LAISHA
Open Laisha

GATE ST Bayesian Inference: practice questions

GATE - Statistics (ST) · India · IITs and IISc on rotation (2027: IIT Madras)

Original practice questions written to the published syllabus · not past papers
10worked here, free
71written to this exam in this subject
11syllabus units

About GATE ST

GATE - Statistics (ST) is a postgraduate admission exam run by IITs and IISc on rotation (2027: IIT Madras), for candidates in India. Computer-based, 3 hours, 100 marks (15 General Aptitude + 85 subject). MCQ, MSQ and NAT; negative marking on MCQ only; on-screen virtual calculator.

Source: gate2026.iitg.ac.in.

Checked against the exam body's published information. Formats and fees change; confirm the current edition there before you book.

Bayesian Inference in GATE ST

Inference through prior and posterior: conjugate analysis, prior choice, credible intervals, Bayes factors and decision theory, hierarchical models, and computation by Markov-chain Monte Carlo, with agricultural uses in variety trials, genomic prediction and small-area estimation.

Our 71 GATE ST questions in this subject: 19 easy, 39 medium, 13 hard, across 11 units.

10 practice questions, worked

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 GATE ST and are not shown on any other exam's page.

Question 1 · medium

A seed lot is good (germination 0.9) with prior probability 0.8, otherwise poor (germination 0.7). Seven of 10 seeds germinate. What is the posterior probability that the lot is good?

  1. 0.177
  2. 0.462
  3. 0.0459
  4. 0.800
Answer and working

B. 0.462

Likelihoods: good C(10,7)0.9⁷0.1³ = 0.0574; poor C(10,7)0.7⁷0.3³ = 0.2668. Posterior = 0.8 × 0.0574/(0.8 × 0.0574 + 0.2 × 0.2668) = 0.462. 0.177 ignores the prior (equal prior weights); 0.0459 is the unnormalised joint probability.

Question 2 · hard

Two experiments on germination give proportional likelihoods for θ: one fixed n = 12 trials, the other sampled until the 3rd failure. What does Bayesian inference with the same prior conclude?

  1. Only the fixed-n experiment gives a valid posterior
  2. The posteriors differ, because the stopping rules differ
  3. The posteriors agree only when the prior is uniform
  4. The posteriors are identical
Answer and working

D. The posteriors are identical

Binomial and negative-binomial sampling with the same counts give likelihoods proportional to θ⁹(1 − θ)³, so the posteriors are equal for any prior. Frequentist p-values differ between the two designs because they depend on the sampling plan.

Question 3 · medium

A leaf sample comes from a field that is healthy, mildly or severely infected with prior probabilities 0.5, 0.3 and 0.2. A test is positive with probability 0.1, 0.6 and 0.9 in these states. The test is positive. What is the posterior probability of severe infection?

  1. 0.562
  2. 0.439
  3. 0.180
  4. 0.200
Answer and working

B. 0.439

m(+) = 0.5(0.1) + 0.3(0.6) + 0.2(0.9) = 0.41; P(severe | +) = 0.18/0.41 = 0.439. 0.180 is the joint probability; 0.562 ignores the prior.

Question 4 · hard

No infected plants are found among 10 tested. With Jeffreys' beta(1/2, 1/2) prior, what is the posterior mean infection rate?

  1. 0.0455
  2. 0.0500
  3. 0.0833
  4. 0.0000
Answer and working

A. 0.0455

The posterior is beta(0.5, 10.5), mean 0.5/11 = 0.0455. The MLE is 0; a uniform prior gives beta(1, 11), mean 1/12.

Question 5 · medium

Which statement compares equal-tailed and highest-posterior-density (HPD) credible intervals correctly?

  1. HPD is shortest; equal-tailed is transformation-invariant
  2. They coincide for every posterior distribution
  3. HPD intervals exist only for normal posteriors
  4. Equal-tailed is shortest; HPD is invariant to transformations
Answer and working

A. HPD is shortest; equal-tailed is transformation-invariant

The HPD interval minimises length for given probability but changes when θ is transformed; the equal-tailed interval maps directly under monotone transformations. They coincide for symmetric unimodal posteriors.

Question 6 · medium

The posterior for a variety's mean yield is N(4.517, 0.186²) t/ha. What is the 95% equal-tailed credible interval?

  1. 4.21 to 4.82 t/ha
  2. 4.15 to 4.88 t/ha
  3. 4.45 to 4.59 t/ha
  4. 4.21 to 4.99 t/ha
Answer and working

B. 4.15 to 4.88 t/ha

4.517 ± 1.96 × 0.186 = 4.15 to 4.88. Using 1.645 gives a 90% interval; 4.21 to 4.99 is the data-only interval that ignores the prior.

Question 7 · medium

An MCMC run gives 10,000 draws, and the sum of lag autocorrelations Σρₖ is 4.5. What is the effective sample size?

  1. 1818
  2. 1000
  3. 2222
  4. 10000
Answer and working

B. 1000

ESS = N/(1 + 2Σρₖ) = 10,000/10 = 1000. Forgetting the factor 2 gives 1818.

Question 8 · medium

A posterior SD is 0.5 and the effective sample size is 1,000. What is the Monte Carlo standard error of the posterior-mean estimate?

  1. 0.0050
  2. 0.0158
  3. 0.5000
  4. 0.0005
Answer and working

B. 0.0158

MCSE = SD/√ESS = 0.5/√1000 = 0.0158. Using the raw 10,000 draws ignores autocorrelation and gives 0.0050.

Question 9 · easy

What does thinning an MCMC chain (keeping every 10th draw) achieve?

  1. It saves storage but does not add information
  2. It replaces the need for a burn-in period
  3. It removes autocorrelation and so raises the ESS
  4. It guarantees convergence of the chain
Answer and working

A. It saves storage but does not add information

Thinning reduces correlation between kept draws but discards information; the ESS of the thinned chain is not larger than that of the full chain.

Question 10 · hard

James-Stein shrinkage towards the grand mean, rather than towards zero, uses which constant in place of p − 2?

  1. p
  2. p − 4
  3. p − 1
  4. p − 3
Answer and working

D. p − 3

Estimating the common centre costs one degree of freedom, so the shrinkage constant becomes p − 3 (requiring p ≥ 4 for domination).

A study plan for Bayesian Inference in GATE ST

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.

  1. First, the heavy units (45% of the 71 questions): Conjugate families (9, 13%); Markov chain Monte Carlo: Gibbs sampling and Metropolis-Hastings (8, 11%); Posterior summaries and credible intervals (8, 11%); Bayes' theorem for parameters: prior, likelihood and posterior (7, 10%).

    Read the manual chapters for these first and do every worked question you can find on them.

  2. Then the middle band: Convergence diagnostics (7, 10%); Non-informative and weakly informative priors (7, 10%); Posterior predictive checks and model comparison (DIC, WAIC) (7, 10%); Bayesian hypothesis testing and Bayes factors (6, 8%).

    Aim for one sitting per unit, then mixed questions across them.

  3. Then the lighter units: Empirical Bayes (5, 7%); Hierarchical (multilevel) models (4, 6%); Decision theory: loss functions and Bayes risk (3, 4%).

    Learn the definitions, classifications and standard numbers; these units are more often recalled than worked.

  4. Leave time for the hard ones: Non-informative and weakly informative priors (3 hard of 7), Empirical Bayes (2 hard of 5), Markov chain Monte Carlo: Gibbs sampling and Metropolis-Hastings (3 hard of 8).

    Come back to these after a first pass, untimed, and read every explanation, including for the options you rejected.

  5. Bank the easy marks: Non-informative and weakly informative priors (3 easy of 7); Posterior summaries and credible intervals (3 easy of 8).

    Do these early in revision so they are fast on the day.

  6. Finish timed. Sit the subject as a timed set without notes, mark it, and send each wrong answer back to its unit above.

Last updated · How Laisha writes and checks its pages · Report a mistake

Ask Laisha about this

Opens Laisha's chat with your question ready to send. Sign in or start free.

Unlock every question — GATE ST Bayesian Inference

US$10 a month · one price worldwide · cancel any time

  • Every practice question in this subject, with full working
  • Timed exam papers and progress by unit
  • Ask Laisha to explain any answer — up to 1,000 questions a month
Free or Smarter
FreeSmarter
Grow guides, pest pages, articlesRead allRead all
Ask Laisha5 a dayUp to 1,000 a month
Exam practiceThe sample on each pageEvery question, timed papers
Crop calendar and pest alerts—For your own fields
Calculators and tools—All
Uploads and reports—100 a month, real files
10 days free for every new account, no card.