Experimental Economics (ECON30022)

How the subject works

Most Tuesdays, I run a live experiment in the lecture hall. Students join on their own phones or laptops, then trade, bid, cooperate, or make individual decisions. We look at the results before the session ends. On Friday, I introduce the theory and ask how well it predicts what happened in our class and in published experiments.

All the classroom experiments run on uproot, the open-source framework I develop. I build each one around a question the class can investigate. Some randomly assign different rules to students, giving us a direct comparison. In several experiments, a few students are chosen at random to be paid according to their decisions.

Experiments run in 2026

The 2026 experiments include:

• A framed dictator game in which students are randomly told either to give from their own endowment or to take from the recipient’s, testing whether framing alone changes the final split
• Birnbaum’s number ratings: students rate 9 and 221 in random order; the first ratings give a between-subjects comparison, while the paired ratings give a within-subjects comparison
• Markets: a posted-offer market, a call auction, and a prediction market in which students trade on the outcome of the Cook Islands election
• Cooperation: a prisoner’s dilemma and a public goods game
• Ultimatum and dictator games on fairness and the threat of rejection
• A real-effort task comparing flat pay, piece rates, and tournaments
• Sealed-bid auctions: first-price and second-price sales of gift cards with known face values, followed by a sale where bidders see noisy estimates of an unknown face value
• Macroeconomic expectations with randomly assigned news reports, measuring how information moves forecasts of Australian GDP growth and the share market
• Social norm elicitation with incentives for matching the most common answer in the room
• Risk: Allais’s common ratio problem and an elicitation of risk preferences

Econ Really Matters

Alongside the lectures, I write Econ Really Matters, a series of short, self-contained handouts. Each starts with a model, works through its assumptions, and follows the result into a real decision or policy. The 2026 units include:

• What randomization does and does not identify, from Fisher to blinded drug trials
• Incentive compatibility, and why unpaid survey answers can be more or less informative
• How prices settle competing claims on scarce resources, and what central planning would need to replace them
• Why abstaining from a purchase reduces production, using elasticities and implicit differentiation
• What the equimarginal principle implies for charitable giving, once the welfare objective is made explicit
• Goodhart’s law: why rewarding a measure can worsen the performance it stands for when workers have several tasks
• How a seller who can submit fake bids breaks an auction, and what this means for blockchain fee markets
• Preference falsification, threshold models, and why revolutions surprise observers
• How risk preferences are estimated from choices, and how the model’s assumptions shape what the estimates mean

My aim is for students to leave able to design an experiment, say exactly what its comparisons identify, and explain where the conclusions might fail outside the laboratory.