What inspired this Wish List Request?
Customer request (Tesla) to understand the impact of measurement uncertainty on Type 1 and Type 2 error. JMP can do this through the MSA platform and data collected from an MSA design, but the idea here is to take that result from the MSA and apply it to a column of future measurements to understand what impact that uncertainty has on measurements in spec becoming out of spec, and out of spec measurements becoming in spec.
Please describe the new or improved feature or solution you would like to see.
A column property would be available to define a sigma of the measurement uncertainty and a sigma multiplier to be used for the confidence intervals. The sigma would be the GRR sigma that the MSA platform provides and the multiplier would be based on the users desired confidence level.
What would the value be to you and/or other users if this new feature is provided?
This would save a lot of time. Currently Tesla is manually shifting their measurements by +/- the GRR sigma, resulting in 3 columns per measurements (raw measurements, meas + 3*GRR, meas - 3*GRR), and then comparing the in/out of spec rates to determine what the Type 1 and Type 2 error would be from the GRR. This is very tedious/manual process, and would be nice if JMP had this functionality built in.
Is this an external customer request?
Yes, this was specifically requested by the Quality group (George Li) at Tesla.
Is there a competitive software package that has this feature or solution that we should be aware of?
Not to my knowledge
Will this impact a renewal, expansion, or adoption at a customer, and does this put an account at risk if we do not address this?
No
Hi Mark, I believe this is already available (mostly) in JMP 18 as an "MSA" column property.
@mia_stephens I discussed that column property with @tonya_mauldin and she said the existing MSA column property does not do what I am describing above. But what I am describing could be added to the existing MSA column property. The idea being if you had that property specified, doing a distribution analysis with that column would give you the Type 1/2 errors.
