Ethical Issues in the Dynamic Environment for Public-Stewardship Statistical Information

This paper explores some ethical issues encountered by managers and others in statistical and data science organizations. It does not spell out detailed ethical requirements. Instead, it outlines factors that are important in developing and implementing a practical positive-sum utilitarian approach to ethical conduct for work in public-stewardship statistical information systems. The exploration highlights cases that are encountering rapid changes in their underlying methodological, technological, data and societal environments. Ethical obligations to external information users receive principal attention, with special emphasis on constructive multi-way communication. The development of these ideas begins with a general conceptual framework for ethical work by organizations. Next, the paper analyzes the allocation of benefits, costs and risks that flow from statistical information systems. These issues hinge on decisions about which statistical information to produce; on multiple dimensions of quality, risk and cost in the resulting information production system; and connections of those dimensions with the use value and option value delivered to stakeholders. Evaluation of these issues requires examination of the design and environmental factors that affect the properties of the system, and that shape the tradeoffs at the heart of the resulting ethical decisions. Three running examples motivate and illustrate the principal concepts under consideration.