Análise multivariada aplicada ao estudo de dados qualitativos ordinais utilizando os métodos de quantificação Ridits e números inteiros ordenados

Abstract

The purpose of this final paper is to compare Ridits and Whole Numbers Ordered quantification methods for ordinal qualitative data analysis using Simple Correspondence and Grouping Analysis multivariate analysis. The work was divided in two articles, the first one conceptualizes the multivariate method of Simple Correspondence Analysis and Whole Ordered Numbers, and the second one conceptualizes the quantification Ridits and Grouping Analysis. The respective methods have been applied to a set of data coming from a research of satisfaction of the users contemplated with the project My Home My Life of the Federal Government, of the habitational set Jardim das Orquídeas, located in Toledo/PR. The criteria evaluated were the internal spatial organization of the house and the surroundings. The first article indicated that the association between variables is high because the X2 test was 242,87 with 44 degrees of freedom for the rooms and 245,36 with 36 degrees of freedom for the surroundings. The two-dimensional percentage map indicated that the service area, bathroom, kitchen and bus lines are the variables that generated the greatest dissatisfaction. In the second article the Ridits quantification also presented strong adequacy for the analysis of this data set, with a cofenetic correlation coefficient very close to the unit. In this method of analysis the same variables mentioned above were grouped by dendrogram. Thus, it is possible to afirm that the different quantification methods associated to different multivariate analysis methods can be used to analyze ordinal qualitative data.