Understanding asymmetric effects of attribute performance on tourist satisfaction with island tourism using user-generated data
GuangMeng Ji,Siew Imm Ng,Jun-Hwa Cheah,Wei-Chong Choo
DOI: https://doi.org/10.1108/jhti-06-2023-0416
2023-12-02
Journal of Hospitality and Tourism Insights
Abstract:Purpose Past research often relies on linear relationship assumptions from the perspective of managers when studying the relationship between attribute performance and satisfaction. However, this study extracts tourists' online reviews to explore asymmetric relationships and identifies island tourism satisfiers, hybrids and dissatisfiers. Design/methodology/approach The research uses 3,523 reviews from Tripadvisor to examine Langkawi Island's tourist satisfaction. Latent Dirichlet allocation (LDA) machine-learning approach, penalty–reward contrast analysis and asymmetric impact-performance analysis (AIPA) were employed to extract and analyse the data. Findings Langkawi's dissatisfiers included "hotel and restaurant", "beach leisure", "water sport", "snorkelling", "commanding view", "waterfall", "sky bridge walk", "animal show", "animal feeding", "history culture", "village activity" and "duty-free mall". Amongst these, five were low performers. Hybrids encompassed "ticket purchasing", "amenity" "traditional food market" and "gift and souvenir", all of which were low performers. Only one attribute was categorised as a satisfier: "nature view" which performed exceptionally well. Practical implications This study provides recommendations to enhance tourist satisfaction and address tourist dissatisfaction. The elements requiring immediate attention for enhancement are the five low-performance dissatisfiers, as they represent tourists' fundamental expectations. Conversely, the satisfier or excitement factor (i.e. nature views – mangroves and wildlife) could be prominently featured in promotional materials. Originality/value This research constitutes an early endeavour to categorise attributes of island tourism into groups of satisfaction, hybrid or dissatisfaction based on user-generated data. It is underpinned by two-factor and three-factor theories.