A more direct approach would consist of estimating the developed information
search model on observed stated choice data. The difference with the experimental
economics perspective lies in the fact that stated choice-experiments do not
induce preferences; instead, they are designed to infer preferences from choices.
In order for the experiment to ‘match’ the developed model, participants must
face a number of alternatives with uncertain attributes, and be given the opportunity
to either acquire information or directly execute an uncertain alternative.
Chorus et al. (2007a) provide an example of such an experiment in the context
of travellers’ route- and mode-choices. Having collected data on choices among
alternatives as well as on information search patterns, our model can be tested
by estimating separate parameters (e.g. a parameter reflecting travel time importance
in a travel context) using the model of choice among alternatives on the
one hand, and the information search model on the other hand. Statistical analysis
of the two estimated parameter sets may then determine whether or not the
estimated parameters differ between the two models. If the two parameters sets
do not differ in a statistical sense, this implies that for the given dataset, inferring
preferences from information search yields the same results as conventional choice
model estimation. This would be evidence of the validity and applicability of the
proposed model.
A more direct approach would consist of estimating the developed informationsearch model on observed stated choice data. The difference with the experimentaleconomics perspective lies in the fact that stated choice-experiments do notinduce preferences; instead, they are designed to infer preferences from choices.In order for the experiment to ‘match’ the developed model, participants mustface a number of alternatives with uncertain attributes, and be given the opportunityto either acquire information or directly execute an uncertain alternative.Chorus et al. (2007a) provide an example of such an experiment in the contextof travellers’ route- and mode-choices. Having collected data on choices amongalternatives as well as on information search patterns, our model can be testedby estimating separate parameters (e.g. a parameter reflecting travel time importancein a travel context) using the model of choice among alternatives on theone hand, and the information search model on the other hand. Statistical analysisof the two estimated parameter sets may then determine whether or not theestimated parameters differ between the two models. If the two parameters setsdo not differ in a statistical sense, this implies that for the given dataset, inferringpreferences from information search yields the same results as conventional choicemodel estimation. This would be evidence of the validity and applicability of theรูปแบบนำเสนอ
การแปล กรุณารอสักครู่..
