DECONVOLUTING PREFERENCES AND ERRORS: A MODEL FOR BINOMIAL PANEL DATA

In many stated choice experiments researchers observe the random variables <italic>V</italic>, <italic>X</italic>, and <italic>Y</italic> = 1{<italic>U</italic> + <italic>δ</italic>^⊤<italic>X</italic> + ε<italic>null</italic> < <italic>V</italic>}, <italic>t</italic> ≤ <italic>T</italic>, where <italic>δ</italic> is an unknown parameter and <italic>U</italic> and ε<italic>null</italic> are unobservable random variables. We show that under weak assumptions the distributions of <italic>U</italic> and ε<italic>null</italic> and also the unknown parameter <italic>δ</italic> can be consistently estimated using a sieved maximum likelihood estimation procedure.