Output details

A rational model of preference learning and choice prediction by children

<22> Originality: Presents the first unified model of preference understanding in children.

Significance: Previously there was no coherent account of how children learn about other agents' preferences. Explains several empirical results; model's predictions inspired studies now underway at the University of Toronto and UC Berkeley.

Rigour: Evaluates a Bayesian account of preference learning using multiple past studies. Sensitivity analysis shows that performance is robust to changes in parameters. No ad-hoc priors or other assumptions: all model elements follow from first principles or the structure of the problems being solved. Published in a peer-reviewed conference with a low acceptance rate (24%).