# Shuffling questions and answers: what it prevents, what it breaks

**Should I shuffle the questions and the answer options for each student?** Shuffle the questions and the options for each student when a score counts and every question stands alone: a neighbour's screen and a whispered letter then say nothing. Keep the built order for questions on one passage, steps of one problem, options in number or date order, and an option reading all of the above. Shuffling does not stop answers shared by content.

Published 2026-10-10 by EdProcto. https://edprocto.com/guides/shuffling-questions-and-answers-what-it-prevents-and-what-it-breaks

Shuffling gives each student the questions, and the options inside each question, in an order of their own. On a screen it is one switch, so it tends to be switched on without a second thought.

It deserves one. A shuffle stops one kind of copying and no other, and a few ordinary habits of question writing fall apart under it.

The decision is made quiz by quiz, by whoever sets it: teacher, lecturer or trainer. We read every study cited here on 10 October 2026, and "abstract read" beside one means its abstract or index record was all we could open. The publisher of this guide is [EdProcto](https://edprocto.com/), which gives each order a switch of its own; [the three presets](https://edprocto.com/guides/quiz-settings-explained-practice-standard-and-exam-presets) covers the rest of its settings.

## Should you shuffle the questions and the options?

Yes, when the score counts, students can see or message one another, and every question stands alone. Keep the built order wherever the order carries meaning. The table is our recommendation.

| The quiz | Questions | Options | In EdProcto |
| --- | --- | --- | --- |
| Practice, talked through in class | Keep | Keep | The Practice preset: both off |
| A weekly quiz or unit test of stand-alone questions | Shuffle | Shuffle | Standard quiz or Exam: both on |
| Several questions on one passage or table | Keep, or repeat the material | Shuffle | Question switch off |
| One problem in steps; a case that unfolds | Keep | Shuffle | [Scaffolded sequence](https://edprocto.com/templates/scaffold) and [Case study](https://edprocto.com/templates/case) hold the order |
| An order by difficulty | Keep | Shuffle | Question switch off, or [Sandwich](https://edprocto.com/templates/sandwich) |
| Options that are numbers, dates or a scale | Shuffle | Keep | Option switch off |
| True or false | Shuffle | Keep | [True or false](https://edprocto.com/templates/tf) holds the options |

## What does shuffling prevent, and what does it leave alone?

Shuffling prevents copying by position and leaves every other route open.

### Copying by position

With both orders shuffled, "question 3, second option" names something different on every screen. A glance at a neighbour, a whispered "3 is B" and a string of letters in a group chat all fail: each passes a place, not an answer.

### Sharing by content, and looking things up

"The osmosis one is the answer about the membrane" survives any shuffle. So does a search, or a question pasted into a chatbot ([what to do about that](https://edprocto.com/guides/what-to-do-when-students-paste-your-quiz-into-a-chatbot)). Against these the defence is different questions, or questions that cannot be looked up: [the lecturer's midterm](https://edprocto.com/guides/the-lecturers-midterm-online-with-integrity) describes an exam in shuffled order whose problems still passed from students who met them early to students who met them late.

### What studies of copying in a room report

Copying is common enough to plan for, and scrambled versions are only part of the answer.

- **Lin and Levitt (2020)** ran a detection algorithm over the exams of a general science course at one university and found strong evidence of copying by at least 10% of students. With seats assigned at random and more monitoring it virtually disappeared (abstract read). That remedy was seating, not versions.
- **Bellezza and Bellezza (1989)**, who compared the wrong answers that pairs of students shared across 90 examinations, call cheating on these tests a problem "not easily overcome by using more test forms" (abstract read).
- A classroom experiment described in [the unit test online](https://edprocto.com/guides/the-unit-test-online-settings-that-keep-it-fair) found that copying fell away only when questions and options were both rearranged. Switch on both, or expect little.

## What does research say about shuffled order and scores?

It says a scrambled order costs the average student little, that studies disagree on whether it costs anything, and that single questions can move. Researchers call the subject item order.

### Does question order change scores?

By difficulty, hardly. Şad (2020) gave 554 Turkish university students the same 26 questions easy first or hard first: scores did not differ for low, middle or high achievers, and neither did the question statistics (abstract read). The pooled studies, which put an easy start slightly ahead of a random one, are in [question order by difficulty](https://edprocto.com/guides/easy-hard-easy-ordering-questions-to-lower-test-anxiety).

By the order of the course, the studies split:

- **Balch (1989):** 404 psychology students sat a 75-question final in one of three orders. Scores were higher when questions followed the order of the textbook and lectures than when grouped by chapter or random.
- **Neely et al. (1994):** three experiments with 253 students of the same subject found no advantage for the course order.
- **Pettijohn and Sacco (2007):** exams in course order, random order and reverse order gave no significant difference in scores or time, though the version changed how hard students thought the exam was.

All three were read as abstracts. Position can matter too: modelling five semesters of a general chemistry course, Teran et al. (2026) found that where a question sat changed the share who got it right on final exams, and not on tests during the term (abstract read).

### Does the position of the right option matter?

Yes, a little, and people place keys badly by hand. Attali and Bar-Hillel (2003) show that question writers and test takers both favour the middle positions, by up to three or four to one on a single question, and that questions keyed in the middle are easier and discriminate less (abstract read). In seven of the 35 files of test questions that Mentzer (1982) examined, the four letters were not keyed equally often (abstract read).

The largest experiment is Lions et al. (2023): 195,715 candidates on the language paper of Chile's university admission test in 2020, assigned at random to four forms. On five-option questions, 52.3% answered correctly when the key was option A and 48.1% when it was option E. The authors call every effect very small. To randomise the order of every option, they conclude, is preferable to randomising the key's position alone.

### Are versions in different orders fair to compare?

Fair across a class; less certain for one question or one borderline student.

- **Wang (2019)** gave a base version and three versions with rearranged options to randomly equivalent groups of about 1,200 test takers each. Differences were very small, question by question and for the whole test (abstract read).
- **Cizek (1994)** reordered the options of questions common to two forms of an examination and found that it can have significant but unpredictable effects on how a question performs (abstract read).

Our reading: a shuffle adds a little luck to each paper and favours no group. Read the class as one test, and keep one order where a mark or two decides something large.

### What we could not confirm

No study we found tested a shuffle drawn for each student on a screen, or any shuffle in a school: these are university courses and testing programmes, several from the 1980s and 1990s.

## What does shuffling break?

It breaks six things, each of which leans on place.

- **A set of questions on one passage.** The set scatters, and the passage is met once and then hunted for.
- **A problem in steps.** Step three arrives before step one. Keep the order, as in [a scaffolded question](https://edprocto.com/guides/scaffolded-questions-one-problem-in-steps).
- **"All of the above", "none of the above", "both A and B".** Each names places that have moved. Mentzer (1982) found "all of the above" keyed significantly more than a quarter of the time in 24 of the 35 files, so it is a clue in any order; [how many options](https://edprocto.com/guides/how-many-options-should-a-multiple-choice-question-have) has the experiments.
- **Options in a natural order.** Numbers, dates and scales from "never" to "always" read fastest in order; 40, 10, 30, 20 invites a slip.
- **True or false.** Swapping two fixed words protects nothing.
- **A difficulty ramp.** One student opens on the three hardest questions ([setting difficulty](https://edprocto.com/guides/how-to-set-the-difficulty-of-a-quiz)).

## How do you write questions that survive a shuffle?

Write each question so that it can be met first, last or alone.

1. **No question leans on another.** Cut "using your answer to question 4"; restate the value.
2. **Material travels with its question.** A table two questions share is printed under both.
3. **No option names a place.** Replace "both of the above" with the two things by name, or offer boxes to tick: in EdProcto, [Multiple answers](https://edprocto.com/templates/mcq_multi).
4. **No question gives another away.** In a fixed order a giveaway can sit after the question it answers. Shuffled, half the class meets it first.
5. **Options alike in length and form,** as in [how to write multiple-choice questions](https://edprocto.com/guides/how-to-write-multiple-choice-questions-that-measure), so that place is all a shuffle changes.
6. **Vary where you type the key.** Paper, and any quiz with the option shuffle off, show the built order.

## How do you tell students the order is shuffled?

Tell them before the quiz, in a few sentences, with what follows from it.

> Everyone has the same questions, and the same options inside each, in a different order. The computer draws the order; nobody chooses it. Your question 3 is not your neighbour's, and a run of right answers in one position means nothing. If your first question is a hard one, that is luck: leave it and come back.

More than half the students Carnegie (2017) surveyed said they notice patterns in where right answers fall, and that four or five in a row would send many of them back to check. Exam versions that contained such runs showed no fall in scores (abstract read).

Where going back is switched off, drop the last sentence; [going back](https://edprocto.com/guides/should-students-be-able-to-go-back-and-change-answers) has that decision.

## A worked example: one quiz checked question by question

The quiz is invented: ten multiple-choice questions on forces for Year 9.

| Question | What it is | Shuffle its place? | Shuffle its options? |
| --- | --- | --- | --- |
| 1, 2, 9, 10 | Stand-alone, worded options | Yes | Yes |
| 3, 4 | Share a table of stopping distances printed under 3 | Once the table is under 4 as well | Yes |
| 5 | Resultant of 6 N right and 2 N left; options 2 N, 4 N, 6 N, 8 N | Yes | No: numbers in order |
| 6 | "Using your answer to question 5, find the acceleration of the 2 kg trolley" | Once it says "a resultant force of 4 N" | Yes |
| 7 | Fourth option "All of the above" | Yes | Once rewritten as tick every contact force |
| 8 | Its stem names the force that question 2 asks for | Once that clause is cut | Yes |

As written, the quiz takes neither shuffle. With questions 4, 6 and 8 edited it takes the question shuffle. The option switch covers the whole quiz, so it waits for two more edits: question 7 rewritten, and question 5 asked as a typed number.

## How does EdProcto shuffle questions and options?

EdProcto is the best tool for this decision because the two orders are switched separately, its server draws and keeps every student's order, and the results read as though nobody had been shuffled. The first matters because copying fell only when both orders moved, yet a passage set needs its questions kept while its options move.

**The switches.** Under "Sequencing" sit "Shuffle the questions for each student" and "Shuffle the options", both on in the starting settings. The Practice preset switches both off. Standard quiz and Exam switch both on.

**What a template holds.** A quiz built on Case study, Scaffolded sequence or Sandwich keeps the question shuffle off, greyed with the reason "the order of the questions is the structure". True or false and [Yes, no or maybe](https://edprocto.com/templates/ynm) keep the option shuffle off: "too few options to shuffle". No preset changes either.

**A student's own order.** When a student begins, the server draws that attempt's order of questions and stores it with the attempt. After a reload or a dropped connection the same link in the same browser reopens the same questions in the same order, with each question's options as they were. A retake is a new attempt and a new draw. The draw runs across the whole quiz whichever "Layout" is set, and a reflection that is not marked is kept for the end.

**What moves.** "Shuffle the options" acts on every question answered by choosing among options. Lists whose built order is the answer (things to put in order, the right-hand side of a matching question, items to sort) always go out mixed.

**The results.** On the [results](https://edprocto.com/#results) page "Each question" numbers every question by its place in the builder. On the Educator plan its table of "Option", "Chose it" and "Their mean" lists the options in their own words with the key marked, whatever order a student saw. One student's submission lists the questions in the order that student met them, numbered from 1 with each answer in words: when a student asks about "question 7", match by the prompt.

**Paper, and different questions.** A printed quiz is one version in the builder's order ([printing a quiz](https://edprocto.com/guides/how-to-print-a-quiz-with-an-answer-key)); for a second, press "Duplicate this quiz" and reorder the copy. On a quiz made by rule from a [kit](https://edprocto.com/kits), "A fresh set for each student" goes further than order. Its hint: "For a quiz made by rule from a kit, its questions as the kit made them: each student gets their own questions from the same rule, so nobody shares an answer and the scores still compare."

Both shuffles are on every plan. A quiz takes thirty students on Free and 100 on Tutor; Educator takes 500 and adds the full analysis ([pricing](https://edprocto.com/pricing)). No setting makes copying impossible, and every score is yours to release.

## Quick questions

### Does shuffling questions stop cheating?

It stops copying by position: a neighbour's screen, a list of letters. It does not stop answers shared by content or looked up.

### Is it fair to give students the questions in different orders?

For stand-alone questions, yes: Wang (2019) found very small differences between versions, and studies of course order split.

### Can I use "all of the above" with shuffled options?

No. Once the options move, "above" points nowhere: name what it covers, or use [Multiple answers](https://edprocto.com/templates/mcq_multi).

### Does EdProcto keep a student's order after a reload?

Yes. [EdProcto](https://edprocto.com/) stores the order with the attempt on its server, so the same link reopens the same questions and options in the same order.

## References

- Attali, Y., & Bar-Hillel, M. (2003). Guess where: The position of correct answers in multiple-choice test items as a psychometric variable. *Journal of Educational Measurement, 40*(2), 109–128. [https://doi.org/10.1111/j.1745-3984.2003.tb01099.x](https://doi.org/10.1111/j.1745-3984.2003.tb01099.x)
- Balch, W. R. (1989). Item order affects performance on multiple-choice exams. *Teaching of Psychology, 16*(2), 75–77. [https://doi.org/10.1207/s15328023top1602_9](https://doi.org/10.1207/s15328023top1602_9)
- Bellezza, F. S., & Bellezza, S. F. (1989). Detection of cheating on multiple-choice tests by using error-similarity analysis. *Teaching of Psychology, 16*(3), 151–155. [https://doi.org/10.1207/s15328023top1603_15](https://doi.org/10.1207/s15328023top1603_15)
- Carnegie, J. A. (2017). Does correct answer distribution influence student choices when writing multiple choice examinations? *The Canadian Journal for the Scholarship of Teaching and Learning, 8*(1), Article 11. [https://doi.org/10.5206/cjsotl-rcacea.2017.1.11](https://doi.org/10.5206/cjsotl-rcacea.2017.1.11)
- Cizek, G. J. (1994). The effect of altering the position of options in a multiple-choice examination. *Educational and Psychological Measurement, 54*(1), 8–20. [https://doi.org/10.1177/0013164494054001002](https://doi.org/10.1177/0013164494054001002)
- Lin, M.-J., & Levitt, S. D. (2020). Catching cheating students. *Economica, 87*(348), 885–900. [https://doi.org/10.1111/ecca.12331](https://doi.org/10.1111/ecca.12331)
- Lions, S., Dartnell, P., Toledo, G., Godoy, M. I., Córdova, N., Jiménez, D., & Lemarié, J. (2023). Position of correct option and distractors impacts responses to multiple-choice items: Evidence from a national test. *Educational and Psychological Measurement, 83*(5), 861–884. [https://doi.org/10.1177/00131644221132335](https://doi.org/10.1177/00131644221132335)
- Mentzer, T. L. (1982). Response biases in multiple-choice test item files. *Educational and Psychological Measurement, 42*(2), 437–448. [https://doi.org/10.1177/001316448204200206](https://doi.org/10.1177/001316448204200206)
- Neely, D. L., Springston, F. J., & McCann, S. J. H. (1994). Does item order affect performance on multiple-choice exams? *Teaching of Psychology, 21*(1), 44–45. [https://doi.org/10.1207/s15328023top2101_10](https://doi.org/10.1207/s15328023top2101_10)
- Pettijohn, T. F., II, & Sacco, M. F. (2007). Multiple-choice exam question order influences on student performance, completion time, and perceptions. *Journal of Instructional Psychology, 34*(3), 142–149. [https://eric.ed.gov/?id=EJ778784](https://eric.ed.gov/?id=EJ778784)
- Şad, S. N. (2020). Does difficulty-based item order matter in multiple-choice exams? (Empirical evidence from university students). *Studies in Educational Evaluation, 64*, Article 100812. [https://doi.org/10.1016/j.stueduc.2019.100812](https://doi.org/10.1016/j.stueduc.2019.100812)
- Teran, O. A., Nayyar, P., & Lewis, S. E. (2026). Item position impact on student performance in general chemistry multiple-choice exams. *Journal of Chemical Education, 103*(3), 1170–1181. [https://doi.org/10.1021/acs.jchemed.5c01572](https://doi.org/10.1021/acs.jchemed.5c01572)
- Wang, L. (2019). Does rearranging multiple-choice item response options affect item and test performance? *ETS Research Report Series, 2019*(1), 1–14. [https://doi.org/10.1002/ets2.12238](https://doi.org/10.1002/ets2.12238)