How Do Online Exams Work? The Five Stages, and What Breaks in Indian Conditions
Online exams from question bank to published result - how each stage works, why heavy proctoring fails the students with the weakest devices, and the connection-failure rule to decide before exam day.
How an Online Exam Actually Runs
An online examination replaces the answer script with a screen, but the surrounding process is largely the same as a paper exam. Understanding the sequence makes the difference between a smooth exam day and a stressful one.
The sequence has five stages: build the question paper, schedule and allocate, conduct with supervision, evaluate, publish.
Stage 1 — Building the Paper
Questions are created in a bank rather than a document, which is the first significant change from paper. A question bank means each question is stored once with its subject, chapter, type, difficulty and marks attached, and papers are assembled from it.
The practical benefit arrives in the second year. A bank built over one session generates papers in minutes thereafter, and the same question can appear in a practice test, a unit test and a revision set without being retyped.
Question types typically supported:
- Objective — multiple choice, true/false, matching, fill in the blank. Marked automatically.
- Short answer — a few words or a number. Auto-markable with a tolerance for spelling.
- Descriptive — paragraphs, derivations, diagrams. Requires a teacher.
Most school exams are a mix, which means most online exams are partly auto-marked and partly manual. Any system claiming to eliminate teacher evaluation entirely is describing an objective test, not a school examination.
Stage 2 — Scheduling and Allocation
The paper is assigned to classes or sections, with a date, a start time, a duration and a window. Two settings matter more than people expect:
Question randomisation. Each student receives the same questions in a different order, or a different subset drawn from the bank. This is the single most effective anti-copying measure and costs nothing.
The window versus the duration. A 60-minute exam with a 90-minute window lets students who face a power cut or a connection drop still complete it. A window equal to the duration guarantees complaints.
Stage 3 — Conducting It
Students log in, the paper opens at the start time, and a timer runs. Good implementations save answers continuously rather than at submission, so a browser crash at minute 55 does not destroy the attempt.
What supervision realistically achieves. Full-tab lock, camera proctoring and screen recording exist, and for a school unit test they are usually disproportionate. They also fail in exactly the circumstances Indian students most often face — a shared device, an unreliable connection, no webcam.
Proportionate measures for a school exam: randomised order, a per-question timer for objective sections, a single-attempt rule, and logging of tab switches for review rather than automatic disqualification. Heavy proctoring belongs to high-stakes examinations, not weekly tests.
Plan for failure. The connection will drop for some students. The rules should say in advance what happens — resume where they left off, extend by the lost time, or reschedule — so it is not decided under pressure with a parent on the phone.
Stage 4 — Evaluation
Objective sections are marked on submission. Descriptive answers go into a queue for teachers, and this is where the real time is spent.
Two things make it faster: marking one question across all students rather than one script at a time, which keeps the mark scheme loaded and makes marking more consistent; and a rubric attached to the question so partial marks are applied the same way by every teacher.
Stage 5 — Publishing
Results release to students and parents, usually with the marks, the correct answers, and a per-topic breakdown showing where the class as a whole struggled.
That last item is the part schools most often ignore and should not. A report showing that 70% of the class failed the same question is a teaching instruction, not a statistic.
What Online Exams Are Genuinely Good For
- Practice and revision — unlimited attempts, immediate feedback, no correction load
- Objective unit tests — marked instantly, results the same day
- Diagnostic tests — the per-topic breakdown tells you what to re-teach
- Make-up tests for absent students, without setting a separate paper
They are weakest where handwriting, diagrams or long derivations matter, and for board-pattern practice, where students need the experience of writing at length on paper.
Questions Schools Should Ask
- What happens when a student's connection drops at minute 40?
- Are answers saved continuously or only on submission?
- Can the paper mix auto-marked and teacher-marked sections?
- Does it work on a low-end Android phone, which is what many students have?
- Can teachers mark question-by-question across students?
- Does the result carry a per-topic breakdown, or only a total?
How It Works in EdPayU
The exams module builds papers from a question bank, randomises order per student, saves answers continuously, auto-marks objective sections, and queues descriptive answers for question-wise evaluation. Results publish to parents on WhatsApp with the per-topic breakdown, and feed the analytics that show which chapters need re-teaching. Our guide on report card formats covers how the marks reach the report card.
The Short Version
Build from a question bank, randomise the order, give a window wider than the duration, save answers continuously, and decide the connection-failure rule before exam day. Keep proctoring proportionate — heavy surveillance fails the students with the weakest devices. And read the per-topic breakdown; it is the part that changes teaching.