A Data-Driven Approach to Development of a Taxonomy Framework for Triple Bottom Line Metrics

摘要

This paper proposes a data-driven approach to develop a taxonomy in a data structure on list for triple bottom line (TBL) metrics. The approach is built from the authors reflection on the subject and review of the literature about TBL. The envisaged taxonomy framework grid to be developed through this approach will enable existing metrics to be classified, grouped, and standardized, as well as detect the need for further metrics development in uncovered domains and applications. The approach reported aims at developing a taxonomy structure that can be seen as a bi-dimensional table focusing on feature interrogations and characterizing answers, which will be the basis on which the taxonomy can then be developed. The interrogations column is designed as the stack of the TBL metrics features: What type of metric is it (qualitative, quantitative, or hybrid)? What is the level of complexity of the problems where it is used? What standards does it follow? How is the measurement made, and what are the techniques that it uses? In what kinds of problems, subjects, and domains is the metric used? How is the metric validated? What is the method used in its calculation? The column of characterizing answers results from a categorization of the range of types of answers to the feature interrogations. The approach reported in this paper is based on a screening tool that searches and analyzes information both within abstracts and full-text journal papers. The vision for this future taxonomy is that it will enable locating for any specific context, discern what TBL metrics are used in that context or similar contexts, or whether there is a lack of developed metrics. This meta knowledge will enable a conscious decision to be made between creating a new metric or using one of those that already exists. In this latter case, it would also make it possible to choose, among several metrics, the one that is most appropriate to the context at hand. In addition, this future framework will ease new future literature revisions, when these are viewed as updates of this envisaged taxonomy. This would allow creating a dynamic taxonomy for TBL metrics. This paper presents a computational approach to develop such taxonomy, and reports on the initial steps taken in that direction, by creating a taxonomy framework grid with a computational approach.