Sign up free

Retakes and make-ups: a policy that is fair and runs itself

What studies and published rules say about make-up tests, dropped scores and retakes, and how to write a six-line policy that ends case-by-case emails.

The EdProcto editorial teamWritten and checked against the research it cites. Last checked . How the guides are made

Guides · Teaching from the results · · 13 min read

How do I write a fair policy for make-up tests and retakes?

A fair policy for make-up tests and retakes is written before the first test, in six lines: who qualifies, by when, what paper, what mark counts, how many times and how to ask. Allow one missed sitting with no proof, hold make-ups on one common day on a parallel paper, and record the first sitting.

A make-up test and a retake are different things. A make-up is a first sitting for a student who missed the test; a retake is a second sitting for one who sat it. One is about an equal chance, the other about a second one.

This guide is for a teacher or lecturer tired of settling each case by email. It is mostly about the make-up, and about writing both rules down once: how many attempts a quiz should allow has the studies on retakes. Sources were read on 10 October 2026; "abstract read" marks one where only an abstract or index record was open.

EdProcto, which publishes this guide, runs a make-up as a copy of the quiz with a code of its own, and a retake as a setting.

Which make-up and retake policy fits your test?

The one that matches what the score decides and how many students you have. The table is our recommendation.

The test A missed sitting A retake In EdProcto
Weekly quiz, any class No make-up: the lowest two are dropped Practice only "Download the marks"; the sum is yours
Unit test, one class The same paper within two days, scores held; later, a parallel paper One, under the pass mark, on a new paper "Show the score" at "When released"; a copy for the make-up
Midterm, 100 students or more One common make-up day, a parallel paper None, or a second paper for all "Duplicate this quiz": one copy, one code
Final or certificate Your institution's procedure As its regulations say A copy opened by hand for the students named

What do studies say about students who sit a test later?

A later date is rarely an advantage in itself; what travels is the paper.

Is a later sitting an advantage?

Not in the three records we found.

  • Chen et al. (2017) analysed 29,492 exams sat in a supervised computer room at times students chose within a period of days. Scores fell through the period, even allowing for ability on an exam everyone sat together (abstract read).
  • Tarasi (2020) compared biology undergraduates' make-up scores with their scores on the course's other exams: no significant difference. Those away for a school-sponsored activity scored worse than those away for illness or a family emergency (abstract read).
  • Bultas et al. (2024) took 156 examinations sat late in one large nursing programme: scores did not differ significantly from the same students' on-time average or the class average (abstract read).

None assigned the date at random, and Tarasi adds that some advantage cannot be ruled out.

Do the questions travel?

Sometimes. Colliver et al. (1991) followed five classes of final-year medical students whose examination on standardised-patient cases took three weeks, a fifth of the class at a time: case scores showed no consistent change from first group to last (abstract read). Written problems fared worse in the course described in the unit test online, where students tipped off by earlier sitters gained 13 percentage points.

In Perlman and McCann's (2005) experience it is very rare to miss an exam for extra study time, and commoner to ask classmates what was on it. Plan for the paper, not the calendar.

What are the alternatives to a make-up test?

Four are in common use, and only the first has been tested.

Drop the lowest score

Two studies disagree. In a controlled experiment, Sewell (2004) found that dropping the lowest grade made students no more likely to miss an exam or write one off, yet lowered scores on a comprehensive final (abstract read). MacDermott (2013) taught intermediate macroeconomics at one small college under three rules in three semesters: dropping the lowest of three exams improved the cumulative final, and an extra exam replacing the lowest score had no significant effect (abstract read).

Dropping suits many small quizzes, as in weekly quizzes in a lecture course; of three tests, a dropped one is a third of the course.

Let the final exam stand in

The missed mark is replaced by the final's, or by the final's questions on that unit; Perlman and McCann (2005) list both. England's exam boards do something similar for GCSEs and A levels: a student absent from an exam for a reason that meets the published criteria can still be awarded the qualification, provided at least one component was completed (Office of Qualifications and Examinations Regulation, 2025). We found no study comparing a stand-in with a make-up.

One common make-up day

Every missed test of the term is sat at one or two announced times. Perlman and McCann (2005) describe one midway through the semester and another at its end, and warn that students forget it: say it again on every test day.

An oral make-up

Ten minutes of questions replace the paper. Quick to arrange and hard to copy, it is a different test: the same authors warn that it may punish students who "do not think well on their feet". We found no study of oral make-ups; the oral exam is back covers the format.

What happens when a no-questions-asked rule replaces excuses?

Attendance barely moves, on the little evidence there is, and you stop judging stories.

Sewell's (2004) experiment is the nearest test: a dropped score, which needs no excuse, did not change the decision to miss an exam.

Judging excuses has its own cost. Abernethy and Padgett (2010) followed over 400 students in principles of marketing classes across several semesters: the first exam and the final were missed significantly less often than the others, which the authors take as evidence that some students fake an excuse to choose when they sit (abstract read). Proof settles little: Perlman and McCann (2005) note that any student can forge a doctor's letter.

We found no study that counted requests before and after proof stopped being asked for.

One published rule sets a cap: a student may declare an absence once a term, for up to seven consecutive days, with no supporting documentation; a further absence may need documents, and a final exam is outside the rule (University of Toronto Mississauga, n.d.).

Who gains from retakes and mastery rules?

The students who use them, and not all do. Goodman et al. (2020) found that only half of those who sit the SAT, an American university admission test, sit it again, and fewer among low-income and under-represented minority students; retaking raised scores and four-year college enrolment, most for those groups (abstract read). A retake offered on request reaches those who ask.

In schools, mastery programmes, where a class is retaught and retested to a standard, have weaker evidence than their reputation. Slavin (1987) reviewed group-based mastery learning in elementary and secondary schools: essentially no effect on standardised tests, moderate gains on the experimenters' own, and little sign that they lasted (abstract read). The three presets reports a later, more favourable meta-analysis.

So write the retake as a default: a second paper for everyone under the pass mark, after a reteaching lesson.

How do you write the policy in six lines?

Answer six questions, a line each, before the first test. The choices are our working rules.

  1. Who qualifies. Everyone, once, with no reason given; after that, a reason your school accepts.
  2. By when. The common make-up day, or a fixed number of school days after returning.
  3. What paper. A parallel one: the same topics, length, shape and time limit, different questions.
  4. What mark counts. The make-up's, in full, as a first sitting.
  5. How many times. One make-up a test; a missed make-up falls to the stand-in, the final's mark or a dropped score.
  6. How to ask. One route with a deadline: a form, or one line by email, before the make-up day.

One school's handbook is this short: a student who misses a summative exam, having been in the lesson before it, takes it at lunch on the day of return unless a time is arranged with the teacher, and missing the make-up costs an automatic 10 per cent (Geneva School of Boerne, n.d.). Your institution's rules on illness, disability and religious observance come first; this guide is not legal advice.

How do you keep a make-up from being the same paper?

Change the questions, or keep the answers back until the make-up is in.

  • Hold the answers. If the class has seen no score and no answer, a make-up two days later can be the same paper.
  • Write parallel questions. New numbers and a new case, the same idea. Perlman and McCann (2005) want an alternate form of the same exam, no harder and in the same format.
  • Keep the conditions. The same number of questions, rules and time limit.
  • Do not rely on order. Shuffling moves a question and leaves what it asks unchanged.

In EdProcto, "Duplicate this quiz" gives the copy to rewrite.

How do you say it in the syllabus?

In the six lines, in the second person, beside the dates. For a course with three tests:

If you miss a test, you sit its make-up on Wednesday 2 December at 12:30 in room 14. Your first missed test needs no reason: fill in the form before that day. A second needs a note the school accepts. The make-up has the same topics and time and different questions, and its mark counts in full. Miss the make-up, and your final exam mark stands in for that test. If you sat a test and scored under 50%, you sit a second paper the week after, and the average of your two marks is recorded.

Perlman and McCann (2005) suggest having a few students read it first.

What does one term with three missed sittings look like?

Three absences, and no decision made by email. The class and every figure are invented: 27 students, three tests of 25 marks, and the policy above.

  1. Test 1, October. Amira is ill. She fills in the form: her first, so no note.
  2. Test 2, November. Ben is away with a school team and Chloe gives no reason. Each is a first: the form, no note.
  3. Before 2 December. In EdProcto, "Duplicate this quiz" copies tests 1 and 2, and the questions are rewritten: the class saw the answers weeks ago.
  4. Make-up day. Amira scores 19 of 25 on her copy and Ben 21 on his; both count in full. Chloe does not come.
  5. The final. Chloe scores 64%, so 16 of 25 stands in for test 2.

How does EdProcto run a make-up and a retake?

EdProcto is the best tool for a policy that must run without you, because its server keeps the rule: a closed quiz stays closed, a copy has its own code, and attempts are counted where no student can reset them. The risk is the paper, not the date: the later sitters in Chen et al. (2017) scored lower.

A make-up

  1. Before the class sits, set what it will see. "Show the correct answer" takes "At once", "After submitting" or "Never"; "Show the score" takes those and "When released". Choose that if the make-up may be the same paper: "nothing of the score, the marks or the answers reaches a student until you press "Release the scores" on the results page."
  2. After the sitting, press "Close the quiz". A student who opens the link reads "This quiz is not open" and "Its instructor has not published it, or has closed it."
  3. Press "Duplicate this quiz" at the foot of the builder. The copy is a draft titled "… (copy)" with the same questions, keys and settings, a code of its own and no submissions, so "Time limit", "Minutes overall" and "Shuffle the questions for each student" stand as the class had them. Rewrite in place whatever must differ.
  4. On the make-up day, press "Publish & share" on the copy, give its link only to the students who missed, and close it when they have handed in.
  5. Afterwards, each quiz has its own results page and "Download the marks": a spreadsheet file with a row for each sitting, its columns beginning "Name", "Email", "Sitting", "Started (your time)", "Handed in (your time)", "Score %", "Points" and "Out of". Set the make-up's rows beside the class's in your own sheet.

A retake

A retake of the same paper is a setting. Under "Attempts and retakes", "Whole-quiz retakes" takes "None", "1", "3" or "No limit" and starts at "None". "Hours between retakes" runs from 0 to 720: "A retake is refused until this long after the last hand-in." The start screen shows "Attempts" as "1", "2 in all" or "No limit", and with one retake the house rules add "1 retake allowed."

The server counts a sitting once it is handed in, under the email address the student typed. After the last, a start is refused: "You have used every attempt at this quiz." Inside the wait the student reads a line like "Your next attempt opens on Saturday, October 10 at 10:51 AM, by this device’s clock." The address is typed, not checked: the count is a record, not proof of who sat.

On the results page a later sitting is listed under "Students" with "sat again" and left out of the class's figures. A retake on new questions is another copy.

All of this is on every plan. Free takes thirty students a quiz, and a copy counts as one of its ten quizzes a month; Tutor takes 100; Educator takes 500 and adds the full analysis. See pricing and the pages for teachers and for universities. No setting makes cheating impossible, and every score is yours to release.

Quick questions

Is a make-up test the same as a retake?

No. A make-up is a first sitting for a student who was absent; a retake is a second sitting for one who sat the test.

Should a make-up test be harder than the original?

No. Perlman and McCann (2005) want an alternate form of the same exam, no more difficult.

Do students who sit a test late score higher?

Not in the records we read: late scores matched the students' other scores in two courses (Tarasi, 2020; Bultas et al., 2024).

How do I set a make-up test in EdProcto?

In EdProcto, press "Duplicate this quiz", rewrite the questions that must differ, and open the copy for the students who missed with "Publish & share".

References

The settings

Time, attempts, feedback, grading, access, identity, security, tools, accessibility: every rule of a quiz, set by itself.

On every plan

See the settings

Read next