My data is attribute but the distribution is normal

mohamad1988 · Jan 20, 2021

Hi everybody,
We want to use six sigma project, and our data is binary. We are assessing a production line which produce about 20000 products (it is not constant) with about 200 defective products, and we have exact data with the help of the technology.
So, based on my knowledge I used P chart, the problem is that the distribution is not binomial, and the distribution is normal. So Is it possible to use another tools like X-bar or not? Because, it is better for us to have Cpk, too.
Looking forward to your comments.

Miner · Jan 20, 2021

With large sample sizes the binomial distribution and the normal distribution become approximately the same. See Normal approximation to the binomial. In fact, Dr. Donald Wheeler has written about using the IMR chart for binomial type data. With a large sample size such as yours, I recommend using the proportion defective to back into a Cpk, if really necessary, rather than estimate it using the normal calculations.

Log in or Sign up

My data is attribute but the distribution is normal

mohamad1988 New Member

Miner Moderator Staff Member

Log in or Sign up

My data is attribute but the distribution is normal

mohamad1988 New Member

Miner Moderator Staff Member

Useful Searches