Control charts are one of the hardest things for those studying six sigma to understand. Here’s an easy Control Charts Study Guide for you. When I was studying for the Six Sigma Black Belt Exam I noticed there were a lot of questions on control charts. Besides that, I noticed that there were a lot of different types of control charts. You had to use one for certain kinds of data or number of units in the sample and then you had to use other control charts in different instances.

Beyond that, there were different distribution types associated with several of the data sets -something you take for granted when asked a question specifically geared to data, but I found it was easy to get tripped up on the exam when combining with control chart questions.

Very confusing.

So, in the spirit of continuous improvement, I made the following study guide. A visual decision tree is under development and will be coming soon.

Control Charts Study Guide Overview

Control Charts – A great place to start for an overview on control charts.

Statistical Process Control – Control charts are an aspect of statistical process control.

Control Charts by Data Type

Not all control charts are the same. Different Data Types require different charts.

Control Charts for Continuous Data

Once you know that you are making a control chart for continuous data, you need to determine if your population is normal or not and the sample size (n) you are charting.

Special Note: If your data is Not Normal

Note that there is some robust conversation in the industry on this. For your application purposes, I suggest you read the following and make your own decisions (if you’re just studying for certification, you can probably skip):

No matter what you decide, I think you could start off just making a basic Run chart and seeing where that brings you.

The Data is Normal

If n = 1, then use: X-MR or I-MR chart: Also known as Shewhart’s Control chart.

Else, If 1 < n < 10, then use: X bar-R chart

Else, If n >= 10, then use: X bar-S chart

Control Charts for Discrete / Attribute Data

For discrete data you will be using one of 4 Attribute charts. Generally you are measuring things like defects. You decide which one to use by noting how many defects are possible per unit measured and if the sample size is constant or not.

When there are Multiple Defects (i.e. chances to be defective):

The population will follow the Poisson Distribution.

If there are multiple defects and sample size is constant, then use: c- chart

  • Plot the number of defects.

If there are multiple defects and sample size varies, then use: u- chart

  • Plot the average number of defects per sample unit.

When there are NOT Multiple Defects:

This population sorts defects into 2 piles (it’s Binomial). Samples are either good or bad, positive or negative, right or wrong.

If there are NOT multiple defects and sample size is constant, then use: np- chart

  • Plot the number of defects.

If there are NOT multiple defects and sample size is varies, then use: p- chart

  • Plot the ratio of defects found. (number of defects found / sample size).

Control Charts Study Guide Videos

Other Helpful Articles

I love MoreSteam’s visual that shows the layering of multiple tools here.


Comments (5)

I’m taking the green belt test next week and the class was an accelerated class (6 weeks)! Ninety percent lecture. I have yet determined a project.

Please correct this :
If your data is Not Normal
Use a Run chart.

If your data is “Not” Normal You want to said : Your data is Normal
Thank you

Thanks for bringing this up, Rachid. There were a number of errors on the page that I’ve found and updated thanks to your comment.

Please check out the new section on non-normal control charts and let me know your thoughts.

Hello Ted,

I like your website and I am visiting it from time to time.

Well, I have a small question, I can’t find its answer either in your website or in other books.

Could you help me to answer it?

Select the best option

Question- A percent defective control chart designed for

Option One- Reduce variation of project

Option Two- Identify the cause of variation that lies out the system

Option Three- Ensure that variation of output is due to common causes

Option Four- Detect the existence at special causes of variation



Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.