I’ve spent enough time reading assessment research to know that most institutions still treat unequal outcomes as a student problem. When students don’t perform, we add accommodations, extra time, extra support, and we leave the assessment design exactly where it was. Tai, Ajjawi, Bearman, Boud, Dawson, and Jorre de St Jorre’s (2023) paper on assessment for inclusion explains why that approach has been wrong from the start, and why the AI moment is making the question impossible to dodge.
What Assessment for Inclusion Actually Means
The authors argue for a shift in where we locate the problem. If a learning outcome is “can analyse a primary source,” then the test should measure that, full stop. If the test also requires fast handwriting, fluent native-speaker English, the ability to remain composed under exam stress for two hours, or rapid recall under pressure, the test is silently judging students on traits that have nothing to do with the outcome.
Tai et al. (2023) call this a validity problem. The test is telling us things we did not ask it to tell us. Their position is direct: “choices in assessment design are never neutral, as each may promote or constrain inclusion differently, and affect different people” (p. 493).
Why Accommodations Are Not the Fix
The accommodations model treats the core assessment as untouchable. Students with disabilities, carers, students working in their second language, students with fluctuating health conditions: they each get a workaround, and the underlying task continues to demand traits that may not be part of the outcome at all.
Tai et al. (2023) argue that this puts extra paperwork and emotional cost on students who already carry the most, while protecting an assessment design that was probably never inclusive in the first place. I’ve made a similar point in my earlier coverage of Nieminen and Eaton’s (2024) work on how assessment accommodations get framed as cheating, and the parallel is hard to miss.
The Cheating Argument, Reframed
The strongest section of the paper is on cheating. Tai et al. (2023) argue that fear of cheating drives many of the assessment choices that most exclude diverse students: closed-book exams, single sittings that lock out carers, remote proctoring that uses eye-tracking and assumes neurotypical eye movements.
The defense is always integrity. The cost is always inclusion. They put it directly: “The upholding of integrity is undoubtedly important, but it must be enacted in ways that embrace inclusion. Any proposed integrity or security practice must be judged against its effects on the most diverse students “(p. 490).
Three years later, that argument cuts even harder. Universities are reaching for AI proctoring, in-class handwritten exams, and lockdown browsers as anti-AI defenses. Dawson et al. (2024) made the case in a separate paper I covered earlier that validity, not cheating, should drive assessment design. Tai et al.’s point lands in the same place. If the response to AI is to default to surveillance and timed handwriting, we are recreating the very exclusions the paper warned about.
The Three Designs They Recommend
Tai et al. (2023) point to three approaches built for other reasons that can still serve inclusion if applied with intention: authentic assessment, programmatic assessment, and assessment for distinctiveness. Authentic assessment grounds tasks in real-world professional practice. With programmatic assessment, the focus moves across an entire degree, so not every capability needs to be tested in every unit.
Assessment for distinctiveness, the third approach, asks students to demonstrate what makes their thinking different, which becomes increasingly important as cohorts grow and qualifications start to look identical on paper. Each of these connects with Bearman, Nieminen, and Ajjawi’s (2023) argument I’ve covered before about designing assessment in a digital world: when the medium changes, the design has to change with it, not the student.
Where I Would Extend the Argument
The paper is strong on critique and cautious on operational guidance. The authors acknowledge this themselves, noting that they have not involved voices from outside the academy. I’d add a sharper reading. Inclusion lives at the teaching level too.
A redesigned assessment still gets delivered by a teacher, in a classroom, under time and budget pressure. The authors flag “work as imagined” against “work as done” early in the paper, and that gap is exactly the issue. Without sustained professional development for teachers, even a beautifully designed inclusive assessment can collapse on contact with everyday classroom reality.
The paper is now several years old, and the AI conversation has moved past where the authors were standing. Their argument for redesigning the test before redesigning the student has aged well.
The closed-book exam is making a comeback, and so is invigilation. Single-mode written assessment is being revived alongside them. Each of those choices is being defended in the name of integrity, and each does the exclusionary work the paper warned about.
If we want assessment to serve diverse learners in 2026, we need to keep asking the question Tai and her co-authors put on the table: who does this design protect, and who does it exclude?
References
- Bearman, M., Nieminen, J. H., & Ajjawi, R. (2023). Designing assessment in a digital world: An organising framework. _Assessment & Evaluation in Higher Education_, 48(3), 291-304. https://doi.org/10.1080/02602938.2022.2069674 .
- Dawson, P., Bearman, M., Dollinger, M., & Boud, D. (2024). Validity matters more than cheating. _Assessment & Evaluation in Higher Education_, 49(7), 1005–1016. https://doi.org/10.1080/02602938.2024.2386662
- Nieminen, J. H., & Eaton, S. E. (2024). Are assessment accommodations cheating? A critical policy analysis. _Assessment & Evaluation in Higher Education, 49_(7), 978–993. https://doi.org/10.1080/02602938.2023.2259632
- Tai, J., Ajjawi, R., Bearman, M., Boud, D., Dawson, P., & Jorre de St Jorre, T. (2023). Assessment for inclusion: Rethinking contemporary strategies in assessment design. Higher Education Research & Development, 42(2), 483-497. https://doi.org/10.1080/07294360.2022.2057451