The focus of this paper is on the microeconometric analysis of spatial
choice in a cross section. Nested multinomial legit models are used t
o analyze the determinants of individual choice among destinations and
vacation activities. Cramer and Ridder's likelihood ratio test for po
oled alternatives in multinomial legit models is sequentially applied
in order to determine the adequate aggregation level of the mutually e
xclusive alternatives in the choice set. The specification test sugges
ted by Chesher and Santos Silva (1992) is used to investigate whether
or not the IIA proposition is fulfilled within the assumed choice subs
ets. (C) 1997 Elsevier Science B.V.