Experiences in Succession: An Analysis of Longitudinal Consumer Upgrading with Hedonic Experiences

摘要

How do consumers upgrade hedonic experiences over time and what are the characteristics of previous purchases that influence that behavior? We analyze our data using Generalized Estimated Equations (GEE) to estimate parameters for both of our datasets. We assume an exchangeable correlation matrix which is most appropriate for the inherent unequal spacing of purchase i timing and unbalanced number of purchases consumers make. We also specified a Gaussian family distribution and an identity link (Diggle, Liang, and Zeger 1994).