Statistics, in the SSC CGL context, is not the deep academic discipline — it is a focused set of tools for describing a dataset with a single number or a small collection of numbers. The questions test whether you can extract information from grouped or ungrouped data quickly and accurately.
Think of it this way: you have 200 children's ages scattered across a table. Nobody wants to read 200 values. Statistics gives you summary weapons:
Here is the analogy that makes this stick. Imagine a class of students. The mean is the "fair share" — if you pooled everyone's marks and redistributed equally, what would each person get? The median is the "middle child" — the person sitting exactly in the centre when everyone lines up by marks. The mode is the "most popular score" — the value that appears most often, the one everyone seems to have converged on.
SSC CGL leans heavily on mean-based questions — consecutive number averages, weighted averages, and the "sum equals mean times count" identity. Mode and median appear in frequency distribution tables. Standard deviation and variance appear occasionally, but the formulas are rarely needed in raw form — pattern recognition and elimination handle most such questions.
The key insight you should carry into every question: mean is always about the total sum. Every clever trick in this chapter is ultimately a manipulation of Sum = Mean × Count.
For ungrouped data with values x₁, x₂, ..., xₙ:
For grouped data with frequencies f₁, f₂, ..., fₖ and corresponding values x₁, x₂, ..., xₖ:
The single most powerful rewrite: Total Sum = Mean × n. You will use this in almost every SSC CGL statistics question.
Mean of first n natural numbers: Sum of first n naturals is n(n+1)/2. Dividing by n gives (n+1)/2. This is a direct formula — no derivation needed in the exam hall.
Mean of consecutive even/odd numbers: For any sequence of consecutive even numbers (or odd numbers), the mean equals the average of the first and last term, which is also the average of any symmetric pair around the centre.
Arrange data in ascending order. If n is odd, median = value at position (n+1)/2. If n is even, median = average of values at positions n/2 and n/2 + 1.
For grouped frequency data, the median class is the class where the cumulative frequency first crosses n/2. You then apply interpolation — but this is rarely tested in SSC CGL in its full form. More common is the direct "find the middle value" style question.
Mode is the value (or class) with the highest frequency. In a symmetric distribution (like a normal bell curve), mean = median = mode. This symmetry property is tested directly — if a question gives you a symmetric frequency distribution, you can equate the three.
For grouped data, the modal class is the class with the highest frequency. The exact mode formula uses the frequencies of the modal class and its neighbours — but again, SSC CGL almost never asks you to compute this. What it does test is whether you can identify the modal class and compare it with the computed mean.
Variance (σ²) measures the average squared deviation from the mean:
Standard deviation σ = √(variance).
Key properties you must know for SSC CGL:
k, the mean increases by k but the standard deviation does not change. Adding a constant shifts the data, it does not spread it.k, the standard deviation gets multiplied by |k| and the variance gets multiplied by k².A and B, if set A = {2, 2, 2} and set B = {1, 2, 3}, set A has σ = 0 and set B has σ > 0, even though both have mean = 2.A large chunk of SSC CGL statistics questions involve arithmetic sequences — consecutive integers, consecutive even numbers, consecutive odd numbers. Here is what you must know cold:
a + (k-1)/2a + 2(m-1)Look — when the question says "8 consecutive even numbers and the average of the 4th and 5th is 41", you do not need to write out the series. The 4th and 5th terms are symmetric about the mean of the series, so their average IS the mean of the entire series = 41. But be careful: the mean of an even-count series sits between two terms, so here the 4th = 40 and 5th = 42, and you reconstruct from there.
When combining groups: Combined Mean = (n₁x̄₁ + n₂x̄₂) / (n₁ + n₂). This generalises to any number of groups. SSC CGL loves multi-group average problems where you find a missing value — always convert to sums first.
Before touching any average/mean question, rewrite it as: Total Sum = Mean × Count. Every piece of information in the question translates to a sum. Then add/subtract sums to find the unknown sum, and divide by the count at the end.
Example: "15 numbers, mean 80" → Total = 1200. "First 6, mean 72" → Sum of first 6 = 432. You never need to hold fractions mid-calculation. Standard method (working with averages throughout): ~60s and prone to fraction errors. This method: ~20s, integer arithmetic only.
For any arithmetic sequence (consecutive integers, evens, odds), the average of the kth term from the start and the kth term from the end equals the mean of the entire series. So in an 8-term series, avg(1st, 8th) = avg(2nd, 7th) = avg(3rd, 6th) = avg(4th, 5th) = overall mean.
Use this when the question gives you the average of a symmetric pair — immediately write down the series mean, then reconstruct the first term. This collapses a 4-step problem into 2 steps. Standard reconstruction: 5 steps. Symmetric pair insight: 2 steps.
If a frequency distribution table is symmetric (frequencies increase to a peak and then decrease in mirror image), then mean = median = mode. You can compute the easiest one and assign the same value to the rest.
Example: frequencies 20, 50, 70, 50, 20 for values 1, 2, 3, 4, 5 are perfectly symmetric around value 3. Mode = 3 by inspection (peak frequency). Mean = 3 by symmetry. No calculation needed. Standard mean calculation: 4 multiplications + 1 division = ~30s. Symmetry inspection: ~5s.
If values are large (e.g., 198, 200, 202, 204), subtract a base value (say 200) from each: new values are −2, 0, 2, 4. Mean of new values = 1. Add back the base: actual mean = 201.
This is the "assumed mean" or "step deviation" method. For variance, the shift does not change σ² at all — compute σ² on the simplified values directly. Standard calculation on 3-digit numbers: ~50s. After shifting: ~15s.
If every value is increased by k: new mean = old mean + k, but new SD = old SD. If every value is multiplied by k: new mean = old mean × k, new SD = old SD × |k|.
In MCQ format, this eliminates 3 options immediately. The question will typically give you the original SD and ask for the new one after a transformation. Identify the transformation type, apply the rule, done. No calculation needed — pure rule application. Time: ~10s vs ~90s for full recalculation.
When you see a statistics question in the exam hall, run this decision tree:
Step 1 — What is being asked?
Step 2 — Mean question type:
(n+1)/2, zero calculationStep 3 — Always verify by back-substituting into the original condition. On a 3-variable problem (like 13th, 14th, 15th number), assign a variable to the one most constraints are written in terms of, write one equation, solve linearly. Do not set up three equations — that is exam-hall time you do not have.
Budget: straightforward mean question — under 60 seconds. Multi-group complex average — under 90 seconds. If you are past 2 minutes, eliminate and move on.
Why this question: Tests the symmetric pair property of arithmetic sequences — the single most common statistics pattern in SSC CGL.
Solving path: The 4th and 5th numbers are the two middle terms of an 8-term series. Their average = 41 = mean of the series. Since these are consecutive even numbers, the 4th term = 40 and 5th = 42 (they straddle the mean). Count back: 3rd = 38, 2nd = 36, 1st = 34. Count forward: 6th = 44, 7th = 46, 8th = 48. Sum of smallest and largest = 34 + 48 = 82.
Why this question: Tests the multi-group weighted average setup — three interlinked conditions, one equation. Classic SSC CGL complexity.
Solving path: Total sum of 15 numbers = 80 × 15 = 1200. Sum of first 6 = 72 × 6 = 432. Next 6 average = 72 × 1.25 = 90, so sum = 90 × 6 = 540. Sum of last 3 (13th, 14th, 15th) = 1200 − 432 − 540 = 228. Let 15th number = x. Then 14th = x − 10 and 13th = x + 8. So (x + 8) + (x − 10) + x = 228 → 3x + (−2) = 228 → 3x = 230 → x ≈ 76.67. Then 14th = 66.67, 13th = 84.67. Average of 13th and 14th = (84.67 + 66.67)/2 = 151.34/2 = 75.67.
Why this question: Tests whether you can identify a symmetric frequency distribution and equate mode with mean — zero computation needed if you see the symmetry.
Solving path: Frequencies are 20, 50, 70, 50, 20 — perfectly symmetric around age 3. Mode = 3 years (highest frequency). Mean = (20×1 + 50×2 + 70×3 + 50×4 + 20×5) / (20+50+70+50+20) = (20 + 100 + 210 + 200 + 100) / 210 = 630 / 210 = 3. Difference = 3 − 3 = 0. If you spotted the symmetry first, the mean computation was just a confirmation — you already knew the answer.
Why this question: A direct formula recall question. No computation — just knowing (n+1)/2 and being able to derive it under pressure.
Solving path: Sum of first n naturals = n(n+1)/2. Arithmetic mean = Sum / n = (n+1)/2. Check with n = 4: numbers are 1, 2, 3, 4, mean = 10/4 = 2.5 = (4+1)/2 = 2.5. Confirmed. On the options, eliminate n/2 (off by a half), n(n+1)/2 (that is the sum, not the mean), (n−1)/2 (would give 0 for n=1, wrong). Only (n+1)/2 survives.
Confusing "mean of a series" with "mean of a pair within the series." In consecutive number questions, the average of the two middle terms equals the series mean — but only when the series has an even number of terms. For odd-term series, the single middle term is the mean. Do not mix these up.
Forgetting that 25% more than 72 is 90, not 97. "25% more" means × 1.25, so 72 × 1.25 = 90. A surprising number of errors in multi-group average questions come from this one arithmetic slip. Write it out explicitly: 72 + 0.25×72 = 72 + 18 = 90.
Adding constants to standard deviation. If every observation increases by 5, the mean increases by 5 but SD stays the same. Many test-takers add 5 to SD as well. The rule is: SD measures spread, not location — shifts do not change spread.
Writing mode as the frequency, not the value. Mode is the value that appears most frequently, not the frequency count itself. In a table with value = 3, frequency = 70, the mode is 3, not 70.
Using n/2 for the mean of first n naturals. This is the most common wrong answer on that template question. The correct formula is (n+1)/2. Verify with n = 1: mean of {1} = 1 = (1+1)/2 = 1. The formula n/2 gives 0.5 for n = 1, which is obviously wrong.
Setting up three separate equations for three unknowns when one substitution suffices. In the "13th, 14th, 15th number" type question, all three unknowns are defined in terms of one (the 15th). Write one equation in one variable. Students who set up a 3×3 system waste 2+ minutes on algebra that a single substitution solves in 30 seconds.