Alternative statistical techniques for the analysis of customers' satisfaction : factor analysis vs mixtures and shifted binomial models

Abstract

The aim of this dissertation is to study and apply statistical methods which can be used to analyze the categorical data obtained from the Farmers' Market survey. A typical approach in such situations is factor analysis, which we notice throughout our application that it isn't the best option to be used for such an analysis. To overcome this shortcoming we will introduce a mixture of a uniform and shifted binomial (MUB) model, which was originally created to be applied in this kind of analysis. This MUB model will also be used to confirm or contradict the results obtained from our factor analysis. In conclusion, we will consider the relationship between our models and the respondents' profile variables.