PhD defence M. (Marco) Gregori

Summary:

Many questions are nowadays sensitive. Examples are: “How often have you flied on a plane in the past 3 months?”. "Would it be a problem for you to work with a transsexual colleague?". Survey participants, who would affirm these questions in private, might not affirm these questions in public when it goes against social norms or it is illegal. 
The first part of this thesis deals with survey techniques to obtain respondents’ truthful answers. The key idea behind the techniques examined here is that we do not actually observe whether a certain person answers yes or no to the questions of interest; in other words, the individual responses are hidden. However, we are still able to estimate the percentage of people who affirm a sensitive question at group level.  The techniques are applicable to many real-life problems in marketing, health, psychology and economics. The second part of the thesis deals with the selection of optimal interventions to stimulate desirable behavior. For instance, researchers may want to identify what might be the best advertisement to promote a health intervention, or the best design to make carbon taxes acceptable to voters. Large samples are required to test many possible interventions. Here we develop an adaptive technique to sequentially test interventions and evaluate the statistical evidence. The technique aims to make this type of studies cheap and effective.