IIT JAM - Mathematical Statistics (MS) · India · IITs (JAM 2026 organised by IIT Bombay)
Original practice questions written to the published syllabus · not past papersIIT JAM - Mathematical Statistics (MS) is a postgraduate admission exam run by IITs (JAM 2026 organised by IIT Bombay), for candidates in India. Computer-based, 3 hours, 60 questions, 100 marks, English. Section A 30 MCQs (10 x 1 mark, 20 x 2 marks); Section B 10 MSQs (2 marks, no negative or partial marking); Section C 20 NAT (10 x 1, 10 x 2, no negative). MCQ wrong answers lose 1/3 or 2/3 mark.
Source: jam2026.iitb.ac.in.
Checked against the exam body's published information. Formats and fees change; confirm the current edition there before you book.
Most powerful critical regions, size and power computations (including discrete and uniform cases), MLR and UMP tests, LRT statistics and their asymptotic chi-square df, and p-values; multiple-select items on which tests are UMP or unbiased.
The logic and optimality of tests: errors and power, the Neyman-Pearson lemma, uniformly most powerful and unbiased tests, likelihood-ratio tests and their large-sample forms, and Wald's sequential probability ratio test.
Our 59 IIT JAM questions in this subject: 18 easy, 29 medium, 12 hard, across 11 units.
Choose an answer, then open the working. Each explanation says why the right answer is right and why the others are not.
These 10 were written to IIT JAM and are not shown on any other exam's page.
Which property does every most powerful test of size α (0 < α < 1) for simple hypotheses have?
C. Power ≥ α at θ₁
The MP test is unbiased: the trivial test that rejects with probability α regardless of the data has power α, and the MP test cannot do worse. Randomisation may be needed with discrete data.
X₁, ..., Xₙ are iid N(μ, σ²) with μ known. For H₀: σ² = σ₀² against H₁: σ² = σ₁² with σ₁² > σ₀², the most powerful test rejects when:
A. Σ(Xᵢ − μ)² is large
log Λ = −(n/2)log(σ₁²/σ₀²) + (1/2)(1/σ₀² − 1/σ₁²)Σ(Xᵢ − μ)², increasing in Σ(Xᵢ − μ)² when σ₁² > σ₀². The test uses χ²ₙ critical values since μ is known.
Five observations come from uniform (0, θ). For H₀: θ ≤ 1 against H₁: θ > 1, the UMP size-0.05 test rejects when X₍₅₎ > c. What is c?
D. 0.98979
P(X₍₅₎ > c | θ = 1) = 1 − c⁵ = 0.05, so c = 0.95^(1/5) = 0.98979. 0.05^(1/5) = 0.54928 would give size 0.95.
For a normal mean with σ unknown, what is the optimality property of the two-sided one-sample t-test?
C. It is UMPU of level α
With σ unknown no UMP test exists, but the t-test is uniformly most powerful among unbiased level-α tests (and among invariant tests).
For a normal sample with σ unknown, the likelihood-ratio test of H₀: μ = μ₀ against μ ≠ μ₀ is equivalent to which test?
A. The two-sided t-test
Λ = [1 + t²/(n − 1)]^(−n/2), a decreasing function of |t|, so rejecting for small Λ is rejecting for large |t|.
For a normal mean with known σ, how do the Wald, score and likelihood-ratio statistics for H₀: μ = μ₀ compare?
D. They are identical, all equal to n(x̄ − μ₀)²/σ²
The log-likelihood is exactly quadratic in μ with constant curvature n/σ², so all three give n(x̄ − μ₀)²/σ², which is exactly χ²₁ under H₀.
When should Welch's t-test be preferred to the pooled two-sample t-test?
A. When the two population variances differ, especially with unequal sample sizes
Welch's test uses separate variances and Satterthwaite's approximate df, protecting the Type I error rate when variances differ. Paired data need the paired t-test; counts may need a GLM.
The correlation between plant height and grain yield over 27 plots is r = 0.5. What is the t-statistic for H₀: ρ = 0, and its df?
D. t = 2.887 with 25 df
t = r√(n − 2)/√(1 − r²) = 0.5 × 5/√0.75 = 2.887, with n − 2 = 25 df. Dividing by 1 − r² instead of its square root gives 3.333.
A 99% confidence interval for an effect contains 0, but the 95% interval from the same data does not. What can be said about the two-sided p-value for H₀: effect = 0?
A. It lies between 0.01 and 0.05
Excluded from the 95% interval means p < 0.05; included in the 99% interval means p > 0.01.
With 10 independent tests and family-wise level 0.05, what per-test level does Šidák's method use?
A. 0.00512
1 − (1 − 0.05)^(1/10) = 0.00512, slightly above the Bonferroni 0.005 and exact under independence.
This plan follows where our questions actually fall, so the time goes where the marks are likely to be. It is our reading of the syllabus, not the exam body's weighting.
Read the manual chapters for these first and do every worked question you can find on them.
Aim for one sitting per unit, then mixed questions across them.
Learn the definitions, classifications and standard numbers; these units are more often recalled than worked.
Come back to these after a first pass, untimed, and read every explanation, including for the options you rejected.
Do these early in revision so they are fast on the day.
Last updated · How Laisha writes and checks its pages · Report a mistake
US$10 a month · one price worldwide · cancel any time
| Free | Smarter | |
|---|---|---|
| Grow guides, pest pages, articles | Read all | Read all |
| Ask Laisha | 5 a day | Up to 1,000 a month |
| Exam practice | The sample on each page | Every question, timed papers |
| Crop calendar and pest alerts | — | For your own fields |
| Calculators and tools | — | All |
| Uploads and reports | — | 100 a month, real files |
An AI operating system for agriculture: free guides, manuals and tools written for every country.