Gauge R&R for destructive test with no true parts - best JMP approach?

[#1]

Hi everyone,

I'm working on a measurement variability study on a product sample. The test is destructive (each test consumes the full sample), so I do not have "true parts" that can be remeasured. My dataset consists of:

1. Multiple technicians

2. Two instruments

3.  2-3 repeated measurements per technician-instrument combination

4. All sample coming from the same material lot

My primary goal is to understand technician-to-technician and instrument-to-instrument variability, I guess I can't do a part-to-part variation. In JMP Gauge R&R platform, I am unsure where all these parameters should go (how do I correctly define "Part/Sample ID" in this case, should I group the analysis by instruments or technician, etc.). From the R&R result, which values should I care about?

Date	Instrument	Technician	Measurements	Repeats
1/2/2026 11:47	E-1	CG	0.282	1
1/2/2026 11:46	E-1	CG	0.28	2
1/2/2026 8:42	E-1	CG	0.28	3
1/2/2026 11:48	E-2	CG	0.294	1
1/2/2026 8:44	E-2	CG	0.296	2
1/2/2026 8:41	E-2	CG	0.292	3
1/3/2026 9:51	E-1	DM	0.286	1
1/3/2026 8:31	E-1	DM	0.282	2
1/3/2026 8:42	E-2	DM	0.298	1
1/3/2026 7:40	E-2	DM	0.294	2
1/2/2026 17:02	E-1	NZ	0.286	1
1/2/2026 17:19	E-1	NZ	0.288	2
1/2/2026 17:21	E-2	NZ	0.3	1
1/2/2026 17:17	E-2	NZ	0.282	2

Appreciate any help on this!  

[#2]

Can you make an assumption that some of the parts are identical enough to each other, to consider them the same part?

Few links to check out (not sure if they apply to your situation): 

• Advanced Measurement System Analysis for a Destructive Gauge Using a Split-Plot Design 
• Using MSA for Situations Requiring Destructive Testing