Resource icon
- Thread starter Bev D
- Start date
Bill Levinson
Industrial Statistician and Trainer
I looked briefly at the download, and it raises a very interesting point; inspectors are fixed rather than random factors. This means that, if one were to bring in other inspectors, the reproducibility variation component could indeed be different. I wonder if the key deliverable from the R&R study should be, rather than the specific reproducibility component (although this is mandatory and must be reported) is the fact that there is appraiser variation which means we should look at how the measurement is done to remove it regardless of magnitude.
A simple example is a micrometer whose instructions say "finger tight," the perception of which will vary from one inspector to another, and a micrometer with a slip clutch to make the amount of pressure constant.
The Youden Plot looks like it is useful for interlaboratory comparisons (ASTM for example) and I can see how it can apply to MSA as well. 1.3.3.31. Youden Plot
A simple example is a micrometer whose instructions say "finger tight," the perception of which will vary from one inspector to another, and a micrometer with a slip clutch to make the amount of pressure constant.
The Youden Plot looks like it is useful for interlaboratory comparisons (ASTM for example) and I can see how it can apply to MSA as well. 1.3.3.31. Youden Plot
Last edited:
Jack Youden developed the Youden plot when he was the head of NBS and yes it is applicable to any measurement system comparison that utilizes 2 repeated values. There is a great history of his career at the AMSTAT site.
It is important to note that that the ‘popular’ Gauge R&R method is only ‘mandatory’ in some industries. Primarily the automotive industry as that is where it originated. Many industries do not require this approach; those that do often have required without much thought - they just copied the automotive industry. As my Mother used to say: “if all of your friends were going to jump off a bridge would you?” I choose to not jump.
It is important to note that that the ‘popular’ Gauge R&R method is only ‘mandatory’ in some industries. Primarily the automotive industry as that is where it originated. Many industries do not require this approach; those that do often have required without much thought - they just copied the automotive industry. As my Mother used to say: “if all of your friends were going to jump off a bridge would you?” I choose to not jump.
unfortunately too many people just blindly copy the automotive industry. It creates a lot of unnecessary heartache.
Bill Levinson
Industrial Statistician and Trainer
I trust the AIAG's manuals because they make a lot of sense, and are entirely consistent with generally accepted methods. This is not to say there is never room for improvement. As but one example, Donald Wheeler pointed out quite accurately that risk priority numbers are products of ordinal ratings and might therefore not reflect the actual risk level. The AIAG-VDA FMEA manual now addresses this. On the other hand, I am now very interested in the Youden plot and how it might be used to supplement traditional MSA.
Bill Levinson
Industrial Statistician and Trainer
Wheeler makes sense, and I never paid much attention to "percentage of total variance." All most people care about is the precision/tolerance ratio which is based on the gage standard deviation as calculated from the appraiser and equipment variation only.
I don't pay much attention to part variation either because, with only 10 parts, we don't have enough data for a good estimate; we will rely on the process capability study instead (with 30 or more measurements for a good one). And of course part variation, which we expect, does not have anything to do with the gage precision.
Bill Levinson
Industrial Statistician and Trainer
Your article raises some good points; I am going through it in a little more detail.
On page 26, "A subgroup size of 10 can be biased by one outlier" ... the outlier should be evident in the MSA assessment such as the range chart and also in anything supporting, like a normal probability plot. At this point, we know something is wrong with the study (or that there is a risk of non-random assignable cause measurement errors).
Also, 10 is a clearly inadequate sample for estimating the part variation; I see part variation as more of an academic exercise where we compare what we get from the MSA to what we get from a process capability study that uses 30 or more parts. I would not rely at all on the MSA's part variation to reflect the actual process performance. Also, I am not sure why part variation is even relevant (unless one wants to compare part variation to gage variation) because the key metric is the ratio of the gage standard deviation to the specification width or tolerance, and not to the process standard deviation.
I would also, as you point out on page 28, be hesitant to add %EV and %AV as calculated for exactly the reason you describe. The deliverable is the total gage standard deviation for which a precision/tolerance ratio can be calculated. I know that various software reports these things but my webinar on MSA does not bother with them at all because the deliverable is the P/T ratio.
I am going to continue looking through your presentation, and I will also see what I can find on the Youden plot, which looks very interesting.
That's like saying, " I trust the AIAG manuals because they contain the AIAG methods." They got to be "generally accepted methods" because of AIAG requirements.
Similar threads
