Parametric and Nonparametric Statistics for Sample Surveys and Customer Satisfaction Data

This book deals with problems related to the evaluation of customer satisfaction in very different contexts and in many different ways. Analyzing satisfaction is not an easy issue since it represents a complex phenomenon which is not directly measur- able. Often satisfaction about a product or service is investigated through suitable surveys which try to capture the satisfaction about several partial aspects which characterize the perceived quality of that product or service. In this book we present a series of statistical techniques adopted to analyze data from real situations where customer satisfaction surveys were performed. The aim is to give a simple guide of the variety of analysis that can be performed when analyze data from surveys. Experiencing satisfaction when customers buy products or services is an index of how a company is operating. Are goods or services appreciated by customers? Are customers satisfied with goods or services of a specific company? How they respond to similar questions is a crucial point in order to evaluate and analyze the answers. For this purpose preference evaluation methods are good candidates to understand how customers react to the evaluated items. A promising and theory-based method is called CUB model. In the first chapter of this book CUB model has been adopted in order to evaluate two latent variables, feelings and uncertainty, that are supposed to be involved in the choice process of an item. The application field refers to a genuine study conducted in Italy and in Austria about the satisfaction level of customers about food packaging at the grocery store. The second chapter deals with the concept of heterogeneity in satisfaction. Iden- tifying customer groups characterized by “within homogeneity” and “between het- erogeneity” could be a useful starting point of market segmentation. In this chapter the main heterogeneity indices are introduced and testing methods for comparing the satisfaction heterogeneities of two or more customer populations are described also with different practical examples. In the field of satisfaction assessment it is quite common that the final objective is obtaining an appropriate ordering of different products or services under compar- ison. From a statistical point of view the issue of ranking several populations from the best to the worse on the basis of one or more aspects of interest is not so easy. In the third chapter of this book different examples of contexts where the problem of ranking occurs are described and a nonparametric inferential approach is presented with application to the field of food sensory analysis. Another way to assess satisfaction is represented by the so-called composite in- dicators which aggregate different dimensions of satisfaction into a single overall indicator. How to suitably compute such indicator is the topic of Chap. 4. In this chapter the construction of a composite indicator is discussed in general and a non- parametric composite indicator which includes different benchmarks of satisfaction is developed. The properties of the proposed indicator are shown by analyzing data from a university students’ satisfaction survey. Finally Chap. 5 describes some rank-based procedures for analyzing surveys data with the help of a useful R package.