Normalisation of Marks: How Multi-Shift Exam Scores Are Made Comparable

Large recruitment exams cannot be held at one sitting. Lakhs of candidates sit the same computer-based test across days and shifts, each shift with a different question paper. However carefully those papers are set, one will turn out harder than another — and a candidate who drew the harder shift would be ranked lower for no fault of their own. Normalisation is the adjustment that corrects for this before the merit list is drawn up.

To see which recruitments are open to you now, start from 12th pass or jobs for graduates. The two exams this guide draws on are the SSC CGL and the RRB NTPC.

When It Applies, and When It Does Not

Normalisation is triggered by shifts, not by the exam. SSC's wording is conditional, and the condition is easy to miss: marks are normalised "if conducted in multiple shifts". A paper that everyone sits in a single shift has nothing to be compared against, so no adjustment is made at all.

The Railway Recruitment Boards give the same reason for doing it:

When an examination is conducted in multiple sessions for the same syllabus, in spite of all efforts there are chances of variation in the difficulty level of the questions in various sessions. Thus the score obtained by the candidates of same calibre is likely to vary.

So this is not a routine step applied to every recruitment. It is a repair for a problem that only exists once a paper is split across sittings.

What Normalisation Does — and Does Not Do

Normalisation converts each candidate's raw marks into a normalised score that can be compared across shifts. The Railway Recruitment Boards put the point bluntly in their published explanation: it "is a scientific and statistical process. It is not a process of awarding grace marks."

So normalisation does not add marks to everyone, and it does not reward a whole shift. It places each candidate relative to the people who sat the same paper, and then puts all shifts on one scale. Depending on how hard your shift was, your normalised score can come out higher or lower than your raw score — or, in the base session, the same.

It also comes after the answer key. Your raw score is fixed first, once objections to the provisional answer key are settled; normalisation works on that final raw score.

How the Railway Recruitment Boards Describe Theirs

The RRBs publish their method, and have revised it over the years, so the clause in your own Centralised Employment Notice is the one that applies. Their published explanation for CEN 01/2018 shows the logic in full.

• One shift is chosen as the base session: "a Session having highest average among all the Sessions of an exam with a condition that its population should be 70% or more than the average of all Session."
• Every other shift is rescaled to the base session's average and spread, by the formula the RRBs publish: Xn = (S2/S1)(X − Xav) + Yav, where X is your raw score, Xav and S1 are your shift's average and standard deviation, and Yav and S2 are the base session's.
• In the base session itself, normalised marks equal raw marks.

The same document shows how the raw score is worked out first, and it contains a detail worth knowing. Of 100 questions, 2 were "Ignored (on account of question wrong, multiple options correct etc)". A candidate who attempted 60, got 54 right and 6 wrong scored 54 positive marks less 2 negative — a third of a mark per wrong answer — for a net 52, and then (52 ÷ 98) × 100 = 53.06. A defective question is removed from the paper and everyone's score is computed over what remains; it is not awarded to all candidates.

That formula belongs to that recruitment. The RRBs revise the method between recruitments, so read the normalisation clause of the notice you applied under rather than assuming this one still holds.

Why SSC's Method Cannot Be Checked

SSC's position is stated in its exam notices rather than explained in them. The Combined Graduate Level notice says marks from a multi-shift computer-based exam "will be normalized by using the procedure published by the Commission vide Notice dated 02-06-2025 and such normalized scores will be used to determine final merit and cut-off marks." The Combined Higher Secondary Level notice repeats it, and adds that "Normalized scores of candidates will be used to determine merit and for final selection."

What the notices do not do is set out the procedure. They cite it by date. That matters for a practical reason:

• The 2 June 2025 notice is not reachable as a public document on the Commission's website at the time of writing — the notice board is served as a single-page application and the file is not indexed.
• The Commission does not publish your raw score, your shift's statistics, or the workings, alongside your result.

Treat the formulas circulating for SSC normalisation with care. They may well be right, but no candidate can check them against the notice they claim to describe, and a worked example you cannot verify is not worth planning around. The RRB document above is the better guide to how this kind of adjustment behaves, because it is actually published.

Percentile Is Not Percentage

Normalisation discussions lean heavily on percentile, and the two words get confused. A percentile is not the share of marks you scored. If you are at the 80th percentile of your shift, you scored as well as or better than 80% of the candidates in it — whether that took 40% of the marks in a hard paper or 70% in an easy one. A method built on percentiles compares positions, not marks.

How to Tell Your Score Has Been Adjusted

You can settle this yourself, with arithmetic rather than a formula.

Work out what a raw CGL score can arithmetically be. Tier-I is four subjects of 25 questions and 50 marks each — 100 questions for 200 marks, so two marks a question — with "negative marking of 0.50 marks for each wrong answer". Every raw Tier-I total therefore lands on a whole or half mark. In Tier-II, Paper-I deducts "1 mark for each wrong answer in Section-I, Section-II and Section-III", so the same is true there.

Now look at what gets published. Cut-off and score figures in the CGL 2024 final result include:

No combination of right and wrong answers under those marking schemes produces five decimal places. A figure of that shape is an adjusted score and nothing else — which answers the question without needing to know the method.

You Can Challenge a Key, Not a Score

There is a window to contest the inputs, and it closes early.

Tentative answer keys are published after the exam and can be challenged on payment of a non-refundable fee per question — ₹100 for CGL, ₹50 for CHSL. A key that gets corrected changes raw marks for everyone who answered that question, and normalised scores are computed afterwards.

The score itself is final. Both notices state: "There shall be no provision for re-evaluation/ re-checking of the scores of any stage/ Tier(s) of the examination." So the arithmetic cannot be appealed, and since the shift statistics are never released, it cannot be independently recomputed either.

What It Means for You

• Cut-offs are on normalised scores. A raw score you counted from the answer key is not directly comparable with a published cut-off.
• Normalisation does not lift you over the qualifying floor. That minimum is a separate rule fixed in the notice — in CHSL, 30% for UR, 25% for OBC and EWS, 20% for all other categories.
• A "hard shift" is not a disadvantage in itself. If everyone found it hard, your position among them is what counts.
• Your score depends on everyone else in your shift, which no candidate can see.

The cut-off guide explains how normalised scores become cut-offs, and the merit list guide how they become a ranking. For where this sits in the wider sequence, see the recruitment process guide. To work out your raw score from a marking scheme first, use the exam score calculator.

Frequently Asked Questions

Is normalisation the same as grace marks?

No. The RRBs describe it as a statistical process that is "not a process of awarding grace marks". It adjusts for differences in paper difficulty between shifts; it does not add marks across the board.

Is normalisation applied when an exam is held in a single shift?

No. SSC's notices make it conditional on the examination being "conducted in multiple shifts". With one shift there is nothing to compare against.

Can my normalised score be lower than my raw score?

Yes. If your shift's paper was easier than the base session's, raw marks are adjusted down to be comparable with candidates who sat harder papers — and up if your shift was harder.

Are cut-offs based on raw or normalised marks?

On normalised marks, where an exam ran in several shifts. SSC's notices state that normalised scores determine final merit and cut-off marks.

Can I calculate or challenge my exact normalised score?

Neither, in practice. It depends on the whole distribution of scores across shifts, which only the recruiter holds, and both notices rule out re-evaluation or re-checking of scores. What you can challenge, within the time limit and on payment of the per-question fee, is a tentative answer key.

Why does my result show marks with several decimal places?

Because it is a normalised score rather than a raw one. Under the published marking schemes a raw total can only fall on a whole or half mark, so decimals of that kind are the sign an adjustment has been applied.

What happens to questions that were dropped from the answer key?

In the RRB method above, a defective question is removed from the paper and your score is computed over the remaining questions — not awarded to everyone.

In Summary

Normalisation makes scores from different shifts comparable, so that drawing a harder paper does not cost you a rank, and it applies only when an exam actually runs in multiple shifts. It is not grace marks, and it can move a score down as well as up. The RRBs publish their formula and a worked example; SSC cites its procedure by notice date without publishing it, so treat third-party SSC formulas as unverifiable. You can tell an adjustment has been made because published figures carry decimals a raw score cannot produce — and at that point the key, not the score, is the only thing still open to challenge.

Sources

• Staff Selection Commission, CGL Examination 2026 notice (21 May 2026) — paras 13.2, 13.3, 13.7, 13.8 and 13.9.8
• Staff Selection Commission, CHSL Examination 2026 notice (7 September 2026) — paras 13.2, 13.3, 13.7, 17.1 and 17.2
• Railway Recruitment Boards, "Normalization of Marks in CBT2 under CEN 01/2018: A Scientific and Statistical Process"
• Staff Selection Commission, CGL 2024 final result write-up — the published cut-off lists