I find myself in the possition of having to explain how having a 67% of study error in a measurement system (AIAG MSA method - variable data) errodes the confidence in DOE results relying on this system. Can anyone suggest how to prove the negative effects of using this poor system for DOE analysis? Those I need to convince don't seen to think there is too much wrong with this! They feel that the DOE is used to show relative differences between parameter settings, thus the %error is not that important! I've told them that with this much error, you have little confidence about the relative differences you've calculated. Please help me prove it!