A rapid review method for extremely large corpora ofliterature: Applications to the domains of modelling, simulation, and management

摘要

While literature reviews with a large-scale scope are nowadays becoming a staple element of modern research practice, there are many challenges in taking on such an endeavour, yet little evidence of previous studies addressing these challenges exists. This paper introduces a practical and efficient review framework for extremely large corpora of literature, refined by five parallel implementations within a multi-disciplinary project aiming to map out the research and practice landscape of modelling, simulation, and management methods, spanning a variety of sectors of application where such methods have made a significant impact. Centred on searching and screening techniques along with the use of some emerging IT-assisted analytic and visualisation tools, the proposed framework consists of four key methodological elements to deal with the scale of the reviews, namely: (a) an incremental and iterative review structure, (b) a 3-stage screening phase including filtering, sampling and sifting, (c) use of visualisation tools, and (d) reference chasing (both forward and backward). Five parallel implementations of systematically conducted literature search and screening yielded a total initial search result of 146 087 papers, ultimately narrowed down to a final set of 1383 papers which was manageable within the limited time and other constraints of this research work.