ABSTRACT
This chapter presents an overview of measurement system development organized around seven attributes. A statistical model of measurement, presented first, provides a basis for definition of the attributes. Measurement system development proceeds in stages by means of experiments that focus on different measurement system attributes. In addition to approaches usually associated with repeated measures, measurement system experiments often require errors-in-variables and latent variables approaches. The statistical model presented here serves as a basis for the experiments discussed later, and, moreover, serves as a link to related models found in the statistics literature and as a guide to possibilities for overcoming constraints in specific measurement applications. Sensitivity indicates the ability of the measurement system to distinguish units that differ in the property of interest, and thus involves comparison of the deterministic and random parts of the response. Measurement system sensitivity provides a criterion for reconfiguring and adjusting a measurement system.
