https://edexia.com/research Research into AI-powered grading, pairwise comparison methods, and human vs machine reliability in essay assessment. en-AU Tue, 21 Apr 2026 13:14:22 GMT https://edexia.com/research/comparison-is-key https://edexia.com/research/comparison-is-key How pairwise comparison produces more reliable essay grades than direct scoring. Research on Thurstone’s law, Ofqual data, and implications for AI grading systems with limited training data. Sun, 02 Nov 2025 00:00:00 GMT https://edexia.com/research/the-reliability-of-human-judgement https://edexia.com/research/the-reliability-of-human-judgement When trained raters disagree on 35% of essays, which dataset should AI learn from? Exploring inter-rater reliability, cognitive biases, and what ‘accuracy’ means in grading assessment. Tue, 11 Nov 2025 00:00:00 GMT