The human development index : a search for a measure of human values
The thesis investigates methods of evaluating indexes that measure concepts of human values. My understanding of indexes, especially on how they relate to the real world and concepts (that are the objectives of the measurement), is influenced by my study of literature on models used in economic and in physics. We learn from this study of models the following: (1) regularities described in theories do not represent real world phenomena, which consist of many different forces acting simultaneously; (2) but such regularities are true in models, because they describe specific conditions under which regularities in nature are displayed; (3) there are more than one model that can represent the same phenomenon depending on which particular aspect of the phenomenon to focus on; and (4) the success of a model has to be evaluated partly by criteria that are independent from theoretical ones. Since the role indexes play in relation to real world and concepts are similar to the role models play in relation to theories, I have applied the above knowledge to propose the following three criteria to evaluate successful indexes: (1) Purpose-dependent criteria: criteria that are based on particular motivations of the measurement project; (2) Theory-dependent criteria: criteria that are reflected in the theories that expressly or implicitly guide the development of the project of measurement; and (3) Conditions-dependent criteria: criteria that are based on the conditions under which the index measures what it is designed to measure. I apply these three criteria of successful indexes to examine two projects of measuring human values, one called the Human Development Index developed by the United Nations Development Programme and the other called the Life Satisfaction Indicator developed by an officer at the Economic Planning Agency in Japan. Among the findings from the examination of those two indexes are that they can be the products of a mixture of concerns that include convenience, conventions, practicality, politics and consistency with relevant theories, and some of these concerns may conflict with each other. Another important finding is that because there are many assumptions made and simplifications applied in order to choose a quantitative representation of a human value, the application of the measure is limited. I conclude that both in using and in evaluating indexes of human values, it is important that we are aware of such limitations, so that we can more effectively know both how to avoid misusing the indexes and how to improve them over time.