Nolan, T. W. and L.P. Provost. 1990. Understanding Variation. Quality Progress (May): 70-78
Summary by Jennifer Ryan
Master of Accountancy Program
University of South Florida, Fall 2000
Deming Main Page | SPC Main Page
BASICS OF VARIATION
There is variation in all aspects of our lives.
Typical losses resulting from misinterpretation of variation:
● Blaming people for problems beyond their control.
● Spending money for new equipment that is unneeded.
● Wasting time looking for explanations of a perceived trend when nothing has changed.
● Taking other actions when it would have been better to do nothing.
Process – set of causes and conditions that repeatedly come together to transform inputs into outcomes.
System – interdependent group of items, people, or processes with a common purpose.
Note: As used below, the term "process" is taken to mean "process and/or system".
Indicators of performance, or quality characteristics, of any process can be identified and measured. Some examples are number of accidents, number of billing errors, and time of delivery. All vary over time. Often variances are used for action on the process, which is sometimes not appropriate.
The traditional view of variation is that it is either good or bad. This method is used for inspection of products and services and for grading in schools. One shortfall of classifying variation as good or bad is that there is no information on the causes of the variation.
Walter Shewhart developed the concept of variation based on the idea that a quality characteristic has two types of causes:
● Common causes – causes that are always part of the process and affect everyone working in the process.
● Special causes – causes that are not part of the
process normally, or do not affect everyone,
but are attributable
to specific circumstances.
To determine whether the cause of variation is common, Shewhart created the control chart, which consists of three lines and points plotted on a graph. The points reflect measurements of a quality characteristic at regular intervals. Two of the lines represent the upper and lower control limits, which are above and below, respectively, the center line or average.
Concepts were developed for manufacturing, but can be used for processes of management, administration, and service.
A process with only common causes affecting the outcome is stable, or in a state of statistical control. Stable means only that the variation is predictable within statistically established limits, not that there is no variation.
An unstable process is one where the variations can be attributed to both common and special causes.
W. Edwards Deming describes the benefits of a stable process as:
● A process has an identity and its performance is
predictable; therefore there is a
rational basis for planning.
● Costs and quality are predictable.
● Productivity is at a maximum and costs are at a minimum for the process.
● The effect of changes in the process can be measured faster and more reliably.
If a stable process is adjusted, perhaps due to misunderstanding of the variances, variation will increase rather than decrease. (See the funnel experiment for more on the effects of tampering).