This program sets up an individual decision-making task
in which each person is asked to submit a probability (expressed
in terms of "chances in 100") of an binary event that can be observed
subsequently. The event can be specified by the administrator, e.g.
a particular candidate wins an election, which is then entered manually
by the administrator after it is observed. Alternatively, it can be randomly determined by
the computer. If a random event is used, then subjects see a series of N
realizations from a process with an unspecified probability prior to reporting a
probability for the N+1st realization. In order to encourage people to
pay attention, the participants are asked to predict the binary outcomes of
the first N realizations, with a payment that depends on whether or not the
prediction is correct, in a process that has been used in probability matching
experiments. On trials N+1 and after, they report a probabilily that is
incentivized with either a Lottery Choice Menu, a continuous Becker Degroot Marshack procedure with
a random draw, or a Quadratic Scoring Rule.
The continuous BDM elicitation procedure is explained more fully in
Holt and Smith (2009) "An Update on Bayesian Updating," JEBO.
The choice menu presents subjects with an ordered list of simple binary choices;
see "Belief Elicitation with a Simple Menu of Lottery Choices" (Holt and Smith, 2012).
| || || || || ||Elicited probability assessments can be used to illustrate the effects of possible psychological biases.|
Vecon Lab - September 20, 2020