Correcting for Sample Selection Bias in Stated Preference Tourist Surveys

Abstract

This paper makes novel use of Heckman's sample selection methodology in modelling tourists' choices for cultural heritage attractions. The data used were collected as part of a discrete choice experiment which examined tourists' preferences for two heritage sites: the Knossos Palace and the Heraklion Archaeological Museum, both located on the island of Crete. In particular, the study addresses the possible discrepancies that may have arisen from the fact that the sample population may not have been random, but rather selected systematically as a result of the unobserved effect of the repeat visitation phenomenon. The empirical evidence presented suggests that the correction for sample selection provides a venue for further target audience selection for managers and policy makers. In particular, the main implication to emerge from the analysis is that the two heritage attractions, contrary to the general perception, are part of a tourism demand continuum. As a result, policy makers can capitalize on tourists' different perceptions in their promotion strategies for the two attractions by promoting the Heraklion Archaeological Museum to first-time visitors and the Knossos Palace to repeat visitors to Crete.