Quality of Measurement Information in Decision-Making

Abstract

This paper introduces a general-purpose framework aimed at capturing the elusive concept of quality of measurement information, a critical issue for both researchers and practitioners when dealing with measurement information-enabled decision making. The framework is a blueprint for the definition, assessment, communication, and improvement of measurement information quality, as analyzed through a set of general criteria, classified according to the syntactic, semantic, and pragmatic layers of semiotics, as suggested in the ISO 8000-8:2015 technical standard. The top-down analysis, where each criterion is specified in terms of characteristics and each characteristic in terms of domain-related indicators, is complemented with a bottom-up synthesis and operationalized by means of a flow chart. An application example, about the quality of information provided by the networks of measurement instruments reporting pollutants in the air, is presented to test the usefulness and the limitations of the framework.

Fundamental Concepts in Measurement

“…", a now classic paper by Charles West Churchman, 1959). While information is sometimes sought for purely knowledge-related purposes, in an encompassing perspective measurement is a key component of information-enabled decisionmaking (Petri et al, 2021): in order to make appropriate decisions on an object, a rational procedure matches the desired state with the current state of the object, as it can be known by means of measurement. This suggests that the general purposes of measurements-and accordingly the general criteria for assessing the quality of measurement results-can be classified depending on whether (i) measurement is a process of information acquisition as such, or (ii) a process of information acquisition aimed at enabling some sort of decision-making for some given purpose(s) in a given context, as happens when measurement results are used as inputs to a control system (e.g., in a decision rule of the kind: "if the value of the relevant property is less than X than do Y, else do Z") or to set the parameters of predictive models.…”

A2.3 - Information recycling of NDE data sets

“…This already represents a first critical point, because whether data sets are suitable for reuse depends on several factors. The most obvious characteristic is thereby the correctness and valency of the measurement [4]. Since this is influenced by essentially the correct execution of the inspection, this point is unfortunately difficult to verify afterwards.…”