This paper brings together several research streams and concepts that have
been evolving in random utility choice theory: (1) it reviews the literatur
e on stated preference (SP) elicitation methods and introduces the concept
of testing data generation process invariance across SP and revealed prefer
ence (RP) choice data sources; (2) it describes the evolution of discrete c
hoice models within the random utility family, where progressively more beh
avioural realism is being achieved by relaxing strong assumptions on the ro
le of the variance structure (specifically, heteroscedasticity) of the unob
served effects, a topic central to the issue of combining multiple data sou
rces; (3) particular choice model formulations incorporating heteroscedasti
c effects are presented, discussed and applied to data. The rich insights p
ossible from modelling heteroscedasticity in choice processes are illustrat
ed in the empirical application, highlighting its relevance to issues of da
ta combination and taste heterogeneity. (C) 1999 Elsevier Science S.A. All
rights reserved. JEL classification: C25.