I’m finally filling a hole in my list of giants by adding the ‘father of statistical quality control’ (Shewhart. pictured on the left) and his current champion (Wheeler, pictured right).
To be strictly accurate, the timeline linkage would go Shewhart – Deming – Wheeler, but W. Edwards Deming rightfully gets his own ‘giant’ page.
Shewhart was a physicist and statistician. He worked at the famous Hawthorne Works of the Western Electric Company, devising what we now widely recognise as the control chart. This was way back in 1924. He then moved to the Bell Telephone Research Laboratories, where he remained until his retirement.
He understood that every process contains variation, that there were two possible sources of this variation and that it would be important to have a robust method of separating them out so as to understand what is going on and thus form the right conclusions to act upon.
Shewhart was a great friend, mentor and colleague to Deming, with the two spending a great deal of time together deep in statistical discussions. They first met in 1927.
Deming wrote that:
“The great contribution of control charts is to separate variation by rational methods into two sources:
1. The system itself [referred to as noise], which is the responsibility of management; and
2. Assignable [or special] causes specific to some ephemeral event that can usually be discovered to the satisfaction of the expert on the job, and removed.
A process is in ‘statistical control’ when it is no longer afflicted with special causes. The performance of a process that is in statistical control is predictable….”
Shewhart wrote some highly influential books to express his thinking. This work covered far more than control charts. In fact Deming noted that Shewhart is “best known for the least of his contributions”.
You may know the Plan – Do – Study – Act (PDSA) cycle as Deming’s. However, Deming was very clear to attribute this philosophy as the ‘Shewhart learning and improvement cycle’, which combined management thinking with statistical analysis, leading to systems improvement.
Before you think of rushing out and buying either of Shewhart’s most famous books (see below), bear in mind that Deming explained that:
“Although [Shewhart’s] explanations could be simple and clear in a face-to-face discussion, his greatest papers remain as difficult for the reader as they were for him to write. As he told me once, when he writes, he must make it foolproof. I replied in a particular instance that he had made it so foolproof that no one would understand it.”
- The Economic Control of Quality of Manufactured Product (1931); and
- Statistical Method from the Viewpoint of Quality Control (1939)
One of the really big differences between Shewhart and most other statisticians is that he had an ‘operationalist’ outlook – he was concerned with measurement in the real world, so that a system could be improved over time.
Deming later went on to define (i.e. distinguish between) enumerative and analytical studies, which helps make clear what Shewhart was focused on.
- Enumerative studies: focused on a specific frame, and the judgement of results, for the purpose of taking action upon the variables within the frame being studied; and
- Analytic studies: focused on improvement of the system which created, and will continue to create, results. Such studies focus on the problem of prediction into the future – and we are back at control charts.
Deming was making clear that management should be analytical (about improving the system for the future) instead of enumerative (judgement of current results).
Shewhart sounds like he was a fine human being. Deming wrote that:
“[Shewhart] was glad to help anyone. Actually, he never thought of himself as helping anyone: he was simply glad to talk and absorb thoughts from anyone who was genuinely struggling to improve his understanding of the statistical method – interchanging ideas was his way of putting it.”
I really like that he saw it as ‘interchanging ideas’. This fits well with some of my blog posts in respect of dealing with perspectives rather than truths, and knowledge rather than opinions.
If Shewhart’s writings are hard to read, then Donald Wheeler’s work is a delight! Further, he has done a sterling job at righting many of the misunderstandings that have crept into the use of statistics in system improvement efforts – particularly via the Six Sigma movement. One great example would be making clear the correct (and incorrect) calculations of control limits, and if/ when/ how to recalculate them.
Wheeler has spent a distinguished career of over 40 years wonderfully articulating Shewhart’s (and Deming’s) contributions in a clear, concise yet precise manner. He’s delivered seminars around the world, written many excellent books and regularly writes articles and columns.
If you are unsure of his credentials then he:
- was a student and colleague of Deming for 21 years
- is a Fellow of both the American Statistical Association and the American Society for Quality; and
- is a recipient of the ASQ’s ‘Deming Medal’ (joining the likes of Scholtes and Johnson in receiving this honour)
If you are starting out in the world of data and variation, then I’d suggest that the place to start is to read Wheeler’s books. A colleague asked me for advice recently about which Wheeler book to buy, and here’s what I wrote back:
- ‘Understanding Variation – the key to managing chaos’ is (I believe) the one most cited by other authors (such as John Seddon). It’s relatively short, easy to read and gets the points across very well. The book is a sort of ‘management overview’ and an excellent starting place
- ‘Twenty Things you need to know‘ is written as a technical companion-book to partner alongside ‘Understanding Variation’. i.e: if you want to ‘get the point’ then the first book is excellent…but if you want to actually ‘do it’ (i.e. control charting) and do so properly (and avoid some pitfalls) then this technical book is excellent.
- ‘Making Sense of Data’ is a bigger, more involved book that could be bought instead of the first two. It has been written from a ‘service’ (as distinct to manufacturing) perspective, which is rather useful. However, I feel that my reading journey through book’s 1 and 2 beforehand helped me with this 3rd book.
- ….Wheeler has also written a lot of other books on statistical process control, with many focused on manufacturing and advanced statistics.
In addition to these books, you can of course take a look around Donald Wheeler’s website (spc press). The ‘reading room’ tab will take you to his articles and columns.
And to close, Donald Wheeler comes across as a really nice guy, which is always an added bonus 🙂
1. Source of Shewhart info comes mainly from the book ‘The World of W. Edwards Deming’, written by Deming’s long-time assistant Cecelia Kilian. Some chapters of this book are dedicated to honouring the life and work of Walter Shewhart.
2. Source of Wheeler info comes mainly from his (useful) website at www.spcpress.com and his many books.
One thought on “Walter Shewhart (1891 – 1967) and Donald Wheeler”
There is some very nice stuff on the quincunx or Galton board, Pascal’s triangle, Fibonacci Numbers, Golden Ratio at Galtonboard.com I mention this because Don Wheeler shows some nice examples using the quincunx in either Understanding Variation or Normality and the Process Behavior Chart.