The use of the CHAID algorithm for determining tourism segmentation: A purposeful outcome

Heliyon. 2020 Jul 8;6(7):e04256. doi: 10.1016/j.heliyon.2020.e04256. eCollection 2020 Jul.


This paper considers the most suitable market segment(s) from an environmental and local economic development perspective in the specific context of visits to natural environments. More specifically, the paper explores the distinctions and differences between tourists (non-residents) and residents with regard to visit behavior at natural attractions. By using the CHAID algorithm, a decision tree is constructed for means of transportation which serves as a key factor in the segmentation process. However, such a tree for visitors' resident or non-resident status cannot be built as a first explicative variable, unless it is statistically forced. Once it is forced, the tree opens in several sub-segments, for non-residents and residents alike. Finally, it allows understanding of the means of transportation used by visitors according to their geographical origin as well as a set of added independent variables: accommodation establishment, length of stay, season, and other demographic variables (educational level, gender, and age). Also, more importantly, we have obtained segments with no overlap configured according to all the aforementioned variables. This is a very strong result from a methodological point of view and for policy makers in destination settings.

Keywords: Accommodation; CHAID algorithm; Canary resident/non-residents; Economic impacts; Environmental economics; Environmental impacts; Marketing; Means of transportation used; Statistics; Stays; Tourism; Tourism economics; Tourism market segmentation.