Extracting relevant structures from self-determination theory questionnaires via Information Bottleneck method

Abstract: In this paper, we introduce the application of Information Bottleneck as a method to investigate properties of questionnaires developed for the study of motivational profiles based on self-determination theory. Founded on information theory, the Information Bottleneck method compresses multidimensional categorical data into clusters by minimizing the loss of relevant information. It does not require linearity such as Pearson correlation. For exploratory data analysis, it allows investigating the structure of questionnaire items grouping with a hierarchical partition as perceived by the sample, based only on conditional probability distributions calculated from answers to the questionnaire items. We applied the approach to an instrument adapted from the academic motivation scale. The Information Bottleneck approach showed a partition of items aligned with the self-determination continuum. The results of this first application of Information Bottleneck to psychometric studies suggested that this procedure can could be useful as a consistent, complementary analysis for traditional exploratory procedures such as factor analysis.