Online Program Home
My Program

Abstract Details

Activity Number: 437 - Deming Lecture
Type: Invited
Date/Time: Tuesday, July 31, 2018 : 4:00 PM to 5:50 PM
Sponsor: Deming Lectureship Committee
Abstract #330989 Presentation
Title: Improving the Quality and Value of Statistical Information: Fourteen Questions on Management
Author(s): John L. Eltinge*
Companies: United States Census Bureau
Keywords: administrative record data; big data; incentives; integration of multiple data sources; matrix management; organic data; sample surveys; non-designed data; system of profound knowledge; total survey error model; transparency, reproducibility and replicability; use value and option value of public goods
Abstract:

In his landmark book, Out of the Crisis, Dr. W. Edwards Deming presented "Fourteen Points for Management." Taken as a whole, his points reflected a profoundly humane and nuanced perspective on improved management of large and complex organizations. The exposition and application of those fourteen points centered primarily on management of quality and productivity in manufacturing, but also were informed by Dr. Deming's extensive experience with sample surveys. In addition, his commentary on these points included substantial discussion of work by government agencies and other organizations that produce statistical information for public use.

In the decades since publication of that book, economic and societal conditions have changed remarkably in much of the world. This has led to corresponding changes in the environment for statistics, including the improved understanding of statistical concepts and methods by some segments of the public; expansion of relevant data sources; changes in the processes for allocation and management of resources for statistical work; and dramatic growth in the methodology and technology available for the capture, integration and use of statistical information from multiple sources. Taken together, these changes present an extraordinary opportunity for the statistical community to improve the quality and value of information that we provide to the public.

This paper suggests that Deming's "fourteen points" reflect fundamental concepts and principles that can offer important insights for the statistical community as it responds both to the abovementioned opportunity, and to closely related challenges. Initial sections provide brief reviews of some of the relevant literature on data quality, survey methodology and public use of statistics; and present schematic models that may help in the assessment of the quality, risk and cost profiles for statistical information. Subsequent sections consider each of Deming's fourteen points, placed into three inter-related groups:

(A) systematic approaches that link stakeholder value with quality, risk and cost; and continuous improvement thereof (per Deming's points 1-5);

(B) use of the framework from (A) to produce fundamental improvements in management structure and function (points 9-12); and

(C) the critical role of humans, and the talents and leadership thereof (points 6-8 and 13-14)

The discussion links each of Deming's points with an underlying management issue; explains the relevance of that issue to the current statistical environment; and presents a question that may be fundamental in addressing that issue. Current and prospective work with integration of multiple data sources provide illustrative examples for each of the fourteen questions. These examples also indicate the importance of in-depth exploration of both management issues as such, and underlying scientific, methodological and technological problems. Taken together, these ideas are intended to further Deming's stated goal of "work that offers opportunity to create new statistical methods, or to use existing methods . to improve . research; or to acquire new knowledge .; or to improve efficiency, uniformity, quality service, and performance of product; or to achieve smoother operation and more effective administration and management in industry and government."


Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program