Research
Three strands, and they feed each other. The first measures preferences. The second asks whether the measurement is picking up what we claim it is. The third takes whatever survives the first two and puts it in front of someone who has a decision to make.
Preferences and behavioural economics
Discrete choice experiments across organ transplantation, diabetes, prostate cancer, dementia, loneliness, telehealth, medical devices, health workforce and vaccination. The recurring interest is in where observed choices depart from the standard model: heterogeneity that is not noise, inconsistency that is not error, and equity concerns that a simple sum of individual utilities cannot represent.
Attention and information processing
A choice model assumes people read the alternatives. Eye tracking says they often do not, and that what they skip is patterned rather than random. Attribute order alone moves estimates. This strand brings process evidence into preference measurement, so that a coefficient can be read alongside evidence about whether anyone looked at the attribute it belongs to.
Measurement, outcomes and policy
Whether the outcomes we measure are the ones people care about, and whether the evidence reaches the person deciding. In practice this means economic evaluation, outcome measure development, and building decision-support tools that carry the estimates rather than summarising them in a report nobody opens.
Projects
- MandEval, the effects and side effects of Australia's COVID-19 vaccine mandates, and what mandate designs the public will accept.
- FETP business case, scaling field-epidemiology training, with the World Bank.
- Attribute ordering and attention, what moves when you reorder a choice task.
- Health preferences, the choice experiment programme across clinical areas.
- Decision-support tools, turning estimates into instruments.
Also here
- Impact, where the work has been used.
- Data and code, public repositories.
- Grants, funded programmes.
- Papers, the full list.