image Dietrich


back to all projects

Opinion Pooling and Bayesian Groups

I explore the aggregation of probabilistic beliefs from several perspectives. Often, I aim for a 'Bayesian group', i.e., for aggregate beliefs that comply with Bayesianism, the standard rationality norm. Falling short of a Bayesian group means violating a common working hypothesis in social science: the hypothesis that households, countries or other groups are rational agents in the same sense as individuals. A methodological question is therefore at stake: is rational-choice theory only a theory of individual agents or also a theory of group agents? A different goal (pursued with Christian List) is to extend opinion pooling theory to 'general agendas', by dropping the unrealistic assumption that the group cares about arbitrarily complex 'composite' events (i.e., all events of an algebra). For instance, a committee of weather experts might form probabilities of 'wind' and of 'sunshine', but not of 'wind or sunshine', be it because this complex event is uninteresting or because members are unable to assess its probability.

Related Work

back to all projects