A study of human measurement error in a controlled experiment for micrometer measurements
This is an exploratory thesis. The objective of this thesis is to isolate human measurement error in a physical measuring environment. A secondary objective is to evaluate this measurement error, if successfully isolated, with respect to its effect of biasing statistical quality control tests that describe a manufacturing process.
The author of this report designed a specialized measurement jig that was used to isolate human measurement error. Specifically, the tests involved seven human inspectors making micrometer measurements of the diameters of cylindrical brass pieces. Several physical factors were rigidly held in control or eliminated by the design of the jig and the experimental process. This was necessary to obtain as accurate an estimate as possible of the human error.
The human measurement error in this experiment was successfully isolated. Further analysis of this measurement error led to the hypothesis that it probably can bias statistical quality control tests that describe a manufacturing process. This bias is reasoned to have a greater effect on statistics that describe processes with very close physical tolerances.