Happy New Year! It is hard to believe that this marks the end of the 6th year of publishing our monthly SPC newsletter. This is the 72nd newsletter and our readership has grown to over 6500. A special thank you to all our readers.
As 2009 draws to a close, it is time to review our past newsletters. This will help new readers know and remind our long-time readers what is available. Remember special and common causes of variation? How about the contributions of Dr. W. Edwards Deming? Questions on when to use a certain control chart or what the difference is between variable and attribute control charts? How can a histogram be used to monitor incoming material from a supplier? How do I know if my measurement system is any good? What is process capability and how do I determine it? How do I get to root cause? How do I know if these three processes are the same? All this information and much more is in our newsletters.
The newsletter topics have been divided in the following sections:
One of our objectives has been to make the website a place where people can go for information on SPC and related topics. We believe our newsletters have helped us reach this objective. There is a wealth of information in these newsletters. We hope you enjoy the review. You can view all our newsletters by selecting the "Articles & Newsletters" link to the left.
Happy New Year to each of you and may 2010 be a wonderful year for you and yours!
This takes us back to where everyone needs to begin whenever you are dealing with statistical process control - with variation. To effectively use SPC, you must understand the concept of variation and, in particular, the difference between common and special causes of variation. While variation is really part of each of our newsletters, we have six that address the issue directly. If you don't know the difference between common and special causes of variation, this is where you need to begin. The funnel experiment is a great example of what happens when you confuse special and common causes of variation. The red bead experiment is unsurpassed for helping people understand variation. The February 2009 newsletter contained a free red bead experiment that can be downloaded and run in Excel. The newsletters on variation are listed below.
Any study of variation is incomplete without understanding the teachings of Dr. W. Edwards Deming. Our newsletters include a four-part series on the teachings of Dr. Deming. Dr. Deming was a remarkable man. He is often credited with helping the Japanese rebuild economically after World War II. The Deming Prize is still given to companies for major advances in quality improvement. (Note: permission to use the Media Gallery photo of Dr. Deming granted by Diana Deming Cahill of the W. Edwards Deming Institute).
Dr. Deming spent years developing a theory for helping companies move forward into the twenty-first century. Remarkably, it all still applies today. Understanding Dr. Deming begins with understanding his "system of profound knowledge." This system is composed of four bodies of knowledge:
To understand the system of profound knowledge, you must understand what a system is and what the "aim" of a system is. You must also understand variation and realize that the true benefit of this understanding comes from how you lead people. You must understand the theory of knowledge, and finally you must understand motivation and psychology. You don't have to be an expert in each of these areas. However, you must know something about each area because they are interrelated. Our four-part series addresses each of the bodies of profound knowledge:
Control charts, of course, are the workhorse of statistical process control and represent the topic of many of our newsletters. A control chart is a movie of your process over time. It is the way that your process communicates with you. The control chart will tell you if everything is working as the process was designed (just common cause variation present) or if there is a problem (special cause variation present). All you have to do is listen.
To listen, you must be able to interpret a control chart. This was the first of our newsletters on control charts. The newsletter on control strategies provided a method of looking for the causes of out of control points. The January 2005 newsletter covered how to start and maintain control charts as well as provided insights on what should be charted. The newsletter on the impact of statistical control explains why calculated values of an average, a standard deviation, or a Cpk value have no meaning unless the process is in statistical control. The April 2006 newsletter provided a process flow diagram for selecting the right control chart to use. Our general control chart newsletters are listed below.
Variable control charts are used when you have variables data - data that you can measure to any precision (depending on your measurement system). Examples of variables data include height, weight, density and time. We have covered three variable control charts: Xbar-R, Xbar-s and the individuals (X-mR) control charts. Rational subgrouping is important when using Xbar-R and Xbar-s control charts. There are two newsletters that cover rational subgrouping along with a great exercise that shows the impact of rational subgrouping. In addition, we demonstrated how individuals control charts can be used to handle "rare" data as well as "chunky" data. Our variable control chart newsletters are listed below.
As of this year, we have completed newsletters on each of the four attribute control charts: p, np c and u control charts. Our April 2009 newsletter reviews all four. We also covered how the control limit equations for these charts are not valid in the "small sample case" and how to handle these situations. Our seven newsletters on attribute control charts are listed below.
We have seven newsletters that describe how to use control charts in specific industries or situations:
The Pareto diagram and the histogram are often referred to as bar charts. The Pareto is used to separate the "vital few" from the "trivial many" problems or causes. Histograms are used to estimate the location, spread and shape of a set of data. Our second part on the histogram newsletters showed how you can use histograms to monitor incoming material from suppliers.
Collecting and analyzing data are a vital part of process improvement. It is important to be sure that the data we are collecting are accurate and precise. Our first newsletter on test methods examined how to use control charts to monitor the measurement system. We also have a four-part series on the variables measurement system that covers stability, bias, linearity, and gage R&R. The last newsletter listed below gives a procedure to use to determine the % of variance due to the measurement system.
"Is the process capable of meeting specifications?" This is the key question that must be asked in the end. Does the process really do what we want it to? This is the issue of process capability and involves terms such as Cpk and Ppk. Our three-part series on process capability provides an in-depth explanation of process capability.
One major aspect of problem solving is root cause analysis. What caused the problem to occur? How do we know it is the true cause of the problem? Our newsletters cover scatter diagrams to see if one variable is linearly correlated to another. Also included is how to create and analyze cause and effect (fishbone) diagrams. The Failure Mode and Effect Analysis newsletters resulted in the most visits to our website. We also have looked at modeling a process using linear regression. ANOVA was introduced in our December 2008 newsletter.
On occasion, you have the need to compare two or more processes. Are the processes operating at the same average? With the same variance? The newsletters listed below describe how this can be done.
We also have offered a variety of miscellaneous topics. These include an explanation of standard deviation as well as skewness and kurtosis. A ten-step problem solving model, a method for data collection and a model for process management have been covered. The normal distribution and how to make a normal probability plot have also been introduced in our newsletters.
We have reviewed our on-line newsletters and hope you find this information useful. If you have any comments about our newsletters or have suggestions for future topics, please let us know.
Thanks so much for reading our publication. We hope you find it informative and useful. Happy charting and may the data always support your position.
Dr. Bill McNeese
BPI Consulting, LLC