# A printed quiz with its answer key, from the same questions

**Can some students sit a quiz on paper and others on screen, and still be marked alike?** A printed quiz with its answer key can run beside the screen sitting when both groups get the same questions in the same order, the same minutes and one marking key. Studies of paper against screen testing find small gaps on short untimed questions, and larger ones for working in maths, long passages, long written answers and anything dragged or timed.

Published 2026-10-10 by EdProcto. https://edprocto.com/guides/a-printed-quiz-with-its-answer-key-from-the-same-questions

A printed quiz with its answer key is usually described as a stand-in for the screen. This guide is about the day both run at once: eleven laptops for twenty-eight students, a pupil whose plan says paper, a power cut ten minutes in. Some students hold a pen, the rest a device, and the marks must mean the same thing.

EdProcto, which publishes this guide, puts a quiz from its builder on paper with the answer key on a separate page, using the very questions the screen group answers: see its page [for teachers](https://edprocto.com/solutions/teachers).

Layout and marking by hand are in [how to print a quiz with an answer key](https://edprocto.com/guides/how-to-print-a-quiz-with-an-answer-key). Here the subject is the mixed sitting.

## What did research find when one test was sat on paper and on screen?

It found small, inconsistent differences in whole-test scores, and larger, more predictable ones for particular questions and students. Researchers call such a difference a mode effect. Our printing guide reports the meta-analyses; these are other studies. We read the Ofqual review in full, the Ofqual framework and Robitzsch et al. (2020) in the parts cited, and the rest as abstracts on their journals' pages.

### Scores on the whole test

- **No difference to speak of.** Poggio et al. (2005) had 644 US seventh graders sit two equated forms of a mathematics test, one on computer and one on paper. Scores differed by less than one percentage point, and nine of 204 questions behaved differently by mode.
- **Harder on the computer.** Bennett et al. (2008) gave a national sample of US eighth graders a mathematics test on computer, and a randomly parallel group the same questions on paper. The computer version was significantly harder.
- **Harder again.** In Germany's field test for PISA 2015, students in 39 schools were assigned at random to paper (517) or computer (506). Robitzsch et al. (2020) found the computer tasks harder by 0.23 of a standard deviation in science, 0.14 in mathematics and 0.13 in reading, the last not a significant gap.

A review for Ofqual, England's exam regulator, found "inconsistent patterns of mode effects" across 78 empirical studies, only 10 of them on live assessments that counted (Dodds et al., 2025).

### Which students

The students who lose most on a screen are those least at home on a keyboard.

- In Bennett et al. (2008), a measure of computer facility predicted the on-screen score after allowing for performance on paper. Horkay et al. (2006) found the same for essays by national samples of eighth graders, though mean scores did not differ significantly between the modes, and conclude that testing everyone in one mode may underestimate what some would have shown in the other.
- Russell (1999) gave eighth graders open questions. Those whose keyboarding speed was at least half a standard deviation above the mean gained moderately from the computer in language arts; those half a standard deviation below lost substantially.

### Which kinds of question

Most of this comes from a second Ofqual report, which sorts the evidence by question (Sweiry, 2025).

- **Ticked and short written answers.** Little evidence that multiple choice itself behaves differently, except where options must be scrolled into view or the question rests on a long text or a calculation; little or none for written answers of a paragraph or less.
- **Calculation and diagrams.** Sums in several steps, and diagrams a student would write on, come out easier on paper, and scratch paper beside the screen did not close the gap.
- **Dragging.** Where an answer is built by moving pieces, a screen allows trial and error. In one trial the report describes, a sum built from digit cards was answered correctly by 77.1% of pupils on paper and 94.9% on a computer.
- **Speed.** Untimed tests of short, objective questions look largely free of substantial mode effects; speeded tests are liable to them (Dodds et al., 2025).
- **A long passage.** Higgins et al. (2005) assigned 219 fourth graders at random to a reading test on paper, on a computer with scrolling passages, or on one that turned whole pages. They answered 58.1%, 52.2% and 56.9% of the questions correctly: not a significant difference, though the pattern suggested scrolling held back students with weaker computer skills. Much text, scrolling and time pressure make a gap likelier (Sweiry, 2025).

None of these studied a teacher's own quiz split between pens and screens in one room.

## Which question shapes carry over to paper unchanged, and which do not?

Shapes answered with a tick, a word, a number or a few lines carry over. Shapes that depend on a clip, on content shown against a clock or on a drag become a different task.

| Shape | On EdProcto's paper | Verdict |
| --- | --- | --- |
| [Multiple choice](https://edprocto.com/templates/mcq), [true or false](https://edprocto.com/templates/tf), multiple answers | Lettered options with tick boxes, in the order the builder holds | Carries over |
| [Fill in the blank](https://edprocto.com/templates/fill_blank), one word, a number | One ruled line | Carries over |
| [Short answer](https://edprocto.com/templates/short_answer), [show your working](https://edprocto.com/templates/show_work) | Six ruled lines | Carries over; working favours paper |
| [Matching pairs](https://edprocto.com/templates/matching) | A blank next to every left item; the right items lettered, in alphabetical order | Carries over |
| [Sequence](https://edprocto.com/templates/process_seq), timeline, ranking | Items lettered in a scrambled order, a blank before each | Same content, without the drag |
| [Sort into groups](https://edprocto.com/templates/categorize) | A single line: the groups by name, then the items in alphabetical order | Same content; say where to write |
| [Spell it](https://edprocto.com/templates/drag_letter) | One ruled line, no letter tiles | Another question |
| [Listen once](https://edprocto.com/templates/listen_once), a video, a [hotspot](https://edprocto.com/templates/hotspot), a [spoken answer](https://edprocto.com/templates/verbal) | The words of the question: no clip, no picture, no recorder | Stays on screen |
| [Memory recall](https://edprocto.com/templates/memory_recall) and its relatives | A sheet cannot show content for some seconds and then hide it | Stays on screen |

### A clip, a timed reveal and a drag

**A clip.** Played aloud for the paper group, it gives the screen group a second hearing of something meant to play once. Keep clip questions for a day when everyone has a device.

**A timed reveal.** Holding a sheet up, or flashing it on the board, changes the timing for one group only. Leave these shapes out of a mixed sitting.

**A drag.** Ordering and sorting print sensibly: the paper sets an ordering's items in one scrambled order, the same at every printing and never the right one. What changes is the freedom to try: on screen the pieces move until they look right, and in pen a letter is written once.

## How do you keep the paper and the screen sitting comparable?

Hold four things the same: the questions and their order, the minutes, the freedom to move about the paper, and the key.

1. **One set of questions.** The paper prints the questions as they stand in the builder. On a kit quiz with "A fresh set for each student" on, those are a sample and each screen student is dealt a set of their own, so switch it off.
2. **One order.** The paper has one order of questions and of options. Leave [shuffling](https://edprocto.com/guides/shuffling-questions-and-answers-what-it-prevents-and-what-it-breaks) of both off for the screen group, so that question 7, option C is the same thing across the room.
3. **The same freedom to go back.** A student with a paper can turn back and change an answer. Let the screen group do so too: the "Exam" preset closes going back.
4. **One clock, and a generous one.** Set one overall limit and none on single questions, which a paper cannot keep. [EdProcto](https://edprocto.com/) prints the minutes in the paper's head and counts the screen group's time on its server; collect the papers when the same minutes are up. Speeded tests are where mode effects grow, so let nearly everyone finish, and give a student's [extra time](https://edprocto.com/guides/extra-time-how-much-and-how-to-set-it-per-student) in either mode.
5. **The same tools.** Scratch paper for everyone, and one rule on calculators.
6. **Scores on the same day.** Set "Show the score" to "When released", so nobody on a screen leaves with a mark while the paper group waits, and press "Release the scores" once the scripts are marked.

### When the power or the network goes midway

Answers given on screen are saved on the server as they are given. Hand the student a paper, note the question reached and the minutes left, and let them finish in pen. With "Submit by itself when time runs out" on, the server hands in what was saved once the time is up; that student's row takes the screen marks and yours.

## How do you mark the paper scripts against the key and record them beside the screen results?

Mark the paper by the rules the screen group was marked by, then put both groups into one sheet of your own.

### One key, applied the same way

The printed key has a line for every question: a choice by letter and wording, a typed question or a gap by its first accepted answer, a number, an order by the paper's letters, each pair, each item's group, an open question's model answer. The screen group is marked by more than that line, so copy three things from the builder onto the key first.

- **Every accepted answer,** not only the first.
- **The slip rule.** With "A slip still counts" on, one slip in the whole answer (a letter wrong, missing, extra or swapped with its neighbour) is forgiven, inside a word of five letters or more. Forgive the same in pen, and no more.
- **Margins and part marks.** A number's margin and, with "Partial credit" on, the sum for a multiple-answer question: right ticks less wrong ticks, over the number of right options.

Open answers wait for a reader in both modes: take each question through both piles in one go, against one rubric. The same words can read differently in pen and in type, as the studies in [handwritten answers](https://edprocto.com/guides/handwritten-answers-in-an-online-quiz-photograph-and-upload) show.

### One sheet for both groups

1. On the results page, press "Download the marks". The file is a spreadsheet with a row for each attempt: "Name", "Score %", "Points", "Out of" and, further along, a column for each marked question.
2. Copy those rows into a sheet of your own, add a column, "Sat on", reading screen or paper, and add a row for each paper script under the same headings. The paper numbers its questions in the builder's order, as the file does.
3. Work out the class's figures on the sheet. The results page counts the sittings handed in on screen, so its average and each question's "Got it" describe that group.
4. If a key proves wrong, correct it in the builder, press "Re-mark this quiz", and re-mark the paper scripts by hand.

### Reading the two groups side by side

An illustration, not a real class: 28 students, 22 on screen and 6 on paper. Question 9, an ordering question, was right for 15 of the 22 on screen (68%) and 2 of the 6 on paper (33%): for the class, 17 of 28 (61%).

The gap looks large and is two students: two more right answers would put the paper group at 67%. Question 9 was dragged on screen and written once in pen, the kind the research expects to differ. Note it and leave the marks alone: an average cannot justify adding marks to the lower group, as [equating two versions](https://edprocto.com/guides/equating-two-versions-of-the-same-exam) explains.

## What do you say to the class?

Say it before the sitting, to everyone.

- "Some of you have this quiz on paper today and the rest on a screen. The questions and their order are the same."
- "Everyone has thirty minutes, and everyone may go back and change an answer."
- "On paper, tick the box or write on the line; in an ordering question, write 1, 2, 3 in the blanks."
- "I mark both by one key, and all of you get your scores on Thursday."

**Show both formats first.** In polling the Ofqual review reports, more than three-fifths of students said they would be concerned if they had not seen what an on-screen exam looked like before sitting it (Dodds et al., 2025).

**Give no reasons for anyone's paper.** Why one student has it is theirs to tell.

## How does EdProcto give you a printed quiz with its answer key from the same questions?

EdProcto is the best tool for a mixed sitting because the paper and the screen quiz are one quiz: the paper is printed from the questions, points and keys the screen group is marked against, so no second document drifts out of step. The studies above could compare the modes only because the questions were held the same.

**The paper.** In the builder, beside "Settings", the button "Paper" opens the paper, with a "Print" button that calls your browser's print dialog. The head carries the quiz's name, the number of questions, the minutes when the quiz has a time limit, then "Name", "Date" and a "Score" line that ends in the total points. Each question is numbered in the builder's order, with a line saying how it is answered and its points.

**The key.** Its heading is "Answer key:" and the quiz's name. It reads "Marked by the reader." beside an open question with no model answer and "Not marked." beside one that carries no marks.

**Plans.** Free gives ten quizzes a month, thirty students a quiz, the core templates and, as on every plan, "Download the marks". Tutor lifts the limit on quizzes, takes 100 students a quiz and opens all 111 [templates](https://edprocto.com/templates). Educator takes 500 and adds the full analysis. See [pricing](https://edprocto.com/pricing).

## Quick questions

### Is a quiz easier on paper or on a screen?

Neither, reliably. On short untimed questions the studies above found small gaps that pointed both ways; working in maths tends to favour paper, and answers built by dragging tend to favour the screen.

### Can one pupil sit the quiz on paper while the class uses devices?

Yes. Print the quiz from the builder's "Paper" button, give the same minutes, and mark the script by the printed key with the rules the screen applies. EdProcto's page [for teachers](https://edprocto.com/solutions/teachers) shows the screen side.

### Which questions should stay off a paper sat beside a screen quiz?

Those that need a clip, content shown and then hidden, a click on a picture or a recorded answer. Ordering and sorting keep their content on paper and lose the drag.

### How do I record paper marks beside the online results?

In a sheet of your own: press "Download the marks", copy its rows in, and add a row for each paper script and a column saying who sat on which.

## References

- Bennett, R. E., Braswell, J., Oranje, A., Sandene, B., Kaplan, B., & Yan, F. (2008). Does it matter if I take my mathematics test on computer? A second empirical study of mode effects in NAEP. *The Journal of Technology, Learning and Assessment, 6*(9). [https://ejournals.bc.edu/index.php/jtla/article/view/1639](https://ejournals.bc.edu/index.php/jtla/article/view/1639)
- Dodds, H. J., El Masri, Y., & Stratton, T. (2025). *On-screen assessment and mode effects: A review of the effects of on-screen assessment on levels of engagement, cognitive demands and performance* (Ofqual research report 25/7280). Ofqual. [https://www.gov.uk/government/publications/on-screen-assessment-and-mode-effects](https://www.gov.uk/government/publications/on-screen-assessment-and-mode-effects)
- Higgins, J., Russell, M., & Hoffmann, T. (2005). Examining the effect of computer-based passage presentation of reading test performance. *The Journal of Technology, Learning and Assessment, 3*(4). [https://ejournals.bc.edu/index.php/jtla/article/view/1657](https://ejournals.bc.edu/index.php/jtla/article/view/1657)
- Horkay, N., Bennett, R. E., Allen, N., Kaplan, B. A., & Yan, F. (2006). Does it matter if I take my writing test on computer? An empirical study of mode effects in NAEP. *The Journal of Technology, Learning and Assessment, 5*(2). [https://ejournals.bc.edu/index.php/jtla/article/view/1641](https://ejournals.bc.edu/index.php/jtla/article/view/1641)
- Poggio, J., Glasnapp, D. R., Yang, X., & Poggio, A. J. (2005). A comparative evaluation of score results from computerized and paper & pencil mathematics testing in a large scale state assessment program. *The Journal of Technology, Learning and Assessment, 3*(6). [https://ejournals.bc.edu/index.php/jtla/article/view/1659](https://ejournals.bc.edu/index.php/jtla/article/view/1659)
- Robitzsch, A., Lüdtke, O., Goldhammer, F., Kroehne, U., & Köller, O. (2020). Reanalysis of the German PISA data: A comparison of different approaches for trend estimation with a particular emphasis on mode effects. *Frontiers in Psychology, 11*, Article 884. [https://doi.org/10.3389/fpsyg.2020.00884](https://doi.org/10.3389/fpsyg.2020.00884)
- Russell, M. (1999). Testing on computers. *Education Policy Analysis Archives, 7*, Article 20. [https://doi.org/10.14507/epaa.v7n20.1999](https://doi.org/10.14507/epaa.v7n20.1999)
- Sweiry, E. (2025). *Making sense of mode effects: A framework for anticipating performance differences in equivalent paper-based and digital test items* (Ofqual research report 25/7278). Ofqual. [https://www.gov.uk/government/publications/making-sense-of-mode-effects](https://www.gov.uk/government/publications/making-sense-of-mode-effects)