Posted: June 2025
Contact: Max Callaghan
In prioritised screening for systematic reviews, the question is when to stop the screening
process in a way that the reviewer can be confident to reach a desirable recall target.
Building on the work of Callaghan and Müller-Hansen (2000), this project aims to reformulate
and improve existing stopping rules in a Bayesian framework. Bayesian updating of a prior
distribution will be used to derive stopping rules that are both statistically robust and efficient
in reducing screening workload in real-world applications.
Useful skills: Stochastic calculus, Monte-Carlo simulations
Useful prior knowledge: Bayesian statistics, Systematic review automation
