Prioritization of Issues and Requirements by Cumulative Voting: A Compositional Data Analysis Framework

Abstract

Cumulative Voting (CV), also known as Hundred-Point Method, is a simple and straightforward technique, used in various prioritization studies in software engineering. Multiple stakeholders (users, developers, consultants, marketing representatives or customers) are asked to prioritize issues concerning requirements, process improvements or change management in a ratio scale. The data obtained from such studies contain useful information regarding correlations of issues and trends of the respondents towards them. However, the multivariate and constrained nature of data requires particular statistical analysis. In this paper we propose a statistical framework; the multivariate Compositional Data Analysis (CoDA) for analyzing data obtained from CV prioritization studies. Certain methodologies for studying the correlation structure of variables are applied to a dataset concerning impact analysis issues prioritized by software professionals under different perspectives. These involve filling of zeros, transformation using the geometric mean, principle component analysis on the transformed variables and graphical representation by biplots and ternary plots.