Some answers belong on paper. A page of algebra, a reaction mechanism or a force diagram loses something at a keyboard, while the rest of the test is at home on a screen. Handwritten answers in an online exam join the two: the student writes on paper, photographs the page and uploads it into the same timed attempt.
EdProcto, which publishes this guide, carries that as its handwritten work template: written on paper, photographed, uploaded, and kept with the attempt for you to mark. The instructions, the photographing drill and the marking are yours.
Why set handwritten answers in an online exam?
Set a handwritten answer when the thing you are marking is hard to type: working, a diagram, notation, or something made by hand. A keyboard turns a fraction or a curly arrow into a typing puzzle, and a student wrestling with an input box is not showing you chemistry. Where a solution is short enough to type in named steps, our guide to marking the working in maths questions covers that route.
A second reason: a typed essay finished at home can come from a chatbot in a minute, and a page written by hand in a supervised room is much harder to hand over. That is a first defence and no more, as integrity in the AI era explains.
Do not ask for a photograph when typing would serve.
What did research find on handwritten versus typed exam answers?
It found small differences between the two formats and large differences between markers, with handwriting speed and legibility each able to move a mark. Nearly all of it concerns essays by university students, not photographed maths.
Length, marks and the marker
Mogey et al. (2010) offered about 70 first-year divinity students a one-hour mock examination. Thirty-seven sat it: 24 typed and 13 wrote by hand. Every script was then copied into the other format, and four markers each marked every script once. Those who typed generally wrote more words. The same words scored slightly higher in handwriting, a mean of 55.12 against 53.19, a gap that was not statistically significant. The markers were the bigger problem: the four marks given to any one student's script spanned at least 5 marks and, at the extreme, 38.
Powers et al. (1994) had seen the same direction with 32 examinees who each wrote one essay by hand and one on a word processor. Each essay was converted to the other format and rescored, and average scores were higher for essays read in handwriting, however they were first produced. When the readers' training was changed and the experiment repeated, the gap shrank by about 25% and the remaining effects were small (Powers et al., 1992).
What students choose, and how fast they write
Students' choices were not consistent. In the 2010 mock, almost two thirds of those who sat it chose the laptop. When Mogey et al. (2012) encouraged students to choose between the two for an examination, few opted to type. At one British and one Australian university, most students told Mogey and Fluck (2015) that they typed at least as fast as they wrote.
Speed matters, and it is hard to screen for. Connelly et al. (2005) found undergraduates writing by hand at a speed equivalent to published figures for 11-year-olds, with handwriting fluency accounting for large amounts of the variance in tutors' marks for examination answers. Summers and Catarro (2003), with 66 second-year students, found that a short handwriting speed test did not predict how much a student wrote in exams (r² = 0.072). Read together: slow handwriting costs marks, and a quick test will not tell you whose.
Legibility, and what we could not confirm
Greifeneder et al. (2010) ran three experiments in which essays in legible handwriting were evaluated more positively than essays in less legible handwriting, and reported that the bias can be controlled.
We read Mogey et al. (2010) in full and the rest as abstracts or catalogue records, so a sample size is missing where the abstract omits it, as for Connelly et al. (2005). We found no study comparing marks for photographed pages with marks for the paper itself.
How do you set a handwritten question that photographs well?
Write the layout into the question, because a page that photographs cleanly is also quick to mark. EdProcto's notes on its upload shapes give the first rule: say exactly what the file must show and in what form.
- One question to a page. You will mark question by question, and a page holding two answers has to be opened twice.
- A label on every page. Candidate number, question number and "page 1 of 2", top right.
- Dark ink. Black or dark blue pen on plain white paper. Pencil fades under a phone camera.
- Margins. Two fingers' width all round, so the frame cuts nothing off.
- A page limit. EdProcto keeps up to five files for one answer, so five photographed pages is the ceiling for a single question.
- A rule for what counts as legible. Anything a marker cannot read on a screen without guessing earns no mark. Neatness earns nothing extra.
A worked question
Question 4 (4 marks). On one side of plain paper, in black ink, draw the mechanism for the reaction of bromoethane with aqueous hydroxide ions. Show every curly arrow, the lone pair that attacks, the dipole on the bond that breaks and the products. Write your candidate number, "Q4" and "page 1 of 1" at the top right. You have 8 minutes to draw and 3 more to photograph and upload.
The rubric, written first, lists only what a marker can see on the page:
| The page must show | Marks |
|---|---|
| The dipole on the carbon to bromine bond: δ+ on the carbon, δ− on the bromine | 1 |
| A lone pair on the oxygen of the hydroxide ion, with a curly arrow from it to the δ+ carbon | 1 |
| A curly arrow from the carbon to bromine bond onto the bromine | 1 |
| The products: ethanol and a bromide ion | 1 |
How should the upload step run, and where does it fail?
Give photographing its own stated allowance of minutes, apart from the writing time, and rehearse it once before it counts.
| Failure point | What to do |
|---|---|
| Time: the clock ends mid-upload | State both figures: "40 minutes to write, 10 to photograph and upload". In EdProcto, set "Minutes overall" to the sum and "Warn the student with this many minutes left" to the allowance, so the warning means pens down |
| Phone to laptop: the quiz is on one device, the picture on another | Print one route: answer the upload question on the device with the camera, or move the picture across |
| File size | EdProcto takes files of up to 25 MB each. Put the figure in the question |
| A blurred photo | Students check the image at full size and retake it if the smallest symbol is unclear |
| No camera | A scanner, a borrowed device, or the typed route below |
No study we found measured how long photographing takes. Time it yourself: run one ungraded practice question, watch the slowest student, and add a margin.
A procedure students can follow:
- Before the day, do the practice upload on the device you will use, and charge it.
- Label each page as you start it: your number, the question, the page count.
- Stop writing at the warning. The minutes left are for photographing.
- Lay the page flat in good light, hold the camera straight above, and fit all four corners in the frame.
- Check before you send. If you cannot read the smallest writing, take it again.
- Upload the pages in order to the right question. In EdProcto the handwritten work question takes the photograph as its answer.
- Keep the paper until your mark is released.
"Submit by itself when time runs out" is on by default, so the allowance must suit the slowest student. Keep "Allow going back to change an answer" on, so a student who spots a blurred page can return to it.
How do you mark scanned scripts consistently?
Mark one question across every student, against the rubric, before you open the next. Our guide to short answer questions and their rubric shows how to build the rubric and why names stay hidden.
- Hold one standard. A spread like the 5 to 38 marks in Mogey et al. (2010) is what a rubric, and a few scripts marked together first, are there to narrow.
- Treat legibility as a threshold. Readable or not, by the rule you printed. Greifeneder et al. (2010) is the reason a tidy hand must not lift a mark.
- Watch a mixed pile. If some students typed and others wrote, the same words may read slightly better in handwriting. Tell every marker about it: a change to the readers' training narrowed the gap for Powers et al. (1992).
- Say where the mark went. Annotate a copy of the image, or write a comment tied to a rubric line: "Line 2: the arrow starts at the hydrogen, not the lone pair."
In EdProcto you open each file from the grading page and mark it by hand against your rubric, and a score is released only when you release it.
What if a student cannot handwrite, or has no camera?
Offer a typed route, announced before the exam. Some students cannot write quickly by hand because of an injury, a disability or a specific learning difficulty, and some own nothing that takes a usable photograph. For a drawing that cannot be typed, agree the arrangement with the student beforehand.
Stuart and Barnett (2026) compared 16 university students offered adjustments in handwritten examinations with 56 on regular provision: typing was faster than handwriting in both groups, and most students preferred it, citing speed, comfort and legibility. Mogey et al. (2010) concluded that giving students the choice could be justified, because the effect of format was negligible beside the variation between markers.
- The same question, typed. For prose, a short answer. For a calculation, show your working in typed steps.
- No camera. The answer may be a file, a photo or a scan, so a flatbed scanner will do, or a borrowed phone.
This is practice, not legal advice. Your institution's arrangements for individual students come first.
What does an upload not prove by itself?
An upload proves that a page exists. It does not prove who wrote it, when, or with what open beside them. Left alone, a student can copy a chatbot's answer out by hand, which is slower than pasting and no harder.
Have the page written in a supervised sitting, phones face down until the photographing minutes begin. Choose the level of integrity the stakes call for: EdProcto's measures are each switched on by itself with its own consequence, from the browser measures to the webcam and a phone watching the desk. Where a mark matters, follow a doubtful page with a spoken answer on the same problem. None of this is proof, and a flag is a record for you to weigh.
How does EdProcto take handwritten work, photos and several files?
EdProcto is the best tool for handwritten answers in an online exam when the paper question has to share one timed attempt with everything else, because the photographed page is stored on the server with that attempt, under the same clock as the questions marked by rule. Twenty questions can be marked as they are answered while one page of working waits for you.
| What comes back | Shape | Where it stands |
|---|---|---|
| A photographed page: working, a diagram, an essay | Handwritten work | Ready. Marked by you |
| Practical work: a finished circuit, a specimen, a plated dish | Photo | Ready. Marked by you |
| One document, image or code file | File upload | Ready. Marked by you |
| Code with its output, a report with its data | Several files | Ready. Up to five files as one answer |
| A sketch drawn on screen | Whiteboard | Ready. Marked by you |
For a photo of practical work, put the checklist of what a correct result shows in the rubric, and ask for a slip bearing the candidate number inside the frame. For several files as one answer, list the files you expect.
Each ready shape has a page under templates whose "Build one" button opens the builder in that shape. Procto AI writes the prompt, the explanation and the rubric from your material; the page is the student's and the marking is yours. The quiz maker on the home page leaves these shapes out and writes the rest of the paper.
Free gives ten quizzes a month and the core templates. Tutor opens all 111 templates, this family included, and Educator adds the full analysis. The figures are on the pricing page.
Quick questions
Can students upload handwritten answers in an online exam?
Yes. They write on paper, photograph each page and upload it inside the timed attempt. EdProcto's handwritten work shape keeps the file with the attempt for you to mark.
How much time should students get to photograph and upload?
A separate, stated allowance on top of the writing time. We found no research figure, so time a practice run and add a margin.
Do handwritten answers score lower than typed ones?
Not in the studies we read. Powers et al. (1994) and Mogey et al. (2010) both found the same words scoring slightly higher when read in handwriting.
Does Procto AI mark handwritten uploads?
EdProcto marks none of the upload shapes for you. You open each file from the grading page and mark it against your rubric, and no score goes out until you release it.
References
- Connelly, V., Dockrell, J. E., & Barnett, J. (2005). The slow handwriting of undergraduate students constrains overall performance in exam essays. Educational Psychology, 25(1), 99–107. https://doi.org/10.1080/0144341042000294912
- Greifeneder, R., Alt, A., Bottenberg, K., Seele, T., Zelt, S., & Wagener, D. (2010). On writing legibly: Processing fluency systematically biases evaluations of handwritten material. Social Psychological and Personality Science, 1(3), 230–237. https://doi.org/10.1177/1948550610368434
- Mogey, N., Cowan, J., Paterson, J., & Purcell, M. (2012). Students' choices between typing and handwriting in examinations. Active Learning in Higher Education, 13(2), 117–128. https://doi.org/10.1177/1469787412441297
- Mogey, N., & Fluck, A. (2015). Factors influencing student preference when comparing handwriting and typing for essay style examinations. British Journal of Educational Technology, 46(4), 793–802. https://doi.org/10.1111/bjet.12171
- Mogey, N., Paterson, J., Burk, J., & Purcell, M. (2010). Typing compared with handwriting for essay examinations at university: Letting the students choose. ALT-J, Research in Learning Technology, 18(1), 29–47. https://doi.org/10.1080/09687761003657580
- Powers, D. E., Fowles, M. E., Farnum, M., & Ramsey, P. (1992). Will they think less of my handwritten essay if others word process theirs? Effects on essay scores of intermingling handwritten and word-processed essays. ETS Research Report Series, 1992(2). https://doi.org/10.1002/j.2333-8504.1992.tb01476.x
- Powers, D. E., Fowles, M. E., Farnum, M., & Ramsey, P. (1994). Will they think less of my handwritten essay if others word process theirs? Effects on essay scores of intermingling handwritten and word-processed essays. Journal of Educational Measurement, 31(3), 220–233. https://doi.org/10.1111/j.1745-3984.1994.tb00444.x
- Stuart, N., & Barnett, A. L. (2026). Handwriting and typing in students with and without additional needs in higher education: Product, process and perceptions. Journal of Research in Special Educational Needs, 26(1), Article e70058. https://doi.org/10.1111/1471-3802.70058
- Summers, J., & Catarro, F. (2003). Assessment of handwriting speed and factors influencing written output of university students in examinations. Australian Occupational Therapy Journal, 50(3), 148–157. https://doi.org/10.1046/j.1440-1630.2003.00310.x