Improving existing models for predicting task difficulty

Funder: Pearson Inc.(£46,667); Post-doctoral Fellow: Dr Yasmine El Masri; Collaborators: Professor Jo-Anne Baird, Dr Peter Foltz, Professor John de Jong and Dr Steve Ferrara (Pearson).

Various psychometric and statistical techniques have been developed to determine task difficulty in educational assessments. However, all of these methods have been problematic and failed at accurately predicting the nature and level of challenges students face when solving tasks in educational assessments. The research project includes various types of activities:

  • Reviewing the literature on previous approaches adopted for predicting task difficulty
  • Investigating strengths and weaknesses of previously used methods
  • Taking into consideration the distinction between the concepts of task ‘difficulty’ and task ‘demand’, review the literature on previous approaches used in evaluating the nature and level of task demands
  • Proposing statistical models to predict task difficulty, taking into account the nature and level of task demand

Yasmine El Masri keynote AEA-E Nov 2014

Yasmine writes blogs relating to this research which are published on Pearson’s Research and Innovation Network website.