at The University of Queensland and supervised by Associate Professor David Smerdon. The project is connected to active collaborations with WHOOP and Chess.com and sits within a broader ARC DECRA program on lifestyle, decision-making, and behaviour.

You will receive training in applied microeconometrics, causal inference, experimental and quasi-experimental design, and the analysis of large-scale behavioural and biometric datasets. The research environment will emphasise reproducible, transparent, and computationally sophisticated workflows, including R/Python-based analysis, version control, local file-based coding, research documentation, and AI-assisted coding and review practices.

Depending on the final thesis topic and data-access arrangements, the project may involve collaboration with researchers and data teams across economics, psychology, health, wearable technology, and online platforms.

You will be encouraged to build a distinct research identity while contributing to a larger interdisciplinary project, with regular attention to code review, robustness checks, documentation, and responsible use of AI tools in empirical research.