SSC CGL Statistics Statistical Inference Questions 2026 (Estimation, MLE, Unbiased Estimator, Confidence Interval, Type I & II Errors, Z, t, Chi-square Tests MCQ, आकलन एवं परिकल्पना परीक्षण) — Free Chapter Wise Test in Hindi & English

A 95% confidence interval for the mean, the MLE of p, the z-value for a test of a proportion, degrees of freedom for a 3 × 4 contingency table — SSC asks estimation and testing as formula numericals and one-line concepts. Practise 75 questions in 3 tests covering point and interval estimation, unbiasedness, consistency, efficiency and sufficiency, maximum likelihood and the method of moments, confidence intervals, hypotheses, Type I and II errors, power and p-values, large-sample tests, t, F and chi-square tests, and non-parametric tests, written for SSC CGL Tier II Paper II (Statistics) for the Junior Statistical Officer post. Every test is in English and Hindi, and every question has a 6–7 point bullet solution.

3 Tests 75 Questions 25 Questions per Test English & हिंदी

Statistical Inference Tests — Pick a Test and Start

Tap below to see all the tests. They run through the chapter in order — Test 1 covers point and interval estimation, properties of estimators, maximum likelihood, method of moments, Cramér–Rao and Rao–Blackwell; Test 2 covers hypotheses, level of significance, Type I and II errors, power, p-values and large-sample tests for means and proportions; Test 3 covers t-tests, the F-test, chi-square tests of goodness of fit, independence and variance, and non-parametric tests. Choose English or हिंदी on any test.

Statistical Inference All Chapter Tests Checking… View tests

Set at real SSC CGL Tier II Paper II level. Each test mixes the ways SSC asks statistics questions — direct one-line questions, short numericals, statement questions, match the following and assertion–reason — and no fact is asked twice across the chapter.

Loading tests…

Statistical Inference is the “estimation and testing” chapter of the Statistics paper. SSC asks confidence intervals and sample size, properties of estimators and MLEs, Type I and II errors and power, z, t, F and chi-square test statistics, degrees of freedom and when to use which test.

Marks are lost in three places. The first is Type I vs Type II error: rejecting a true H₀ is Type I (probability α); accepting a false H₀ is Type II (β), and power = 1 − β. The second is degrees of freedom: (r − 1)(c − 1) for a contingency table, n − 1 for a one-sample t-test, n₁ + n₂ − 2 for two samples. The third is choosing the test: paired t for before–after data, F for comparing two variances. The tests here are built around exactly these three.

How Many Statistical Inference Questions Are Asked in SSC CGL?

Statistics is Paper II of SSC CGL Tier II, taken only by candidates for the Junior Statistical Officer (JSO) post: 100 questions for 200 marks in 2 hours, with 0.5 marks deducted for each wrong answer. Statistical Inference is one of the 5 Probability, Sampling & Inference chapters of the 12 on this site. SSC does not publish a chapter-wise breakup; the topics below are the ones the JSO papers return to most often, roughly in order.

TopicHow oftenWhat is usually asked
Tests of hypotheses — conceptsVery frequentH₀ and H₁, α and β, power, critical region, one- and two-tailed tests, p-value
Large-sample testsVery frequentz-tests for a mean, a proportion and differences, critical values 1.96 and 2.58
Small-sample testsFrequentOne-sample, two-sample and paired t-tests, F-test, t-test for r
Chi-square testsFrequentGoodness of fit, independence, Yates’ correction, degrees of freedom
Point estimationFrequentUnbiasedness, consistency, efficiency, sufficiency, MLE and its properties
Interval estimation and non-parametric testsRegularConfidence intervals, sample size, sign, run and Mann–Whitney tests

The sampling distributions behind these tests are in Sampling Theory; testing several means at once is in Analysis of Variance.

What Each Statistical Inference Test Covers

The 75 questions run through the chapter in syllabus order, 25 to a test.

TestTopics
Test 1Confidence Interval for Mean (σ known), Width of Confidence Interval, Sample Size and Width of CI, Sample Size for Given Margin of Error, MLE of Binomial p, Invariance Property of MLE, MLE of Normal Variance, Method of Moments, Unbiased Estimator, Relative Efficiency, Asymptotic Unbiasedness, Consistency of an Estimator, Sufficiency, Sufficient Statistic (Bernoulli), Cramér–Rao Inequality, Minimum Variance Unbiased Estimator, Rao–Blackwell Theorem, Mean Squared Error, Properties of MLE, Unbiasedness vs Consistency, Properties of Confidence Interval, Terms in Estimation, Methods of Estimation and Originators, Efficiency of Mean vs Median, Unbiasedness under Transformation
Test 2Level of Significance, Large-Sample Test for a Mean, Type II Error, Large-Sample Test for a Mean (σ unknown), One-tailed and Two-tailed Tests, Test for Difference of Two Means, Simple and Composite Hypotheses, Large-Sample Test for a Proportion, Two-tailed Critical Region, Test for Difference of Two Proportions, Neyman–Pearson Lemma, Standard Errors in Large-Sample Tests, Size of a Test (Binomial), Errors and Sample Size, Critical Values of Z, Probability of Type II Error, p-value, Power of a Test, Large-Sample Tests — Assumptions, Critical Region for a Mean, α, β and Power, Decision Using p-value, p-value Calculation, Interpretation of Rejection, Critical Region
Test 3One-sample t-test Statistic, Choosing the Paired t-test, Paired t-test Statistic, Assumptions of Two-sample t-test, Pooled Variance Estimate, t-test for Correlation Coefficient, t and F Tests — Properties, F-test for Equality of Variances, Degrees of Freedom of Small-sample Tests, χ² Goodness of Fit Statistic, Conditions for χ² Test, d.f. in Goodness of Fit, d.f. for χ² Test of Independence, Yates' Correction in 2×2 Table, Reason for Yates' Correction, χ² Statistic — Properties, χ² Test for Population Variance, Sign Test, Run Test, Mann–Whitney U Statistic, Non-parametric vs Parametric Tests, Median Test, Kruskal–Wallis Test, Non-parametric Tests — Properties, Mann–Whitney U Test — Assumptions

The three traps that cost the most marks

  • Type I or Type II? Type I = rejecting H₀ when it is true (α); Type II = accepting H₀ when it is false (β); power = 1 − β.
  • Degrees of freedom. Contingency table (r − 1)(c − 1); goodness of fit k − 1 − (parameters estimated); paired t n − 1; two-sample t n₁ + n₂ − 2.
  • Unbiased sample variance. Divisor n − 1 gives an unbiased estimator of σ²; divisor n is biased but consistent.

Solved Examples — SSC CGL Statistics Statistical Inference

Q1. An estimator \(T_{n}\) of θ based on a sample of size n has E(\(T_{n}\)) = θ + 2/n. Then \(T_{n}\) is:
(a) Unbiased for every n
(b) Biased, but asymptotically unbiased
(c) Biased, and not asymptotically unbiased
(d) Unbiased only when n = 2
Tₙ is biased, but asymptotically unbiased. Bias = E(Tₙ) − θ = 2/n, which is non-zero for every finite n, so Tₙ is biased. As n → ∞, 2/n → 0, so E(Tₙ) → θ; this is the definition of asymptotic unbiasedness. 'Unbiased for every n' would need E(Tₙ) = θ exactly for all n. Answer: (b)
Q2. A random sample of size 100 from a population with known standard deviation σ = 10 gives sample mean x̄ = 50. The 95% confidence interval for the population mean μ is (use \(z_{0.025}\) = 1.96):
(a) (30.40, 69.60)
(b) (47.42, 52.58)
(c) (48.04, 51.96)
(d) (49.804, 50.196)
The 95% confidence interval is (48.04, 51.96). For known σ, the 95% confidence interval for μ is \(\bar{x} \pm 1.96\,\frac{\sigma}{\sqrt{n}}\). Standard error = σ/√n = 10/√100 = 10/10 = 1. Margin of error = 1.96 × 1 = 1.96, so the limits are 50 − 1.96 = 48.04 and 50 + 1.96 = 51.96. (30.40, 69.60) uses σ instead of σ/√n, i.e. 50 ± 1.96 × 10. Answer: (c)
Q3. Read the Assertion (A) and Reason (R) carefully.
Assertion (A): For a large sample from a normal population, the sample mean is a more efficient estimator of μ than the sample median.
Reason (R): For large n, the variance of the sample median is approximately \(\frac{\pi\sigma^2}{2n}\), which exceeds the variance \(\sigma^{2}\)/n of the sample mean.
(a) Both A and R are true, and R is the correct explanation of A
(b) Both A and R are true, but R is not the correct explanation of A
(c) A is true, but R is false
(d) A is false, but R is true
Both A and R are true, and R is the correct explanation of A. For a normal population, both the sample mean and the sample median are unbiased estimators of μ, by symmetry. Var(x̄) = σ²/n, while for large n Var(median) ≈ πσ²/(2n) ≈ 1.57σ²/n. Since π/2 > 1, the median has the larger variance, so the mean is more efficient, exactly as A states. Answer: (a)
Q4. k स्वतंत्र प्रतिदर्शों के लिए, जिनके आकार पर्याप्त बड़े हों, शून्य परिकल्पना के अंतर्गत क्रुस्कल–वालिस प्रतिदर्शज H लगभग किस बंटन का अनुसरण करता है?
(a) k − 1 स्वातंत्र्य कोटि वाला \(\chi^{2}\)
(b) k स्वातंत्र्य कोटि वाला \(\chi^{2}\)
(c) (k − 1, N − k) स्वातंत्र्य कोटि वाला F
(d) N − k स्वातंत्र्य कोटि वाला t
सही उत्तर k − 1 स्वातंत्र्य कोटि वाला χ² है। सभी N प्रेक्षणों को एक साथ कोटि दी जाती है और Rᵢ, nᵢ आकार के i-वें प्रतिदर्श का कोटि योग है। H = \(\frac{12}{N(N+1)}\sum \frac{R_i^2}{n_i} - 3(N+1)\)। H₀ (सभी k समष्टियाँ समान) के अंतर्गत, जब प्रत्येक nᵢ बहुत छोटा न हो, H लगभग k − 1 स्वातंत्र्य कोटि वाला χ² होता है। उत्तर: (a)

How the Statistical Inference Tests Work

  • 25 questions per test, with a full question palette — answered, skipped, marked for review, not visited.
  • 3 tests, 75 questions, in syllabus order — point and interval estimation, unbiasedness, consistency, efficiency and sufficiency, maximum likelihood and the method of moments, confidence intervals, hypotheses, Type I and II errors, power and p-values, large-sample tests, t, F and chi-square tests, and non-parametric tests, with no fact repeated across the chapter.
  • All SSC question types — direct questions, short numericals, statement questions, match the following and assertion–reason.
  • Optional 15-minute timer — for quick timed practice; the real Paper II gives 2 hours for 100 questions.
  • Marking: +2 correct, −0.5 wrong, 0 unattempted — the same as Paper II.
  • 6–7 point bullet solutions open after you submit — the answer, the rule or fact behind it, related facts asked alongside it and what each wrong option actually refers to.
  • English and Hindi — every test and every solution is available in both; choose the language when you start.
  • Free with a TrickySSC account — sign in once with Google or your mobile number, and every test is open to you.

How to Prepare Statistical Inference for SSC CGL JSO

  1. Learn the critical z-values. 1.645, 1.96 and 2.58 answer most large-sample questions.
  2. Write the test statistic before the numbers. (estimate − hypothesised value) ÷ SE is the pattern for every z and t test.
  3. Make an estimator card. One line each for unbiased, consistent, efficient and sufficient, with a standard example.
  4. Keep a “which test” list. Mean, proportion, two means, paired data, variances, independence.
  5. Read the solutions, not just the answer. Each solution carries six or seven related facts and formulas, which is how 75 questions cover the whole chapter.

All SSC CGL Chapter Wise Tests

Every chapter below has its own free chapter tests with full solutions.

Browse all 131 chapter tests — Maths (35) · Reasoning (28) · English (14) · GK (18) · Computer (7) · Statistics (12) · Finance & Economics (17)
English Language
General Intelligence & Reasoning
Quantitative Aptitude (Maths)
General Awareness
Computer Knowledge
Statistics (JSO)
Finance & Economics (AAO)

SSC CGL Statistics Statistical Inference — Frequently Asked Questions

How many Statistical Inference questions are asked in SSC CGL?
Statistics is Paper II of SSC CGL Tier II, taken only by candidates for the Junior Statistical Officer (JSO) post: 100 questions for 200 marks in 2 hours, with 0.5 marks deducted for each wrong answer. SSC does not publish a chapter-wise breakup; Statistical Inference is one of the 12 chapters of the paper here, in the Probability, Sampling & Inference group.
Which Statistical Inference topics are most important for SSC?
Type I and II errors and power, large-sample z-tests, t, F and chi-square tests with their degrees of freedom, confidence intervals and properties of estimators.
How many tests and questions are in this chapter?
Three tests of 25 questions each, 75 questions in all, in syllabus order — Test 1: point and interval estimation, properties of estimators, maximum likelihood, method of moments, Cramér–Rao and Rao–Blackwell; Test 2: hypotheses, level of significance, Type I and II errors, power, p-values and large-sample tests for means and proportions; Test 3: t-tests, the F-test, chi-square tests of goodness of fit, independence and variance, and non-parametric tests.
Are the Statistical Inference tests available in Hindi?
Yes. Every test is available in both English and Hindi with the same questions and the same answer key, and you choose the language when you start the test.
What kinds of questions are in the tests?
The same kinds SSC uses in Paper II: short numericals, direct one-line questions, statement questions, match the following (test with its use, term with its meaning) and assertion–reason. Many are short calculations, as in the real paper.
What do the solutions contain?
Six to seven short points for every question, in English and Hindi: the answer, the formula or rule behind it (with the working line by line for numericals), related facts that are asked alongside it, and what each wrong option actually refers to.
Is there a timer, and what is the marking scheme?
An optional 15-minute timer per test for quick timed practice, and the Paper II marking of +2 for a correct answer, −0.5 for a wrong one and 0 for an unattempted question.
Are the Statistical Inference tests free?
Yes, every test is free. You need a free TrickySSC account to start one — sign in once with Google or your mobile number and every test is open to you.

More Free SSC CGL Practice