Can you use SPC with discrete data

Can you use Statistical Process Control charts when your measurement data is rounded to whole numbers? That was a great question I got from Charlie, and the short answer is: yes—but only if you meet some conditions.

If your data has at least 8 distinct categories, you’re usually fine to treat it as continuous for SPC and many other statistical tools. So even if your machine rounds off pressure test results to integers, that doesn't stop you from charting the process. It’s all about how much resolution you still have. I explain when that threshold matters and why it works.

But not all discrete or categorical data is fit for SPC.
For quality grading like “A,” “B,” “C,” you’ll need other tools, like Fleiss’ Kappa, to check measurement reliability—not to control the process.
Defect counting can be used for SPC, and there are also Kappa-style method to check measurement reliability.

#ContinuousImprovement #SPC #QualityTools
https://youtu.be/p9zK4PjsTv8

00:00 Discrete data and SPC
01:26 When discrete behaves like continuous
04:00 What you lose with too few categories
06:00 Quality grades and why they don't chart
08:19 Kappa, Gauge R&R and measurement systems
10:00 Defect rates and real-world applications
13:20 The myth of lower control limits on counts
16:01 When rounding still works

Interested in working with me to boost Operational Excellence in your organisation, or perhaps a more personal coaching on your CI career? Check out my website for details or go to my contacts page directly (www.tommentink.com/contact)

This is Paul Allen's video that I refer to - go check that one out too: https://youtu.be/qfWbujjiPUA

Видео Can you use SPC with discrete data канала Tom Mentink