Visual Six Sigma: Making Data Analysis Lean

8.4. Measurement System Analysis

As a general principle, measurement systems for key variables should always be evaluated before engaging in data collection efforts. Since measurement uncertainty or error is suspected as being part of the reason previous attempts to find root causes of the polymer problems failed, it is all the more important that Carl's team thoroughly study the measurement systems for all the variables identified by the process map.

Carl's team learns that recent routine MSAs indicate that CI; polymer weight, which forms the basis for the Yield calculation; SA; M%; pH; Viscosity; and Ambient Temp are being measured with very capable instruments and methods. However, the measurement systems for MFI and Xf have not been evaluated in the recent past. Furthermore, given how these measurements are made, the team realizes that they may be prone to problems.

8.4.1. MSA for MFI

MFI is measured using a melt flow meter during an off-line laboratory test. Four instruments are available within the laboratory to perform the test, and there are three different laboratory technicians who do the testing. There is no formal calibration for the instruments. When Carl's team members interview the technicians who perform the test, they get the impression that the technicians do not necessarily use a standardized procedure.

Carl meets with the technicians and their manager to discuss the desired consistency of the measurement process and to enlist their support for an MSA. He reminds ...