Good-Bye for Now…

As this is my last column for Quality Digest, I’m delighted to announce that my distinguished predecessor, Donald J. Wheeler, will be writing this column again. You’ll be in good hands.

When I got my master’s degree in statistics in 1980, jobs were plentiful for internal statistical consultants in corporate research and manufacturing. Their jobs involved educating scientists and engineers in the power of applied statistical methods, especially experimental design. Excellent research in applied methods was performed and collegially disseminated by distinguished groups, most notably DuPont, Kodak, and General Electric. I was part of a strong internal group at 3M. Those groups, for all intents and purposes, have now vanished.

Statistical process control (SPC) became more dominant during the late 1980s. Everything became “bigger, better, faster, more, now!” Suddenly there wasn’t time to do response surface experimental designs, except maybe 2² or 2³ factorials and a replication--if you were lucky. See “Using Design of Experiments as a Process Road Map” ( www.qualitydigest.com/feb06/articles/02_article.shtml [1] ).

The influence of W. Edwards Deming peaked, and people misinterpreted his philosophy to mean that everyone must be trained in statistics, resulting in training juggernauts that were (and continue to be) legalized torture.

Quality guru wars began as people argued Deming vs. Juran vs. Crosby. Seemingly overnight, personal computers were easily available, as was statistical software to do cumbersome calculations, and more statistical software that did the “thinking” and “analysis” for you. From this statisticians became their own worst enemies. They turned up their noses at such “insultingly simple” material and deluded themselves that people would still “beat down their doors.” In 1984, I saw some correspondence between Deming and the distinguished applied statistician Gerry Hahn: “Gerry, Sorry about your misunderstanding--TOTAL! When will statisticians wake up?”

The statistical work culture nurtured since the 1950s had allowed statisticians to flourish as “corporate eggheads” who would occasionally save the corporation big time with an opportune experimental design. Eventually, the leaner economic corporate environments of the 1980s and 1990s would no longer tolerate this. In essence, degreed statisticians put themselves out of business in industry.

Then Deming died in 1993. Six Sigma appeared, creating even more statistical training juggernauts... by nonstatisticians. Then came lean, then lean Six Sigma, and more guru vs. guru wars. See “TQM, Six Sigma, Lean and… Data?” ( www.qualitydigest.com/july06/departments/spc_guide.shtml [2]).

Deming would roll over in his grave if he could see the ubiquitous “hacks” (his term) who have been trained and deemed competent in statistical methods. As a friend of mine once said, “Would you rather your children had sex education or sex training?” I hope I’ve given you some good statistical “education” these past four years through greater awareness, understanding, and expansion of your concept of variation. That is the key to true improvement.

I stand by the list published in my November 2006 column, “A Quality Professional’s Holiday Wish List” ( www.qualitydigest.com/nov06/departments/spc_guide.shtml [3] ), which includes books with much-needed change-agent skills for today’s quality practitioner. I would add one book, Statistical Thinking: Improving Business Performance , by Roger Hoerl and Ronald Snee (DUXBURY Press, 2001), which includes clear explanations of traditional basic statistics.

So, where is the quality profession going? I’m not sure, but it’s in serious transition. I’ve done my best to summarize a transformational approach where the words “statistical” and “quality” will be dropped as adjectives because they will be givens and integrated into the leadership/management organizational structure. I’ve just submitted a book manuscript to be published in April 2009 titled, Data Sanity: A Quantum Leap to Unprecedented Results. It also concerns the leadership and psychology issues of transforming an organizational culture.

Remember: First, whether or not people understand statistics, they are already using statistics; and second, don’t teach people statistics--teach them to solve their problems.

Thanks for reading… best wishes… please keep in touch …and Godspeed.

Editor’s note: Davis’ column, “SPC for the Real World,” first appeared in Quality Digest in January 2005. A heartfelt “Thanks!” to you, Davis, for bringing your wit and unique writing to our readers.