- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Report Inappropriate Content
OC Curve
Hi
Does anyone know why I get this error message when I am using OC Curve function in JMP 17 and trying to change AQL and RQL values?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Report Inappropriate Content
Re: OC Curve
Looks to me like the x-axis is actually plotting 1-AQL and 1-RQL. So, 1-AQL should be less than 1-RQL.
1-AQL < 1-RQL implies -AQL < -RQL implies AQL > RQL in this platform. So, assuming everything else is correct, your desired AQL, call this AQL*=1- AQL. As an example, if you want your AQL to be 0.1, put 0.9 in the "AQL" box. Similar for RQL. This is all very counterintuitive...not sure if this is a bug or intentional?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Report Inappropriate Content
Re: OC Curve
Thanks a lot for clarifying this.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Report Inappropriate Content
Re: OC Curve
Hi Behrouz and MRB3855. Thank you for the feedback on OC Curves.
I think what's going on is there are two definitions of AQL and RQL. Suppose we will accept a lot if the defect rate is less than 0.1. From your responses, I believe you define AQL = defect rate = 0.1. In JMP 17.0, the OC Curves define AQL = 1 - defect rate, which here is 1 - 0.1 = 0.9. (RQL is defined similarly)
This definition is used by some (such as in the reference below from the STAT Center of Excellence). However, we have received additional feedback that this definition is counterintuitive and the developer of the OC Curves plans to address this in a future release (so that the OC Curves will use the definition AQL = defect rate).
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Report Inappropriate Content
Re: OC Curve
Hi Jacob
Thanks a lot for explanation and letting me know that I am not the only one that got confused by the definition of AQL/RQL :-). I don't really see the idea of plotting the RQL/RQL lines.