Determining Long-Term Change in Tourism Research Language With Text-Mining Methods

Two text mining methods prove to be capable of detecting significant change in language between early and recent article abstracts and advocate future qualitative analyses for pursuing the reasons and contents of change.

Abstract

Are quantitative text mining methods sensitive enough to recognize change in the language of tourism research? The study of tourism is expected to have shifted focus during the past four decades, and this must be reflected in the abstracts of articles published in a journal of particularly long tradition. Two text mining methods are employed for analyzing change. They prove to be capable of detecting significant change in language between early and recent article abstracts. The study investigates discriminant word items and latent topic structures. The double approach with two computationally unrelated methods (penalized support vector machines and latent Dirichlet allocation) explores (a) single word items that differentiate between earlier and later article abstracts and (b) the relevance of latent topics underlying older and newer abstracts. The results advocate future qualitative analyses for pursuing the reasons and contents of change.