Assessment Compendium

Academic misconduct and misuse of generative AI may be avoided.

Speed of handwriting and legibility can affect performance; some groups, such as students whose first language is not English, or students who suffer anxiety around exams, will be disadvantaged.

The types of notes or texts, or the use students are permitted to make of online resources will need to be planned and clearly communicated. This will also be challenging to monitor.

The same inclusivity considerations largely apply to both traditional and open-book exams (speed of handwriting or typing, English as an additional language, anxiety), but open-book may suit a slightly broader range of needs.

Students will have the opportunity to seek help or to collude on preparation.

Students with commitments such as work or dependents may be unfairly disadvantaged. 

Essay writing is traditionally susceptible to contract cheating, and now also to the misuse of generative AI; encouraging alternative ways of supporting students, such as step-by-step writing of sections of an essay to multiple deadlines before summative submission, or peer support through ‘critical friends’ can help to minimise these risks.

Understanding UK academic conventions can be a challenge to some students, and marking criteria or rubrics will need to be clearly explained to them; students with dyslexia, or whose first language is not English may need to be signposted to additional support, or may have reasonable adjustments granted for academic conditions affecting written expression.

Report writing may be less susceptible to academic misconduct than essays, particularly when original research or a project has been undertaken, or when students work in groups on a report.

If group work is involved in research, how students are allocated to groups and the management of the groups will need to be carefully planned, taking individual needs into consideration.

Give expectations of the structure and language of a report, with exemplars if possible, as some students may not know the components of this form of writing.

There is little risk of academic misconduct in a practical or performance assessment.

Performance on the day can be affected by nerves and anxiety, or circumstantial pressures, and rescheduling may not be straightforward. Enable practice runs before any high-stakes assessment.

If blogs/vlogs are posted on an open forum, there is potential for peer plagiarism.

Since blogs are shorter pieces of writing than essays, and language style can be informal, blogs and vlogs can be useful, low-stakes assessments that get students used to the writing or reflection process, and can help to create communities of learning. As with reports, clarify writing structure and style for those not familiar with this type of writing.

Multiple choice tests may be highly susceptible to academic misconduct. While a time constraint may help to limit the risks, careful consideration should be given to whether or not speed is a component of the learning outcomes.

Getting students involved in the co-creation of MCQs can be an excellent way to identify gaps in knowledge and understanding, aligning with the principle of assessment as learning.

Short timescales for completion will need to be considered carefully as students with dyslexia or English as an additional language will be disadvantaged by a speed of completion requirement.

Academic misconduct and misuse of generative AI may be avoided.

The same inclusivity considerations largely apply to all exam types (speed of handwriting, second language, anxiety), but short answer exams put the focus more on memory and recall than on communication skills.

Computer based tests may be highly susceptible to academic misconduct, but a time constraint and carefully constructed critical/reflective questions may help to limit the risks.

Digital literacy can vary.

Those with visual impairment or dyslexia, or have mobility issues in relation to mouse use may struggle in speed situations.

Ensure you design accessible features such as font sizes, and colour contrast, to enable screen readers for visually impaired individuals, and keyboard shortcuts for those with mobility challenges to access the material. Follow principles of UDL.

Portfolio work can be highly individualised, making it less susceptible to academic misconduct.

Submitting work across a range of modes may allow students to play to their strengths, although this may be overwhelming for some students.

There is a case for increased use of oral assessment in conjunction with long-form writing in light of present-day generative AI contexts.

Understanding and accessing questions may be problematic. Consider issues around hearing and/or speech impairment, and anxiety or mental health conditions. Provide questions in advance, and/or allow access to materials or notes, if possible.

Presentations can be conducted in a live or pre-recorded setting. Can be susceptible to genAI use, though if monitored and used appropriately risks can be minimised. The inclusion of the ability to respond to spontaneous questions in the assessment criteria can minimise these risks further.

Usually relies on verbal articulation of ideas in real time disadvantaging those with English as an additional language or speech impairment. Can cause anxiety. Clarify the specific presentation skills being assessed (action and expression). Provide scaffolded formative practice: in pairs/small groups/whole class, with peer and then tutor feedback using assessment criteria. Give options for live or recorded and face to face or online (action and expression).

There are limited risks associated with poster assessments.

Relying solely on a visual form can challenge those with visual impairment

Staff may also wish to design a non-visual text version of this assignment with alternative marking criteria, as an alternative assessment.

Project work often allows for creativity and originality, which may be a check on the use of generative AI.

Projects often involve group work, so how students are allocated to groups and the management of the groups will need to be carefully planned, taking individual needs into consideration.

Similar to essay writing, dissertations can be susceptible to misconduct and now also to the misuse of generative AI; encouraging alternative ways of supporting students, through effective feedback and guidance on thesis writing is essential here.

Students working towards large scale projects may feel isolated or lonely. Feedback dialogues are important here where face to face guidance and encouragement is essential. Thoughtful consideration should be given to factors such as accessibility, guidance and flexibility.

These assessments may be relatively resistant to academic misconduct.

Can cause anxiety. Nerves may come into play and affect ‘on the day’ performance. Ensure opportunities for practice and feedback are provided before the summative assessment.

These assessments may be relatively resistant to academic misconduct.

Nerves may come into play and affect ‘on the day’ performance. Ensure opportunities for practice and feedback are provided before the summative assessment.

Reflective journals have been suggested as a way to avoid misuse of AI since they depend on personal experience; however, reflections may not be AI proof. See here for more on this.

Consider allowing both written and oral (video) submission of reflective journals since either format should meet learning outcomes.

The creativity and problem-solving skills required for critical incident analysis are currently less susceptible to misuse of generative AI.

Creates anxiety

Feedback dialogue and opportunity for practice before the summative assessment is important here.

Seminars can be difficult to mark ‘live’ and may be better submitted in a pre-recorded format. Students may rely on genAI in a remote setting to create pre-prepared scripts on a given topic ahead of the assessment.

Students with anxiety, autism spectrum disorder, ADHD, hearing or speech impairments or with English as an additional language may encounter difficulties with this type of oral assessment. Distribution of accessible topics to a large group of students may be challenging.

AI tools can often produce clear and effective summaries of texts.

This may be a very unfamiliar type of assessment to the majority of students, and expectations and process will need to be clearly explained.

This type of assessment is relatively resistant to academic misconduct since it relies heavily on reflection and decision making.

These assessments require working under time pressure with a lot of information to manage and respond to, and may cause high levels of anxiety in some students.

Producing artefacts can now be susceptible to the misuse of generative AI; encouraging students to demonstrate their ability to use AI tools effectively and critically, and to generate creative solutions can be supported by staff.

Digital literacy can vary so this may cause digital accessibility issues. Though assistive technologies could also promote inclusive and accessible applications for all learners.

Work based learning assesses both theoretical knowledge as well as the application of skills in practical scenarios which limits opportunity for academic misconduct.

Creates anxiety;

Assessments should be designed with flexibility to accommodate diverse needs, ensuring that all learners can demonstrate their skills and knowledge effectively.