Uncertainty assessment of performance indicators

Abstract

Purpose: This paper defines a model to evaluate the uncertainty in performance indicators (PIs) based on Uncertainty Components (UCs). Methodology: The proposed work consists, in a first stage, of an assessment of the level of influence that each UC has in a given PI. Based on the questionnaire responses a matrix of UCs vs PIs is presented to show the relevance of the contribution of each UC to the uncertainty associated with a PI. The second stage of the methodology consists on the development of a model to infer the uncertainty level on a PI based on the uncertainty level of the identified UCs. Findings: A questionnaire referring to the assessment of PIs was applied, and the results provide evidence that UCs influence the PI. A model was developed based on logical relations between the UCs and the overall PI uncertainty, and the number of empirical analyses contribute to validate it. Originality/value: This paper presents a model to infer the uncertainty level of a PI based on UCs. The model can also be applied to propagate uncertainty among multiple related PIs. UCs definitions can guide the development of actions to reduce uncertainty in PIs, thus reducing the risk in the decision making process.