Reviewing Make-Up Exam Eligibility Evidence

Protocol status

The reviewable question is not whether a make-up was convenient, but whether the sitting can be reconstructed from a dated eligibility decision, a defined assessment equivalent, and a proctor record that matches the student and the time allowed.

This page describes a review design and its proposed controls. It does not report an executed sample, measured outcomes, or evidence that the controls improved performance.

Review protocol

Planned review sequence

1. Define the population from a fixed term extract of make-up sittings recorded in the assessment or testing-center system. Include instructor-arranged make-ups, accommodation sittings, and departmental retakes, and record the count by type. State the extract date, the courses, and the systems joined. Exclude practice or ungraded sittings only under a written rule and keep the excluded count.
2. Link each make-up to its original assessment. Record the course, the assessment identifier, the original date and time, the delivery mode, and the approved duration. A make-up that cannot be linked to an original sitting cannot be tested for equivalence. If the original assessment was replaced or cancelled, record that decision instead of forcing a link.
3. Identify the eligibility decision and its owner. Record who approved the make-up, the date, the scope, and whether the approval covered the date, the format, or both. An accommodation approval from an accessibility office is not the same decision as an instructor's course-policy approval. Record both when both exist, and flag a case where only one is present but the policy requires two.
4. Collect the reason category and the evidence status without copying sensitive detail into the review file. Categories may include illness, family emergency, institutional conflict, religious observance, or accommodation. Record whether the required evidence was present and who reviewed it. The review does not assess whether the reason was true or whether the documentation was clinically sufficient.
5. Test the format equivalence. Record whether the make-up used the same form, a parallel form, an oral format, or another approved variation, and who approved the change. Compare the time allowed, the permitted materials, and the delivery platform to the original assessment. A difference is not automatically a defect, but it must be documented and authorized.
6. Trace the proctor and room record. Record the proctor identifier, the room or platform, the scheduled start and end, the actual start and end, and the person who confirmed attendance. Match the student identifier to the eligibility decision. A proctor record for the right time but the wrong student is a serious mismatch and should be sampled separately.
7. Check the security chain for the assessment materials. Record who accessed the secure file, when it was released to the proctor, and whether the same form had already been returned to another group. If the original form was exposed, record the assessment owner's decision about reuse. The review can show the exposure history; it cannot determine whether a student actually saw the material.
8. Define outcome categories before comparing to the gradebook
9. Stratify by course, assessment type, delivery mode, and approval owner. Include at least one case with an accommodation, one with a same-day request, and one that was denied. A sample of only completed sittings will overstate the reliability of the process. If the volume is small, review every sitting and label the work a census.
10. Compare the student-facing instructions to the scheduling record. Record the date, time, location or link, allowed materials, and the contact path that were sent. Check whether the message described the correct assessment and duration. A correct booking with an incorrect student message can still produce a missed sitting.
11. Review a sample of no-shows and cancellations for a process cause. Record whether the student confirmed receipt, whether the link or room changed, and whether the proctor was present. Do not treat a no-show as a student choice until the scheduling evidence has been checked. The review should distinguish an access failure from an unexplained absence.
12. Test the join between the assessment system, the testing-center schedule, and the gradebook before sampling. Record whether the original assessment identifier is present on the make-up record and whether the student identifier is consistent across systems. Report unmatched and duplicate rows. A make-up that cannot be joined is not automatically invalid, but it cannot be tested for format or outcome equivalence until the link is corrected.
13. Pilot the field list on one department or testing center. Record the sittings where the approval type was unclear, the proctor record was missing, or the original assessment had more than one version. Revise the definitions once, then freeze them. Keep the pilot cases separate so the final counts use a single method.
14. Define decision rights for a disputed sitting. Name who can reopen an eligibility decision, approve a retroactive make-up, or correct a gradebook entry after the term closes. Record the authority and the date. The assistant can prepare the case file, but should not grant an exception because a review found a gap.
15. Review the student communication sample for clarity and consistency. Check whether the message named the course, the assessment, the date, the duration, the location or link, and the contact path. Record whether the student confirmed receipt and whether a later change was sent through the same channel. A correct booking with a wrong message can still produce a missed sitting.
16. Set the reporting audience and cadence. A testing center needs scheduling counts, an instructor needs section-level outcomes, and an accommodation office needs its own approval records. Report aggregates and de-identified examples. Identify the person who will follow up on remedies and set a date for the next review. The report should not become a ranking of individual proctors or instructors.
17. Check the assessment version log for the original and make-up sittings. Record whether the same form, a parallel form, or a revised key was used, and whether the version was approved before the sitting. A make-up that used an unapproved revision cannot be treated as equivalent merely because the title matched. Cases that cross a version change should be reported separately.
18. Document the sample selection record. Describe how the sittings were selected and what happened when a selected case could not be reviewed. Keep the seed or selection file so the review can be repeated. A convenience sample of sittings that were easy to locate will not support a conclusion about the population.
19. Test the correction and appeal trail. For any sitting that was rescheduled, regraded, or disputed, record the original decision, the new evidence, the approver, and the date. Confirm that the gradebook and the student record reflect the correction. A fix that exists only in an email leaves the official outcome unresolved.
20. The review does not judge whether an absence should be excused, and it should say so plainly. The review does not decide whether an absence should be excused, whether a grade was fair, or whether a student met a clinical fitness standard. Those decisions belong to the named academic or clinical authority. The review reports whether the eligibility and scheduling evidence can be reconstructed.
21. Report the population, the sample, the completeness of each required field, the match rate across decision, sitting, proctor, and gradebook, and the limits. Retain the extracts, the sampling frame, the approval references, and the review notes. The protocol measures whether the process can be reconstructed; it does not measure learning, fairness in grading, or whether the original absence was excusable.

Operational implications

A make-up sitting should be linked to a specific original assessment before equivalence can be reviewed.

Instructor approval and accommodation approval answer different questions and should be recorded separately.

The approval should state whether it covers the date, the format, or both.

Format and time differences are reviewable only when the approving owner and the change are documented.

The proctor record must match the student, the assessment, and the approved duration.

Assessment material exposure should be recorded even when the same form is reused with approval.

Completed but ungraded sittings are a distinct outcome from unscheduled or denied requests.

A no-show needs a scheduling check before it is treated as a student choice.

Student instructions are part of the evidence because a wrong message can invalidate a correct booking.

A same-day request should be sampled separately from a planned accommodation sitting.

The review measures process reconstructability, not the merits of an absence or a grade.