how do you rule out measurement error vs real problem

A subtle, high-value skill — especially for you as a metrology person. Your method: 1) First suspect the measurement — it's the cheapest thing to check and the most common false alarm. Verify the tool is calibrated and the method is sound. 2) Re-measure to confirm repeatability — is the reading stable or noisy? 3) Cross-check with a second method or reference standard if available. 4) Only then conclude it's a real process/equipment problem. The key line: 'Before I chase a process problem, I prove the measurement is telling the truth — a false signal sends you chasing a problem that isn't there.' This is where your reference-standard calibration work is gold — you literally improved measurement accuracy/repeatability. It shows rare judgment.