Assessment mapping and audits

Source: Rundle, K, Curtis, G and Clare, J. (2020) ‘Why students choose not to cheat’. Chapter in Tracey Bretag (Ed) A Research Agenda for Academic Integrity.

The diagram above demonstrates the 'Swiss cheese' approach designing for academic integrity at the programme level. The basic idea is to have a diversity of assessments to ensure that students will not be able to graduate from a programme with one or two types of assessment which many be vulnerable to AI.

This needs to be carefully planned and implemented with programme teams. If module leads design their own assessments in isolation, the programme lacks a coherent structure and students can be doing a range of different types of assessments without building skills and practice needed for subsequent assessments - too much diversity is as problematic as too little.

Reducing the volume of assessment at the programme level is also key to addressing students’ motivations to use AI inappropriately as well as allowing more space within a curriculum to develop AI and academic literacies. Guidance from the QAA provides more details.

See Assessment for Learning at King's for more information about programme mapping or contact kings-academy@kcl.ac.uk to talk more about this.

Synoptic assessment

These are assessments across a number of modules or for a whole year which assess a number of programme level learning outcomes. See more about this approach as demonstrated at Brunel and Bradford universities.

Nested tasks

These are a series of tasks sequenced across a module which culminate in one summative grade for a larger piece of work. They are sometimes known as staged or sequential tasks or ‘continuous assessment’. The most obvious benefit is that we get to know students’ work as it progresses, and they can use feedback iteratively to develop their work. Generative AI tools can be incorporated into the process this way as students can use some of the formative tasks suggested in the 'using AI in assessment design' guidance.

Processfolios

A processfolio is similar to a portfolio in that it is a collection of artefacts. But rather than a collection of pieces of work, it shows a student's journey towards producing ONE assignment. It is an accompaniment to a product or piece of work without having to use nested tasks in a more formalised way.

Artefacts can include anything from the course that helped them to produce the assessment (drafts, outlines, course readings, their summaries or research, feedback from you or their peers, even some possible ChatGPT written work). They should write about each artefact, justifying why they have chosen it and how it helped them to produce their assignment. They can even choose to display and talk about anything that did not help them and why.

This helps you to make your assessment less vulnerable to AI in that the student has to select and justify how they have undertaken (and understood) the process of producing a piece of coursework over time.

The folio does not need to be assessed in itself but could be a mandatory accompaniment to a summative assignment. Or you could give it a smaller percentage of the final grade than the main summative work. It has the benefit of helping students raise metacognitive awareness of their learning journey, the writing process and developing their evaluative judgment of standards expected of them. As with nested tasks, the use of generative AI can be part of the work provided students include a reflective commentary on its use.

Multimedia/hybrid assignments

Generative AI tools can be very good at producing written texts in different styles such as blogs. And 'real life' texts such as executive summaries, scripts, policy briefs are not AI-proof.

This does not mean they are not good forms of assessment, however, because when setting these types of output for external audiences, we often work with students to deconstruct and create writing styles. They are also inherently more motivating, and students are less likely to want to outsource these to AI as they can see the direct employability value of them.

Other media assignments include:

• Vlogs
• Podcasts
• Graphic abstracts/posters
• Short Films

Assessment for Learning at King's has a range of examples of these more innovative types of assessment.

Oral assessments

Oral assessments are often used retrospectively when students are suspected of plagiarism. But bringing oral assessments, beyond the usual 10 min presentation, into an assessment, can be very useful for developing students' skills*.

Generative AI tools can also produce slides and presentation scripts in any style so it is important to consider assessments which again, test higher order skills such as creativity and evaluation/analysis, and preferably use their own generated data.

Examples of oral assessments include:

• Debates
• Pecha kucha or pitch
• Research or business proposals
• Role plays/simulations/OCSEs
• Interviews (one to one or peer to peer to peer) about learning
• Vivas or oral discussion of their work with a peer or assessor.

For examples and further details of how you might use these, see Assessment for Learning at King's or contact jayne.pearson@kcl.ac.uk.

*Many students find oral assessments anxiety-inducing so it would not be a good idea to make the majority of your assessments oral. Read more about how to scaffold and support students to engage with oral assessments.

Concept maps or visual representations of data

A concept map is a graphic tool which helps students to visually represent a concept. It can help distinguish whether students have rote-learned, partially understood or really internalised a concept/idea/topic/theory. If students draw their concept map and then explain it to others, it is easier to assess the normally internalised processes of learning because learning itself is externally represented.

You can read more about concept maps at Assessment for Learning at King's. Most generative AI tools are not currently very good at reading or producing complex images using data.

A graphical abstract is a single concise, pictorial and visual summary of the main findings of a published research article. Graphical abstracts are a relatively well-established form used by large publishers of scientific journals to increase the access to and reach of scientific findings.

See example. 

Assessment for Learning at King's Collection

King's Academy's resource 'Assessment for Learning at King's' has a collection of discipline-specific assessments collected from our King's community, all of which deviate from the more traditional forms of assessment which may be more vulnerable to AI. Names of academic staff are provided for you to contact if you wish.