Authentic Assessment: Evaluating Learning Through Projects, Portfolios and Real Tasks

Start with the capability

An authentic assessment asks learners to apply knowledge in a task that resembles a meaningful disciplinary, professional or civic activity. The important feature is the intellectual work: interpreting evidence, making a justified decision, producing a defensible design or communicating for an audience. A realistic-looking assignment can still test little more than presentation skills. Begin by completing the sentence: after this course, a learner should be able to do what, under which conditions, and to what standard?

Carnegie Mellon University's assessment guidance connects assessment choice with intended learning objectives.^[1] For your course, identify two or three priority capabilities and specify the evidence that would demonstrate each. If the outcome is evaluating policy options, a beautifully formatted description of one policy is insufficient. The task must require comparison, criteria and a reasoned recommendation.

Define the task and its boundaries

Write a short assessment brief containing the problem, audience, deliverable, available information and constraints. Tell learners which assumptions they may make and which they must justify. A constrained simulation can be more educationally useful than an undefined instruction to solve a real-world problem. Set a scope that fits the course timetable and the learners' access to resources.

Decide whether external engagement is necessary. A live client introduces scheduling, confidentiality and expectation risks. A well-designed case can assess the same reasoning without those dependencies. If external participants are involved, agree permission, supervision and the permitted use of material before students collect information. Do not make a student's grade depend on whether a partner implements their recommendation.

Combine product and process evidence

A final report alone may conceal how a student reached a conclusion. Request selected process evidence: a problem definition, annotated source selection, an early design decision, a response to feedback and a short individual reflection. These items should illuminate the assessed capability, rather than become a large administrative diary.

For group work, separate shared outcomes from individual evidence. A common project score can be combined with an individual explanation or oral discussion, subject to the institution's assessment rules. Announce the arrangement at the beginning. Peer contribution reports can inform a review, but disagreements should be investigated rather than mechanically converted into penalties.

Set criteria students can use

Describe quality through a small rubric covering the capabilities that matter. For a policy task, criteria might include evidence selection, comparison of alternatives, feasibility and clarity for the intended audience. Weightings should reflect educational priorities. Do not allow visual polish to dominate unless design communication is itself a learning outcome.

Ohio University's authentic assessment guidance recommends sharing performance criteria and supporting students through preparatory tasks.^[2] In practice, let learners examine a short exemplar, discuss why a claim is weak and rehearse one part of the task. Give feedback while improvement remains possible. Explain the boundaries of collaboration and any permitted AI assistance, including what must be disclosed.

Illustrative example: a campus water proposal

Imagine a postgraduate sustainability class asked to recommend one water-saving intervention for a fictional campus. Students receive the same consumption dataset, indicative costs and operational constraints. Teams submit a decision memo comparing three options, a calculation sheet and a plan for checking results. Each student then explains a different assumption in a brief individual conversation.

The assessor can distinguish data interpretation, financial reasoning and communication. Students with no access to a real campus facilities department remain able to participate. An accessible spreadsheet and a text description of all charts are supplied. The task does not claim that any option will achieve a particular saving: uncertainty is part of the decision students must explain.

Keep delivery feasible and fair

Estimate marking time using an actual sample, including moderation and individual discussions. Where the workload is excessive, reduce the number of deliverables rather than remove the evidence needed for a fair judgement. Provide accessible formats and equivalent ways to demonstrate the same outcome where appropriate. Adjustments should preserve the assessed standard without requiring identical presentation methods.

Check that expensive software, personal contacts or unpaid travel do not become hidden requirements. Publish deadlines for feedback, extensions and questions. External feedback can enrich a task, but academic judgement should remain with the authorised assessors.

Review the assessment after use

Compare the evidence produced with the intended outcomes. Ask which instructions students misunderstood, where assessors disagreed and whether some learners faced avoidable barriers. Review a range of achievement levels, not only excellent work. Retain anonymised examples where permissions allow and revise the brief before the next delivery.

An authentic task is not automatically more valid than an examination. Different outcomes may require different methods. The test of quality is whether the assessment produces credible, interpretable evidence of learning while remaining fair, accessible and workable.