Statistical Analysis in Low Quantity Production of 10 to 20 products each year


We produce 10 to 20 products each year. We measure width of a specific slot on all products. Can we assume these measurements as whole statistical population?
If yes, how can we use these measurements for capability study and sigma level calculation?
We have 45 data of this measure of last products.

20 data points is not a lot of data, and you will need to be crafty with it’s use. You state you are interested in performing a study. Are you trying to minimize variance of the widths for each part? Have you noticed any cycles in the data, month to month, quarterly, etc.?

You stated you have 45 previous sets of data. Is that two years worth of data?

There are some very knowledgable individuals on Hopefully, they will provide more helpful information

Your making only about 20 parts a year tells me that this one set of measurements represents the entire process run. I am guessing that means one series of set ups per part.

SPC and traditionally understood process capability (computed as cpk), however, assumes there will be many parts running on the same setup. So, such elaborate statistics don’t appear to suit your needs.

However, there is a misunderstanding that one needs x-bar and r charts and other elaborate data washing for control and improvement projects like 6 Sigma. Pareto charts, distribution charts, cost analysis and ROI predictions for choosing targeted controls and improvements can serve your needs without a whole lot of statistical gymnastics. The Cove has many tools that can help you. Click on the Post Attachments button in the site header and you can access a long list of tools to help understand and manage process capability.

Thank you for the post.
Hi guys, Im a newbie. Nice to join this forum.