Correspondence Analysis: Theory, Practice and New Strategies

Foreword

Correspondence analysis aims to lay bare the structure underlying two-way and higher way contingency tables and so allows us to explore the nature of the association between two or more categorical variables. It is not one single technique but has many variants. The authors of this book counted 26 for two-way tables alone, thus not even including those variants that are used for multiway tables. This versatility is fascinating, and led the French statistician Jean-Paul Benzécri to remark that it is

[la] méthode qui bien mieux que tout autre nous a permis de découvrir les faits de structure que recèle un tableau de données quel qu'il soit [the method which, better than any other, allows us to uncover the actual structure which underlies a data table, whatever it contains].

As Beh and Lombardo convincingly show in their book, correspondence analysis has been the subject of a never-ending stream of publications, both theoretical and applied. They mention bibliographies (until 2012) that list over 7000 publications with the term ‘correspondence analysis’ in the titles, abstracts and/or keywords. The consequence of this large body of literature is that a complete book is needed to do full justice to correspondence analysis as a tool for data analysis. Thus, it will come as no surprise that several monographs on correspondence analysis have appeared over the years.

It is impossible to speak about correspondence analysis without mentioning the French connection, because ...