Visual Inspection (Attribute Data) Gage R&R

T

TWIBlogger

Re: Visual Inspection Gage R&R

Of course is nice to know the overall status of agreement but.... if one of the testers is doing badly, how to detect it (non continuous R&R)?

This was a question I struggled with as well Darius. But after getting the inspectors to try and figure out what they disagreed about regarding the samples, I am of the opinion that we will have to conduct periodic R&R indefinitely until the task of inspection can be simplified or automated.

This will at least ensure that we are considering accept/reject bias and getting people together in order to clarify and simplify the visual inspection criteria.

"Philosophical" question for the group: is the concept of Judgment a Component or a Result of visual inspection?
 
1

1frenchie

Re: Visual Inspection Gage R&R

Realize i am coming late to this discussion but it is important to point out that visual inspections and the determination of a GR&R should come directly off your PFMEA. If the feature is a key internal feature or COC to the customer then a study should be performed.
If you identify visual inspection on the PFMEA without Key or COC selected then review those controls with customer and determine if they want a GR&R. (This strengthens your TS requirement of meeting customer expectations and will blow the auditors mind by having this detail documented)
 

Jim Wynne

Leader
Admin
Re: Visual Inspection Gage R&R

Realize i am coming late to this discussion but it is important to point out that visual inspections and the determination of a GR&R should come directly off your PFMEA. If the feature is a key internal feature or COC to the customer then a study should be performed.
If you identify visual inspection on the PFMEA without Key or COC selected then review those controls with customer and determine if they want a GR&R. (This strengthens your TS requirement of meeting customer expectations and will blow the auditors mind by having this detail documented)

Maybe I'm just a little slow this morning, but I can't for the life of me figure out what "COC" stands for in this context.
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
Re: Visual Inspection Gage R&R

Any time you inspect anything, you really need to have some idea as to the effectiveness of the inspection system. Some folks - especially tool room people - believe their instinct is equivalent to a gage R&R, but it has never been proven as far as I know. In fact, gage system variation is often so large it masks process variation. Visual inspection - especially with its 9-10 PFMEA detection score - really will get some customer scrutiny. But, your real risk is the cost of a rejection and sorting at the customer location. Would that justify a GR&R? I know my vote.
 

Bev D

Heretical Statistician
Leader
Super Moderator
Re: Visual Inspection Gage R&R

Maybe I'm just a little slow this morning, but I can't for the life of me figure out what "COC" stands for in this context.

My guess would be Critical Output Characteristic.
 
S

Sturmkind

Re: Visual Inspection Gage R&R

The recent 4th edition of the AIAG (Automotive Industry Action Group) Measurement Systems Analysis manual page 137 that "A general rule of thumb is that values of kappa greater than 0.75 indicate good to excellent agreement...; values less than 0.40 indicate poor agreement."

This latest edition does an excellent job of an attribute example using the Crosstabulation method. Very good discussion on the part of all posters! Thanks!
 

Bev D

Heretical Statistician
Leader
Super Moderator
yes but a kappa of .75 is lousy. This is (roughly) equivalent to diagreement rate of 25%!

this can result in really bad false rejections and false acceptance rates...

if the characteristic is critical this rate is completely unacceptable. if the characteristic is trivial then maybe its OK (but of course then one must ask why you are inspecting fo rit in the first place)

We need to inject logic and reason into our test resutls and not just accept some random arbitrary goodness number our statistics.
 
S

Sturmkind

Well put, Bev. Kalpol92 had asked the question if there was a written standard. Right, wrong, or indifferent there is the MSA 4th ed. I agree with you that 'higher is better' when it comes to appraiser convergence in decision making.
 
Top Bottom