Online exam software for universities: a buyer's guide

Home
StudyDrome Guides
University exam software

What should you compare first?

Most university exam software evaluations go wrong the same way. The demo looks fine, the contract gets signed, and the problems arrive at scale: accommodations that need a workaround, an item bank you cannot export, and statistics nobody can explain to an appeals panel. This guide covers the seven things that actually differ between systems, and how to test each one before you sign.

Start with the things that are expensive to fix later. Feature checklists are cheap. Migration, accessibility and defensibility are not.

Exam integrity is a policy choice, not a feature score.
Item-bank portability decides whether your next procurement is free or captive.
Accommodations must be a workflow, not a support ticket.
Statistics must be named, not implied.
Integration cost usually exceeds licence cost in year one.

What actually differs between systems?

Seven criteria separate real options. Everything else is table stakes.

Identity and supervision:

Some vendors record candidates and review the footage. Others verify identity before the exam. Others detect browser-level events and adjudicate server-side. These are different philosophies with different legal, ethical and false-positive profiles.

✔

Item banking and blueprinting:

Can items nest, carry typed tags, and map to a competency framework? Can you see coverage as a matrix before the paper is assembled?

✔

Accommodations:

Extra time, rest breaks, alternative formats. Ask to see the workflow, not a slide.

✔

LMS and SIS integration:

Rosters in, results out. Ask which direction is automated and which is a CSV.

✔

Offline and low-bandwidth behaviour:

What happens when a candidate's connection drops mid-exam? What happens when the room's Wi-Fi drops?

✔

Analytics and psychometrics:

Which coefficients are computed, and are they named? A vendor that will not name its reliability coefficient probably does not compute one.

✔

Data residency and retention:

Where do the servers sit, how long are audit logs kept, and can you choose?

✔

How do the categories compare?

Vendor feature grids go stale within a quarter. Categories do not. Shortlist by category first, then compare products inside it.

Criterion	LMS quiz tool	General exam platform	Proctoring-first vendor	Exam manager
Item banking	Basic, per course	Structured banks	Usually thin	Nested banks, typed tags
Blueprinting	No	Sometimes	No	Coverage matrix
Supervision	None	Browser controls	Webcam and identity checks	Browser controls, server-adjudicated
Psychometrics	Score reports only	Basic item stats	Rarely any	Named coefficients per exam
OSCE or practical	No	No	No	Stations and circuits
Accommodations	Per-quiz overrides	Per-candidate	Varies	Per-candidate, roster-level
Migration out	Course export	Varies	Varies	Bank export
Typical buyer	Individual faculty	Exam office	Compliance-driven	Assessment lead

Read the rows, not the totals. A department running low-stakes quizzes and a medical school running finals should not shortlist the same category.

How should you run the evaluation?

1
Write the exam you actually run. Not a generic requirement list. One real paper, one real cohort size, one real accommodation case.
2
Import your own items. Any vendor can demo their sample bank. Ask to load 200 of yours, with images, and see what survives.
3
Sit the exam as a candidate. On the oldest laptop in the building, on the room's real Wi-Fi.
4
Read the post-exam statistics. Ask which coefficient was used and why. If the answer is vague, the number is decorative.
5
Check the exit path. Ask for an export of everything: items, results, audit logs. In what format, and how long does it take? A vendor who cannot answer this quickly has told you something.

Who is in this market?

All figures below were checked on 2026-08-07 and each is the vendor's own published claim. Vendors change; re-check before you cite.

ExamSoft, part of Turnitin, delivers locked-down offline exams via its Examplify app and reports 2,100+ programs and 90M+ exams. It publishes WCAG 2.2 AA conformance and VPATs, and names no SOC 2 or ISO 27001 on any page fetched.
risr/ covers high-stakes medical exams and offers item response theory and Rasch modelling. It holds ISO 27001, SOC 2, Cyber Essentials Plus and CSA STAR. Its UK G-Cloud listing publishes £12 to £49.50 per user per year.
Speedwell has roughly three decades in medical education and owns the OSCE niche. It ships native iOS and Android examiner apps with true offline marking. It names no certification on its site.
MaxExam, from Maxinity, covers healthcare education with Ebel, Hofstee, Angoff and borderline regression. It holds Cyber Essentials and Cyber Essentials Plus. It publishes a 99.95% uptime guarantee, and lists £20 to £50 per licence per year on G-Cloud, with no free trial.
Cirrus targets higher-education written assessment and holds ISO 27001:2022 and WCAG 2.1.
ClassMarker is the self-serve option at $19.95 to $79.95 per month, with camera monitoring charged per attempt.

Note the units differ. Per user per year and per licence per year are not comparable to per tenant per month until you convert them against your own cohort size.

The procurement checklist

Ask these before the demo, and again before you sign.

Ask for the export formats and the time to produce a full export of items, results and audit logs.
Ask for the guaranteed availability level in writing.
Ask where the data centres are, and whether you can choose.
Ask what WCAG level the candidate interface meets, and to see the VPAT.
Ask how long audit logs are retained.
Ask what the offboarding process is, and what happens to your data after it.
Ask which reliability coefficient the system computes, and when each one applies.
Ask to watch an accommodations case set up end to end, not described on a slide.

Where does this buying model break down?

Four failure modes recur, and none of them show up in a demo.

Proctoring is not integrity. Surveillance detects some behaviours and misses others. It also generates false positives, raises candidate anxiety, and creates a data-protection obligation you now own. Decide what you are actually trying to prevent before you buy a camera.

Item-bank migration is the real switching cost. Licence fees are visible and negotiable. Re-authoring 4,000 items is neither. Test the export before the import.

Pilots are biased by design. A vendor-run pilot uses vendor-chosen items on vendor-chosen hardware. Insist on your items, your room, your worst laptop.

Accessibility is often retrofitted. A VPAT is a document, not a workflow. Ask a disabled student to sit the pilot exam.

Do you need a lockdown browser?

That depends on what you are defending against, and it is worth deciding on purpose rather than by default. A lockdown browser stops a candidate opening another application on the machine running the exam. It does nothing about a second device, a phone, or a person in the room.

We keep a full guide to Safe Exam Browser because many institutions do use one, usually via Moodle, Canvas or Blackboard. There is also a serious argument for the opposite approach, set out in exam integrity without webcam surveillance. Read both before you write the requirement.

Frequently asked questions

What is an exam management system?

An exam management system handles the whole exam lifecycle rather than just delivery: authoring and storing items in a bank, assembling papers, scheduling and rostering candidates, delivering the exam, marking it, and producing the statistics that justify the result. A quiz tool inside a learning platform typically covers only delivery and marking.

How do you keep an online exam secure?

Layer controls rather than relying on one. Access codes decide who can start. IP restriction decides where from. Device binding ties an attempt to the machine it started on. Browser controls block copy, paste and window-switching, and record violations server-side. Surveillance is a separate, optional layer with its own trade-offs, and it is not a prerequisite for a defensible exam.

How much does online exam software cost?

Published figures cluster around £12 to £50 per user per year for institutional vendors, and $19.95 to $79.95 per month for self-serve tools, all checked 2026-08-07. Most enterprise vendors publish nothing and quote per institution. UK buyers can often find real bands on the G-Cloud listings even when the vendor's own site is quote-only.

Can exam software integrate with our student record system?

Usually, but check the direction. Roster import is common. Automated results write-back into a student record system is much rarer, and is frequently a scheduled file transfer rather than a live API. Ask which specific fields sync, in which direction, and what happens when a grade is later amended.

What is item analysis, and why does it matter?

Item analysis measures how each question performed. Difficulty shows what proportion answered correctly. Discrimination shows whether strong candidates outperformed weak ones on that item. Distractor analysis shows whether the wrong options attracted anyone. Together they tell you which questions to keep, revise or retire, and they are the evidence an appeals panel expects.

How do you measure exam reliability?

Two coefficients dominate. KR-20 applies when every item is scored right or wrong. Cronbach's alpha applies when items carry partial credit. Both run from 0 to 1, and above 0.7 is generally treated as acceptable for classroom testing. Ask any vendor which one they compute and when each applies. A system that reports a number without naming the method is not giving you evidence.

Disclosure

This guide is published by StudyDrome, which sells a product in the exam-manager category described above. Everything in it is sourced and dated so you can check it independently. The category table is deliberately neutral, and the vendor figures are each vendor's own published claims.