Measure Phase: Data Collection, MSA, and Process Capability
What the exam tests
Measure covers data collection, measurement system analysis, and baseline capability. Expect operational definitions, sampling, Gage R&R interpretation, control versus capability, and Cp/Cpk versus Pp/Ppk.
Core rules
- Define each metric operationally before collecting data; otherwise two people measure different things.
- Run MSA before trusting data. For Gage R&R, percent GRR under 10% is generally acceptable, 10-30% is conditional on the application and cost, and above 30% is unacceptable.
- MSA examines repeatability, reproducibility, bias, linearity, and stability.
- Cp = (USL - LSL) / 6 sigma measures spread against tolerance; Cpk = min(USL - mean, mean - LSL) / 3 sigma adds centering.
- Pp/Ppk use overall long-term variation; Cp/Cpk use within-subgroup variation. Report which one, and only after the process is stable.
Worked example
A process has mean 50.2, standard deviation 0.8, LSL 47, and USL 53. Cp = (53 - 47) / (6 x 0.8) = 6 / 4.8 = 1.25. Cpk = min(53 - 50.2, 50.2 - 47) / (3 x 0.8) = min(2.8, 3.2) / 2.4 = 1.17. The reasoning: the width is adequate because Cp is above 1, but off-centering reduces Cpk; shifting the mean to 50 would give Cpk = 1.25.
Traps
Calculating capability before confirming stability and MSA; using Cpk on non-normal data without a transformation or non-normal method; dividing by 6 sigma when the formula calls for 3 sigma; and confusing control limits with specification limits.