Determination of derived importance of a tourism destination: A comparison of indirect methods

Two types of attribute importance, stated and derived, have been studied in marketing and tourism studies using Importance-Performance Analysis (IPA). Derived importance is thought to have an advantage over stated importance in reducing survey fatigue and social desirability biases, and thus is more predictable to an outcome variable. Derived importance is also used in a Relevance-Determinance Analysis to determine higher-impact core attributes or in an Importance Grid Analysis to explore the asymmetric relationship between attributes and overall satisfaction. Since derived importance is usually estimated through statistical methods, it is necessary to determine which method is the most appropriate. However, few studies have examined the appropriateness of indirect methods. To fill this research void, this paper judges the appropriateness of three statistical methods (i.e. multiple regression, partial correlation, and simple regression) based on data collected from Savannah, GA, using a new approach by which results of derived importance were compared against diagnostic attributes in the “keep up the good work” quadrant of IPA. Diagnostic attributes are attributes with higher stated importance and higher predictability of an outcome variable. Results show that urban forests and cultural heritage are such attributes, and simple regression is the best to infer attribute importance.