Measurement System Analysis (MSA) is the assessment of whether a measurement process produces data suitable for a defined operational or analytical decision. It examines variation and error introduced by the equipment, the measurement procedure and the conditions under which readings are taken.
In a DMAIC project, MSA belongs in the Measure phase, before the team relies on collected data to assess process performance. For example, a team investigating shaft diameter might have 3 operators measure the same 10 parts twice, producing 60 readings for a crossed Gage R&R study. The parts should represent the process variation relevant to the investigation. Randomising the measurement order and concealing earlier results helps limit recognition effects.
The analysis informs a practical decision: can the measurement system distinguish differences that matter? If operator technique contributes excessive variation, the team may need to revise the measurement procedure and repeat the study. A calibration check against a reference value addresses another question, whether readings show systematic bias.
Gage R&R is one tool within MSA, focused on repeatability and reproducibility. Other MSA studies examine bias or stability over time. For categorical inspection, attribute agreement analysis evaluates consistency in decisions such as accepting or rejecting a weld. MSA supports process capability analysis and hypothesis testing by establishing whether measurement error threatens the conclusions drawn from process data.
