QUALITY IN THE FIRST PERSON
The Right Place at the Right Time
by J. Edward Jackson
When I was working on my master’s degree at the Uni-versity of North Carolina (UNC) in the late 1940s, it was an exciting time to be associated with quality control.
World War II had ended only two years earlier and had been responsible for a tremendous advance in the acceptance of the science of quality control and the development of the tools necessary to carry it out.
One course I took at UNC used Eugene L. Grant’s Statistical Quality Control. It was the first edition and probably the first printing. Another class was sequential analysis, which used Abraham Wald’s pioneering book with the same name as the course. Again, it was the first printing; students who took the class before I did had only mimeographed notes.
Although I originally had thought of working in business statistics and forecasting, these two courses introduced me to an exciting new world. I did my thesis on a comparison of various acceptance sampling schemes available at the time. This was before the U.S. defense standard MIL-STD-105 was created.
Interest in quality control was so high in those days I got four job interviews in my hometown of Rochester, NY. I accepted a job as a quality control engineer with Eastman Kodak Co.
Many companies that had government contracts during World War II were required to use some quality control techniques. After the war, some of these companies said, in essence, “Now that the war is over, we won’t have to do these things anymore.” But a great many of them took advantage of the experience they had gained and built on it as the peacetime economy began to develop.
Eastman Kodak was one of these companies. My assignment was in a department that did R&D related to manufacturing and processing color film. Our group helped the processing organization set up control charts and sampling procedures.
Our group leader was very supportive of innovative approaches, and when a colleague and I suggested investigating some of Harold Hotelling’s new multivariate quality control procedures, he encouraged us to proceed.
We ended up combining these procedures with another of Hotelling’s techniques, principal components, to develop new methods for multivariate process control. Our department also designed several experiments to optimize process levels. The nature of the company’s product line also led me into the field of psychometric methods.
In the late 1950s, I returned to school, getting a doctorate in statistics at Virginia Tech. While there, I worked part time in the quality assurance department for the Hercules Powder Co. at the Radford Arsenal. The arsenal produced rockets for the Army such as Honest John, a long-range artillery rocket capable of carrying an atomic or high explosive warhead, and the booster for the Nike-Hercules missile.
These were produced in lots. The Army requested the variability in the lot acceptance testing be broken down into what related to the product and what related to testing and measuring the product. My main assignment was to develop methods to perform this allocation.
One of my other assignments suggested a sequential multivariate acceptance procedure would be appropriate, but none existed at the time. This was the genesis of my dissertation. It was the time of the Cold War, as well as the beginning of the space age, and many of us were able to get our research underwritten by one of several government agencies. Mine was underwritten by the Navy.
I then went to work in a consulting organization that had recently been hired to furnish statistical, mathematical and computing assistance throughout Eastman Kodak, my old stomping grounds. Upon my return to Eastman Kodak, I was gratified the company employed these quality control techniques on one of its government contracts.
By then, emphasis was changing from quality to production capacity (this seemed to be a national phenomenon), and many of my assignments went elsewhere, primarily into psychometrics and market research. Fortunately, that trend reversed, and the last few years before I retired in 1985 resulted in more quality control work again.
It was a joy to work in the group that produced, for ASQ alone, a Shewhart Medalist, Lancaster Medalist, three William G. Hunter Award winners, three Brumbaugh Award winners and 10 ASQ fellows.
After I retired, I consulted for a few years, mostly on work related to multivariate quality control, and wrote a book on principal components analysis, with many examples coming from the field of quality control.
It was a great run, and I would have changed very little. It all started with that quality control course at UNC.
J. EDWARD JACKSON is retired. He has a doctorate in statistics from Virginia Tech in Blacksburg, VA, and is the author of A User’s Guide to Principal Components (Wiley, 1991). He is an ASQ fellow and during his working years was a certified quality engineer. Jackson received ASQ’s Brumbaugh Award in 1978 and William G. Hunter Award in 1994.