Judging the state of a bridge based on SHM observations is an inference process, which should be rationally carried out using a logical approach. However, we often observe that real-life decision makers depart from this ideal model of rationality, judge and decide using common sense, and privilege fast and frugal heuristics to rational analytic thinking. For instance, confusion between condition state and safety of a bridge is one of the most frequently observed examples in bridge management. The aim of this paper is to describe mathematically this observed biased judgement, a condition that is broadly described by Kahneman and Tversky’s representativeness heuristic. Particularly, we examine how this heuristic affects the interpretation of data, providing a deeper understanding of the differences between a method affected by cognitive biases and the classical rational approach. Based on the literature review, we identify three different models reproducing an individual behaviour distorted by representativeness. We apply these models to the case of a transportation manager who wrongly judges a particular bridge unsafe simply because deteriorated, regardless its actual residual load-carrying capacity. We demonstrate that application of any of the three heuristic judgment models correctly predicts that the manager will mistakenly judge the bridge as unsafe based on the observed condition state. While we are not suggesting in any way that representativeness should be used instead of rational logic, understanding how real-life managers actually behave is of paramount importance when setting a general policy for bridge maintenance.
- Bayesian inference
- bridge management