MSA (Measurement Systems Analysis) is the study that determines how much of the variation produced by a measurement system comes from the actual product and how much from the measurement system itself.
Difference from calibration
Calibration shows whether the instrument reads correctly against a known reference. MSA measures how repeatable and reproducible the instrument’s results are under real conditions of use, with real operators, on real parts.
Three operators measuring the same part with a caliper that has a valid calibration can get different results. Calibration does not show this; Gage R&R does.
Properties examined
Bias, linearity, stability, repeatability and reproducibility. For pass/fail type measurements, attribute agreement analysis is applied.
Why is it the precondition for SPC?
Capability values calculated without verifying the measurement system may be mistaking measurement noise for process variation. The signals on the control chart may also be coming from the measurement rather than the process.
The results are reflected in the control plan and the PPAP package.
Gage R&R studies and the interpretation of results are practised on real data sets in the MSA training. For the laboratory-side counterpart, see the ISO 17025 training.