Abstract
Group work, where students work on projects to overcome challenges together, has numerous advantages, including learning of important transferable skills, better learning experience and increased motivation. However, in many academic systems the advantages of group projects clash with the need to assign individualised marks to students. A number of different schemes have been proposed to individualise group project marks, these include marking of individual reflexive accounts of the group work and peer assessment. Here, we explore a number of these schemes in computational experiments with an artificial student population. Our analysis highlights the advantages and disadvantages of each scheme and particularly reveals the power of a new scheme proposed here that we call pseudoinverse marking.
Abbreviations
SOPP: Self organised peer assessment; RA: Reflexive accounts; MRA: Mark-adjusted reflexive accounts; NPA: Normalised peer assessment; PR: Peer ranking; PiM: Pseudoinverse marking
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes on contributors
Hugh Harvey is a final year student completing his MEng in the Engineering Mathematics department at the University of Bristol. His research area during his master’s year is similarity logic for explainable AI.
James Keen is a final year student completing his MEng in the Engineering Mathematics department at the University of Bristol. His research area during his master’s year is persuasion in multi-agent AI systems.
Chester Robinson has obtained his BEng in Engineering Mathematics from the University of Bristol. Chester is about to start as an Australian Power derivatives analyst in Singapore, after spending seven months in Coal and Freight Market risk.
James Roff is a final year student completing his MEng in the Engineering Mathematics department at the University of Bristol. His research area during his master’s year is the instability of pilot operated pressure relief valves.This paper was written during undertaking a Mathematical and Data Modelling unit in the second year of study. All authors contributed equally to the completion of this project.
Thilo Gross studied Physics at the Universities of Oldenburg and Portsmouth. After completing his PhD in Oldenburg in 2004 and postdoctoral work in Potsdam and Princeton he joined the Max-Planck Institute for Physics of Complex Systems in Dresden as a group leader. In 2012 he joined the University of Bristol’s Engineering Math Department, where he was a Reader until 2018. He has since left Bristol to become a Professor of Computer Science at the University of California at Davis. In his work he uses a combination of nonlinear dynamics, network science and statistical physics to explore complex systems in a wide variety of applications.