Published online by Cambridge University Press: 22 March 2010
In many stated choice experiments researchers observethe random variables Vt,Xt, andYt =1{U +δ⊤Xt+ εt <Vt},t ≤ T, whereδ is an unknown parameter andU andεt are unobservablerandom variables. We show that under weakassumptions the distributions of Uand εt and also theunknown parameter δ can beconsistently estimated using a sieved maximumlikelihood estimation procedure.
We are grateful to Bo Honoré, the referees, andthe coeditor Jinyong Hahn for helpful comments.Mogens Fosgerau has received support from theDanish Social Science Research Council.