How to Design Make-Up Exams in the Age of AI
How college faculty can design fair, equivalent make-up exams in the age of AI using oral defenses, scenario tasks, and asynchronous review.
Every semester, the same logistical problem lands in a professor's inbox on exam morning. One student wakes up with the flu, another is traveling with a varsity team, and a third could not get a seat in an overbooked campus testing center. You still need to assess those students fairly, yet the exam you gave to the rest of the class is compromised the moment the first group walks out of the room.
When faculty ask how to design make-up exams in the age of AI, they are trying to solve two problems at once. They need an equivalent assessment that measures the same learning outcomes as the missed exam, and they need a format that a student cannot complete in thirty seconds by pasting a prompt into a chatbot.
The most reliable approach shifts the make-up from a static written test to a short interactive check of reasoning: a structured oral make-up exam, a fresh data walk-through with spoken explanation, or a recorded synthesis session for a missed lab or seminar.
Why Traditional Make-Up Exam Habits Break Down
Most faculty improvise make-up exams using one of three workarounds, and generative AI has made each of them fragile.
The first workaround is reusing the original test in an office or testing center a few days later. Word always traveled fast through group chats and shared study folders. Once students can transcribe remembered questions and ask a language model to generate targeted study sheets for peers, delayed sittings with identical items lose validity quickly.
The second workaround is writing a fresh written version of the exam for two or three absent students. Crafting equivalent multiple-choice distractors or new quantitative problems that match the difficulty of the main exam takes hours. Under mid-semester time pressure, the alternate version often turns out noticeably harder or easier than the original, inviting grade equity complaints.
The third workaround is assigning a take-home essay or an unproctored online quiz as a substitute. Because language models produce clean prose and step-by-step derivations on demand, a take-home replacement leaves the instructor guessing whether the student learned the material. Relying on automated AI text detectors introduces false positives without proving understanding, while requiring webcam proctoring for a sick student adds technical friction and privacy concerns.
Establishing Equivalence Around Learning Outcomes
A common mistake when designing make-up exams is trying to mirror the exact format and question count of the missed test. A fifty-minute written exam with forty multiple-choice items and two short essays cannot be squeezed item for item into a twelve-minute make-up conversation. Chasing surface parity leads to rushed questions and exhausted students.
Equivalence rests on construct validity: the make-up must test the same core learning objectives at the same cognitive level. As UC Berkeley's Center for Teaching and Learning notes in its guidance on alternatives to traditional testing, alternative formats often measure higher levels of Bloom's taxonomy, including analysis and evaluation, more directly than traditional paper tests.
Start with the syllabus blueprint for the missed exam and pull out the three or four non-negotiable competencies a passing student had to demonstrate. In an introductory biology midterm, those competencies might be tracing cellular respiration under oxygen deprivation, interpreting an enzyme kinetics plot, and diagnosing a flaw in an experimental control. Once those outcomes are clear, you can test them directly without authoring forty new items.
Before designing custom make-ups for every low-stakes quiz, eliminate unnecessary make-ups through syllabus policy. Dropping the lowest weekly quiz score or shifting the weight of a missed midterm onto a cumulative final section keeps routine absences out of your inbox. Save dedicated make-up assessments for high-stakes exams, accredited clinical requirements, and documented accommodations where a student needs an independent score.
Three Practical Make-Up Formats That Hold Up With AI
When a student needs an individual make-up assessment, three formats hold up reliably even when generative AI tools are available during study.
1. Structured Oral Make-Up Exams
An oral make-up exam, often run as a twelve-minute mini-viva, asks the student to talk through core problems and answer follow-up questions. Cornell University's Center for Teaching Innovation highlights oral assessments as particularly strong at measuring complex thinking skills such as analysis and evidence-based reasoning, while encouraging deeper preparation because students must state ideas in their own words.
Consider how this works in organic chemistry. Suppose a student misses an exam covering reaction mechanisms. The instructor assigns two synthesis targets and asks for a recorded explanation of the reaction steps, followed by targeted probing questions: why a particular solvent favors SN1 over SN2 here, or how the major product shifts if the temperature rises. A student who memorized chatbot output without grasping carbocation stability runs out of answers on the second follow-up. A student who understands the mechanism can reason through the shift on the spot, even if they hesitate over a minor reagent name.
For practical advice on structuring spoken evaluations across larger courses, see our guide on how to run oral exams in college courses.
2. Fresh Case or Data Walk-Throughs With Spoken Defense
In quantitative, policy, and applied courses, pairing a small novel prompt with a spoken defense gives you both written work and direct verification of reasoning. The student receives a compact dataset or scenario they have not seen before, completes a brief analysis inside a set window, and explains their choices aloud.
In an undergraduate macroeconomics or public finance course, an instructor can give an absent student a one-page municipal budget table showing a sudden revenue shortfall and ask them to recommend two spending adjustments. During the brief follow-up defense, the student must justify the trade-offs, identify the riskiest assumptions, and respond when the instructor drops projected property tax collections by five percent. Even if the student consulted AI while reviewing notes beforehand, they still have to defend the numbers and logic under direct questioning.
3. Recorded Session Synthesis for Missed Seminars and Labs
Not every missed assessment is a formal midterm. In discussion seminars, clinical debriefs, or laboratory briefings, attendance and active participation carry real course weight. When an excused student misses a key session, traditional make-up assignments often become busywork summaries copied from a classmate's notes.
A stronger replacement asks the student to review the class recording or lab protocol and complete a short structured conversation grounded in that material. The prompt can ask the student to explain the central disagreement raised during the session, connect one claim from the reading to the lab data, and answer two follow-up questions anchored in the session transcript. That turns a passive catch-up task into active engagement with the class event.
Keeping Make-Up Grading Fair, Accessible, and Manageable
Faculty sometimes avoid oral or scenario-based make-ups because they worry about grading subjectivity and calendar scheduling. Both issues have straightforward fixes.
Start with a shared rubric before any student sits for a make-up. Anchor the criteria in disciplinary substance: accuracy of core concepts, quality of evidence used, and ability to defend reasoning when challenged. Exclude delivery traits such as speaking speed or accent. Evaluating explanations against clear rubric levels keeps a make-up aligned with the standards applied to the main exam, an approach we examine in our article on AI-supported rubric evaluation.
Design make-up assessments so they remain accessible for neurodivergent students, multilingual learners, and students experiencing acute test anxiety. Share the rubric and general scope ahead of time so the format feels predictable. Offer extended response time, brief pause allowances, or the option to answer in text or audio-only mode when webcam video creates an unnecessary barrier. Our guide on accessibility in oral assessment outlines accommodations that protect academic rigor without penalizing speech differences.
Finally, protect your schedule by defaulting to asynchronous delivery whenever live presence is unnecessary. Finding fifteen-minute office-hour slots across different weeks for absent students creates needless coordination work. Asynchronous spoken or text-based make-up windows let the student complete the assessment within forty-eight hours of recovery, while leaving you with a recording and transcript you can grade in five minutes. Compare scheduling trade-offs in our breakdown of synchronous vs. asynchronous oral assessment.
Where Rocketproof Fits Into Make-Up Workflows
Rocketproof includes workflows built for make-up and alternative assessments so instructors do not have to piece together video calls and spreadsheets.
Inside the platform, faculty can create a dedicated Makeup Assessment assignment, supporting both replacement attempts for missed evaluations and class-makeup syntheses grounded in a recorded session or notes. Students can also submit a makeup request directly from an assignment when they need a replacement attempt, and instructors review, approve, or deny those requests from the assignment page.
For exam-style or applied replacements, Rocketproof's oral exams and AI simulations support three response modes set per assignment: Spoken Response (recorded audio or video), Live Conversation (real-time voice), and Text Conversation. Those modes make it straightforward to provide extended time, pause allowances, or audio-only and typed accommodations without building a separate workflow for one student.
Rocketproof's built-in AI agents, including the Assignment Creator and Rubric Builder, draft make-up prompts, follow-up questions, and rubric criteria from your course objectives, while instructors review and publish every assignment themselves. Once a student finishes, the review screen places the recording, full transcript, and rubric criteria side by side, and the faculty member enters and releases the final grade. For campuses using Canvas, Blackboard, Moodle, or D2L Brightspace, LTI 1.3 LMS integration lets students launch from the course and passes released grades back to the gradebook, while direct assignment links work for courses outside an LMS. To see an asynchronous workflow firsthand, explore the interactive demo or read our guide on how to run an authentic AI assessment pilot.
Frequently Asked Questions
Is an oral make-up exam fair compared to a written exam?
Yes, when both assessments map to the same course learning outcomes and use clear rubric criteria. Because a make-up cannot mirror every multiple-choice item on a written test without leaking questions, focusing the oral exam on three or four core concepts and grading solely on conceptual accuracy and reasoning ensures fair construct equivalence.
How long should an oral or scenario-based make-up exam take?
Ten to fifteen minutes is usually enough for a single midterm or unit make-up. Three well-chosen core prompts, each followed by one or two targeted probing questions, give instructors clear evidence of whether a student understands the material without exhausting the student.
What if a student has speech anxiety or an accommodation that makes live speaking difficult?
Instructors can keep interactive follow-up questioning while adjusting the modality. Offering asynchronous recording with preparation pauses, audio-only responses without webcam video, extended time limits, or text-based dialogue ensures students are evaluated on disciplinary mastery. You can read more about our platform standards on the Rocketproof accessibility page.
How can instructors prevent make-up exam questions from leaking to other students?
Avoid reusing static written prompts across multiple days. Pair a broader pool of core objectives with dynamic follow-up questions that adapt to each student's explanation. Even if a student learns the general topic from a classmate, they cannot script an interactive follow-up exchange around an unfamiliar constraint.
Moving Make-Up Exams From Exception Handling to Course Design
Every syllabus eventually meets real-world disruptions, and make-up exams should not force faculty to choose between spending Sunday afternoon writing a brand-new test and handing out an unverified take-home prompt. When make-up assessments center on outcome equivalence, structured rubrics, and brief interactive explanations, absent students get a fair chance to show what they learned and instructors keep their grading hours intact. Before next semester begins, pick the highest-friction exam in your course and define a three-criterion oral or scenario equivalent in your syllabus so the next illness or travel conflict is already solved.
See Rocketproof in action
Explore the interactive demo as a student and an instructor, or request a pilot for your department.
