Abstract
Titre : Analyse de Khi-carré : fondements, démarche, intérêt et limites
Auteur (s) : Eugénie Kabali Hamuli et José Mangalu Mobhe Agbada
Abstract :
As one of the most widely used data analysis techniques, the Chi-square test unfortunately remains very poorly applied and above all very misinterpreted. Also, in this article, the authors propose to review, in a new and contextualized light, not only the theoretical presuppositions and conditionalities relating to the application of this test, but also, the interest, limits and pitfalls to avoid in its implementation and in the interpretation of the results that come from it. While most books and other articles on the subject are mainly limited to the study of independence Chi-square, in this article it also has other types of Chi-square test such as fit Chi-square and homogeneity Chi-square. In addition, beyond the theoretical discussions, illustrations and concrete examples are proposed to allow readers to put themselves in a situation and apply the proposed theoretical approach. This approach seems more important since the Chi-square test is, in most cases, a prerequisite for the use of the most robust statistical tests such as logistic regression, discriminant analysis, etc.
Editeur : RECOSH
Télécharger l'article complet ici