When an Average Isn’t the Average

Quick Links

SPC for Excel Software

Visit our home page

SPC Training

SPC Consulting

Ordering Information

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

View Bill McNeese

Connect with Us

Comments (8)

  • AnonymousJune 28, 2013 Reply

    Nice work and commentary 

  • AnonymousJune 28, 2013 Reply

    The infirmation was clear and informative. more please.Mike

  • AnonymousJune 28, 2013 Reply

    Hi,You have provided an excellent example of plotting the data. Many are affraid to see words like "statistical" and  "control". This is will be useful to engage them in how to handle their data. 

  • Ashok December 15, 2016 Reply

    Good Info Bill. I am sure 90% of the companies are using wrong data to calculate the Cpk values.

  • Tejas ShahAugust 3, 2018 Reply

    Till date, I merely though that if data is within UCL & LCL that mean that prcess in statistical control. In above example of temperature I though that you have used different rule set to define wheather priocess is in statistical control or not. May i know that rule set?

  • ChebetzApril 23, 2020 Reply

    Dr. McNeese – Let me commence by saying this was a great article! I thoroughly enjoy reading these. I hope you’re still willing to answer questions from people about this post, because I have some. I am hoping you can confirm my understanding of the ideas you presented herein.To help me better understand the concepts you presented here, I took your suggestion and plotted the max temperatures in Minitab 18. The first chart below (if I did this correctly), shows Houston’s average max temperature to be 92.36 degrees. Mother Nature’s process is not in control with multiple data points failing statistical tests. In this example, to say that the average max temp is 92.36 degrees would be incorrect (Is incorrect that the right word? Would misleading be the better word?) because the data is not homogenous. There are multipe changes in the process, indicated by where the data points fail the tests. So, to get a better understand of what is really going on, I need to break this chart up into multiple stages to indicate where the process changes, right?So, I did that (see chart #2). I broke the chart in years ’71-’75 (because those were the 5 consecutive data points that had 4 points below the first standard deviation) and again in years ’08-’12 (because those were the 5 consecutive points that had 4 points above the first standard deviation). Again, assuming I did this correctly, would I be correct to infer from this data that Houston’s more accurate average temperature would be 96.36, because that is the average of the last stage and is closest to now?

    • billAugust 17, 2019 Reply

      Thanks for your kinds words.  Yes I think your analysis  correct.   I did not check the math. The more accurate temperature currently is the last stage.

Leave a Reply

Your email address will not be published. Required fields are marked *