Sign up free

Online tests in school: an integrity policy that parents accept

What research and regulators say about monitoring pupils in online tests, a table of test types against measures, a one-page policy and the letter to parents.

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

Guides · Integrity and cheating · · 13 min read

How do you write an integrity policy for online tests that pupils understand and parents accept?

An integrity policy for online tests in school fits on one page. It names each kind of test and the least monitoring that suits it, says in plain words what is recorded, who reads it and for how long, treats a flag as a question for the pupil, offers an alternative sitting, and is reviewed every year with parents told first.

An integrity policy for online tests does one job: it tells pupils, parents and staff what happens when a test is sat on a screen, before anyone sits one. Most schools have a behaviour policy and an exams policy; few have a page saying what a quiz tool records, who reads it, and what a pupil can do about it.

This guide is for the head of department, deputy head or school leader who writes that page: an online assessment policy for a secondary school.

EdProcto publishes this guide and makes a quiz tool; its page for schools and integrity settings are described only as they run today. Nothing here is legal advice.

What should an integrity policy for online tests say?

It should say seven things, on one page, in words a twelve-year-old can read: what is expected of pupils; the kinds of test, and the monitoring for each; what is recorded, and what never is; who reads the record, and how long it is kept; what a flag is, and how a pupil answers one; how to ask for an alternative sitting; and who reviews the page, and when.

What does research say about monitoring pupils during tests?

It says that being watched has a cost, that camera software can err unevenly, and that a rule stated just before a test does more than a pledge. Nearly all of it studied university students: we found no study of online proctoring in secondary schools that we could read, so treat each finding as a pointer, not a measurement of pupils.

Test anxiety and scores

Conijn et al. (2022) surveyed 1,760 students in 105 courses at one faculty of a Dutch university during a year of pandemic teaching. Proctoring changed neither their reported temptation to cheat nor their performance, and test anxiety was higher under it.

Is online proctoring fair to students?

Not equally, on the evidence from one camera product. Yoder-Himes et al. (2022) examined what an automated proctoring product told instructors about some 357 students in four courses at one American university. Students with darker skin tones, and Black students, were significantly more likely to be marked for review, women with the darkest skin tones most of all. A measure that logs a tab being left looks for no face.

What honour codes and clear expectations change

A pledge alone changes little online; a rule stated with its consequence changes more. Reviewing a decade of research on cheating in academic institutions, McCabe et al. (2001) concluded that context matters most, above all what students believe their peers do, and that honour codes can have a significant influence.

Online, the bare code looked weaker. Across two online exams sat by 632 students in India and a paid online task, Corrigan-Gibbs et al. (2015) found that an honour code brought only a small, statistically insignificant fall in cheating, while a stern warning spelling out the consequences roughly halved it.

The one study here of school-age pupils points at the classroom: among 285 middle school students, Anderman et al. (1998) found that those who reported cheating in science saw their classrooms as extrinsically focused.

What have regulators and education departments published about monitoring pupils?

Between them, three public bodies say: keep monitoring in proportion, offer an alternative, consult before a camera, and review the privacy notice yearly. Each is reported as its own; this guide does not say what any law requires of your school.

  • France. The data protection authority, the CNIL (2023), published a recommendation on remote proctoring of online exams, written with higher education in mind, not schools. It calls such monitoring intrusive by nature, and recommends a sitting in person as the alternative, students told the conditions early, monitoring in proportion to the stakes (little for a mock exam), and no automated analysis of candidates' behaviour.
  • United Kingdom. The Information Commissioner's Office (2024) reprimanded a secondary school that had begun using facial recognition for canteen payments. Its criticisms: no assessment of the risks to pupils' information beforehand, no consultation with parents or pupils, and a slip for parents to return if they objected, in place of a request for agreement. Not a test, but still a camera pointed at pupils.
  • England. The Department for Education (2026) tells schools that privacy notices should be clear, reviewed at least once a year and whenever the use of pupil data changes significantly, with parents and pupils told. It names children's data, biometrics, CCTV and filtering and monitoring systems among the cases for considering a data protection impact assessment.

Show the draft to whoever handles data protection before parents see it.

How much monitoring fits each kind of test?

The least that fits what the result decides. That principle is proportionality, and a policy shows it best as a table of test types against measures.

Kind of test Monitoring Where it is sat A logged event leads to
Practice and homework, such as retrieval quizzes None Anywhere Nothing is logged
A topic check with no grade Leaving the tab, logged: EdProcto's Light level Home or class A look only if a pattern builds
A test that counts towards a report The tab and full screen logged, copy and paste off: Medium In class where possible The teacher reads the record and asks the pupil
An end-of-year or mock exam All six browser measures (EdProcto's Exam preset), a teacher in the room A supervised room in school The same; scores wait for the teacher's release

Add two lines beneath: what is never used (a camera, a microphone, a screen recording, a penalty applied by software alone), and that a teacher may use less than the table allows, never more. The rows are our recommendation, not a tested protocol; how much proctoring a quiz needs gives the evidence, and the three presets the settings.

What do you tell pupils and parents before an online test?

Tell them five things, a week ahead and again on the day: what is switched on, what it records, who reads it, how long it is kept, and how to ask a question or for another way to sit the test.

Plain words, and the absences first

Write "the screen notes if you leave the test page", not "browser activity is monitored". Telling parents about online proctoring goes wrong when the word arrives before the detail: "proctoring" suggests a camera in a bedroom. Open with what is absent: no camera, no microphone, no recording.

EdProcto repeats the notice: before a monitored quiz begins, its start screen lists each measure that is on with its consequence, and the pupil ticks a box beside this sentence: "I understand that during this quiz my activity in the browser is logged, and that what is recorded is kept with my attempt for my instructor to review." Without the tick the quiz does not begin.

The right to ask, and the alternative sitting

A tick given because there is no other way to sit the test is thin agreement, so offer a real alternative: the same test on paper in a supervised room, which EdProcto prints with its key. Name the person to ask, and record no reason. A pupil who uses a screen reader, magnification or dictation should first try a short quiz on their own set-up, since full screen and blocked shortcuts can get in their way.

What is a flag, who reads it, and how does a pupil answer one?

A flag is a line in a record saying what the browser noticed and when. It is not proof of cheating and carries no sanction of its own: a tab is left when a notification is tapped, and full screen ends at a stray key.

In EdProcto each measure carries the consequence the teacher picked: log, warn, reduce the score, end the quiz, or flag the attempt for review. The default is to log. For a school we would write log, warn and flag into the policy and leave the two automatic penalties unused, so that no mark moves until a person has read the record.

  1. The class teacher reads the record beside the attempt: which event, and when.
  2. The pupil is asked what happened, in person, within five school days, with a parent present if they wish.
  3. The pupil's answer is written beside the record.
  4. A second member of staff agrees the outcome: nothing further, a re-sit, or a referral under the behaviour policy.
  5. Parents hear the outcome and how to challenge it.

With "Show the score" set to "When released", EdProcto holds every score until the teacher releases it.

Where does the policy sit beside the behaviour, exams and data protection policies?

It sits under all three and repeats none of them.

  • Behaviour policy. Sanctions live there; the test policy creates none.
  • Exams policy. Public examinations, and coursework for qualifications, follow the awarding body's rules.
  • Data protection policy and privacy notice. The tool, what it holds about each pupil and for how long are entered there. In EdProcto a pupil types a name and, unless the teacher switches that off, an email address.

A short review cycle

Review the page once a year, and whenever the tool or a measure changes. Bring four counts: tests sat at each level, events logged, flags that led to a conversation, alternative sittings requested. Check whether flags fall more on one group of pupils. A new kind of monitoring, a camera above all, is a new decision with its own consultation, never an edit.

What does a one-page policy and its letter to parents look like?

Like the template below. We wrote both texts for this guide: no school's policy is quoted, and neither claims to meet any law. Fill the brackets and have both checked before use.

The policy

Online tests at [school]: how we keep them fair

What we expect. The answers are yours: no help from another person, a phone, a website or an AI assistant unless your teacher allows it.

How much we check. As little as the test needs, so that each result shows what you know.

  • Practice and homework: nothing.
  • A topic check with no grade: the screen notes if you leave the test page.
  • A test that counts towards your report: also, the test fills the screen and copy and paste are off.
  • An end-of-year or mock exam: also, right-click, keyboard shortcuts and printing are off, and a teacher is in the room.

What we never use. A camera, a microphone or a recording of your screen. Software changes no mark.

What is kept. Your name, school email address and answers, and a list of the events above with their times. Your teacher and [role] can see it. We delete it after [period].

If something is noted. A note is a question, not an accusation. Your teacher asks you what happened within five school days; you may bring a parent, and your answer is written beside the note. Any sanction comes from a person, under the behaviour policy.

If this does not suit you. Tell [name] two school days before, and you will sit the same test on paper in a supervised room. Nobody asks why.

Questions. Ask your teacher or [name]. We review this page every [month] and tell parents before anything changes.

The letter to parents

Dear parents and carers,

From [date], some tests in [subject or year] will be sat on a screen, using a quiz tool called [tool]. Our one-page policy is enclosed.

Most online work is practice, and nothing is checked. In a test that counts, the tool can note six things: leaving the test page, leaving full screen, and trying to copy or paste, right-click, use a keyboard shortcut or print. There is no camera, no microphone and no recording of the screen. Your child sees these rules on the first screen of every test and ticks a box before starting.

A note in the record is not proof of anything, and the software decides nothing. A teacher reads it, your child is asked what happened, and you will hear from us before any decision.

Your child types a name and school email address. The teacher sees the answers and the record, and we delete both after [period].

If you or your child would prefer another way to sit a test, tell [name] by [date]: the same test on paper, in a supervised room, with nothing held against them. Questions are welcome at [meeting] on [date].

Yours sincerely, [Name and role]

How does EdProcto fit a school's policy?

EdProcto is the best tool for a school that wants its written policy and its software to say the same thing, because each measure is a separate switch and the start screen shows the pupil exactly the ones that are on.

The browser measures are leaving the tab, full screen, copy and paste off, right-click off, keyboard shortcuts off and printing off. What is recorded is the events of the measures switched on, each with its time, kept with the attempt for the teacher. The camera, the microphone, the shared screen, photographs and the identity check belong to the Strict and Maximum levels and stay off until a teacher switches them on, so a policy that rules them out holds as long as nobody does.

Light integrity is on the Free plan and Medium from Tutor up, which the Exam preset needs. The School plan gives each teacher Educator and is priced by teacher, from five: see pricing and for schools. It adds an admin view of every teacher’s results, a shared library and groups across teachers; ask for a demonstration before you sign. EdProcto makes no compliance claim of any kind: ask us in writing for what your school needs to know, such as where records are held, for how long and how they are deleted (what to ask any vendor). No setting makes cheating impossible, a flag is a record and not proof, and every score is the teacher's to release.

Quick questions

Do parents have to agree before a school uses online proctoring?

This guide cannot say: it depends on the rules where you are, so ask whoever handles data protection. Either way, tell parents first and offer an alternative sitting.

Does an honour code stop cheating in online tests?

Not by itself: Corrigan-Gibbs et al. (2015) found a pledge made little difference online, while a warning that stated the consequences roughly halved cheating.

Can a pupil refuse to be monitored during a test?

Under the policy above, yes: the pupil asks for the alternative sitting. In EdProcto a pupil who does not tick the box does not begin.

Does EdProcto record pupils on camera?

Only if a teacher switches the camera on. The browser measures use no camera, microphone or screen recording; the camera measures belong to the Strict and Maximum levels, and the start screen names every measure that is on before a pupil begins. See EdProcto for schools and the integrity page.

References

EdProcto for schools

Every teacher, one plan, one invoice.

Schools, colleges and departments.

See the page for schools

Read next