Experts, Analogies & Data: Modelling Uncertainty based on Expert Understanding of Analogous Events

Overview

While the digital age and its access to vast data sets has not removed the need for expert judgment, it does offer an opportunity to improve it. This project will explore and develop methods to produce uncertainty assessments, drawing on expert analogies with historical data and statistical methods. This project will be aligned with an existing research project concerning risk analysis on the future of tourism for persons with disabilities in collaboration with the University of Dalhousie.Project Details

While the digital age and its access to vast data sets has not removed the need for expert judgment, it does offer an opportunity to improve it. The development of strategies requires an understanding of causal relationships, to support choice of decisions and anticipate their consequences. Data analysis alone, regardless of the level of sophistication, can detect association only not causation. For this, we require expert judgement. Additionally, data may not be exactly applicable to the context of a decision, and expert judgement is needed to translate data into a meaningful recommendation. However, expert judgement is fraught with well-documented biases, which we propose can in part be minimised with appropriate use of existing data.

Specifically, we are interested in the elicitation of expert subjective probabilities to measure the likelihood of uncertain outcomes. Studies have demonstrated that experts generally feel uncomfortable quantifying their judgment and that their assessments are too narrowly concentrated. Based on the premise that expert judgement is based on analogies, we propose a better-calibrated assessment can be obtained through eliciting the expert’s assessment with historical events and then using statistical methods to effectively model the uncertainty.

This project will be aligned with an existing research project concerning risk analysis on the future of tourism for persons with disabilities in collaboration with the University of Dalhousie, in Canada. The student will have the opportunity to work with a multi-discipline group of academics and pursue a complementary research agenda, gaining access to various experts to conduct an empirical study.

The aim of this research is to assess whether deriving probabilities from appropriate data as determined by experts can produce better calibrated assessments than direct assessment of the probability. We anticipate the student will achieve this through the following objectives:

1. Develop an interview protocol that adequately captures analogous data for the event being assessed, including exploring how experts think about the process of drawing analogies with events for which we have data
2. Develop a statistical methodology for deriving probability, based on the experts’ assessments and analogous data
3. Develop empirical evidence to assess how calibrated and informative experts are through analogous data as compared with their direct assessment, building on methods to validate direct uncertainty assessments from experts

Eligibility

Experience and interest in quantitative methods are desirable for this project.

For entry onto our postgraduate research programmes, we normally look for a first-class or upper second-class UK Honours degree, or overseas equivalent, in a relevant business or social science related subject. For PhD applications, we also normally expect a Masters degree, or overseas equivalent, although there are often exceptions. When reviewing your academic achievements, we're particularly interested in grades which relate to independent research (for example, a research project or dissertation). A strong score in these elements may allow us to consider entry with a lower degree classification.

Strathclyde Business School is committed to supporting a diverse and inclusive postgraduate research population. We make decisions on entry by assessing the whole person and not relying solely on academic achievements. On that basis, please ensure that your application (via your CV and covering letter) can evidence your resourcefulness, commitment and resilience as demonstrated by broader professional and life experiences. This evidence should be centred on your ability to undertake and complete a PhD and contribute to a positive PhD community.

If English isn't your first language, you'll need an IELTS score of 6.5 or equivalent with no individual element below 5.5.