Dimensionality of the Causes of Churning: A Multivariate Statistical Analysis

Abstract

The present study was conducted in Portugal and had, as its object of study, workers from Portuguese companies belonging to several sectors of activity. The main goal of this study was the identification of the dimensions related to the causes of churning and to analyse its applicability in the management of human resources to promote individual and corporate welfare. Its specific targets were (a) to make a sociographic characterisation of the workers; (b) to make their professional characterisation; (c) to analyse the perception workers had in relation to the selected dimensions under study. Through the gathering of data per questionnaire, a sample consisting of 801 answers was considered. First, we resorted to a multivariate statistical analysis through the application of an Exploratory Factor Analysis (EFA) that allowed for the selection of the most relevant dimensions, followed by a descriptive statistical analysis on the collected sample and used items. Finally, we resorted to a TwoStep Cluster analysis that allowed for the identification of two Clusters of workers with a differentiated probability for the occurrence of churning.