Generalizability of a Cluster Solution Over Time: Results from a Visitor Study at the Kennecott National Historic Landmark

Abstract

This study investigated the generalizability across time of a cluster analysis that identified five groups with differing motivations for visiting a National Historic Landmark. Replication of the initial study after a one-year time period suggested the same visitor types (i.e., clusters) were present in both study years; however, the relative size of the clusters varied by up to ten percent. A group defined by motivations scores in the middle of the 5-point scale exhibited weak recovery in the second year. Two scales constructed for this study exhibited high reliability in both study years, but low generalizability across time. Results provide implications for assessing trends, developing visitor typology-based management standards, and the development of scales.

KEYWORDS:

• Cluster analysis
• cross-validation
• generalizability
• REP scales