LAISHA
Open Laisha

IIT JAM Bayesian Inference: practice questions

IIT JAM - Mathematical Statistics (MS) · India · IITs (JAM 2026 organised by IIT Bombay)

Original practice questions written to the published syllabus · not past papers
10worked here, free
69written to this exam in this subject
11syllabus units

About IIT JAM

IIT 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.

Bayesian Inference in IIT JAM

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 69 IIT JAM questions in this subject: 19 easy, 38 medium, 12 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 IIT JAM and are not shown on any other exam's page.

Question 1 · medium

Aphid counts on 5 leaves are Poisson(λ) and sum to 18; λ has a gamma prior with shape 2 and rate 1. What is the posterior mean of λ?

  1. 1.053
  2. 3.600
  3. 3.333
  4. 4.000
Answer and working

C. 3.333

Posterior gamma(2 + 18, 1 + 5) = gamma(20, 6), mean 20/6 = 3.333. The MLE is 3.6; 4.0 forgets to add the prior rate to n.

Question 2 · hard

Pump lifetimes are exponential with rate λ; λ has a gamma prior with shape 2 and rate 10 (hours). Five lifetimes total 40 hours. What is the posterior mean of λ?

  1. 0.175 per hour
  2. 0.140 per hour
  3. 0.200 per hour
  4. 0.125 per hour
Answer and working

B. 0.140 per hour

The exponential likelihood λ⁵e^(−40λ) times gamma(2, 10) gives gamma(2 + 5, 10 + 40) = gamma(7, 50), with mean 7/50 = 0.140. The MLE is 5/40 = 0.125; the prior mean is 0.2.

Question 3 · medium

What is the variance of a beta(9, 6) posterior?

  1. 0.2400
  2. 0.0160
  3. 0.0150
  4. 0.0171
Answer and working

C. 0.0150

ab/[(a + b)²(a + b + 1)] = 54/(225 × 16) = 0.0150. 0.24 omits the factor (a + b + 1); dividing by 15 instead of 16 gives 0.0160.

Question 4 · hard

Why cannot an improper prior be used for the parameter under H₁ when computing a Bayes factor against a point null?

  1. Bayes factors require conjugate priors under both hypotheses
  2. Improper priors always make the posterior improper
  3. Its arbitrary constant does not cancel
  4. An improper prior would make B₁₀ equal to the likelihood ratio
Answer and working

C. Its arbitrary constant does not cancel

An improper prior cπ(θ) has an arbitrary c, which multiplies m₁(x) and hence B₁₀. For estimation c cancels in the posterior; for model comparison it does not.

Question 5 · medium

With equal prior probabilities on H₀ and H₁, the Bayes factor B₁₀ = 9. What is P(H₁ | x)?

  1. 0.90
  2. 0.81
  3. 0.99
  4. 0.11
Answer and working

A. 0.90

Posterior odds = 1 × 9 = 9, so P(H₁ | x) = 9/10 = 0.90. 0.11 is 1/9, a misread of the odds as a probability.

Question 6 · easy

Under absolute-error loss |θ − a|, what is the Bayes estimator?

  1. The posterior mode
  2. The posterior median
  3. The midrange of the posterior
  4. The posterior mean
Answer and working

B. The posterior median

Expected absolute loss is minimised at the median of the posterior distribution.

Question 7 · medium

What assumption justifies modelling variety-trial effects as draws from a common distribution?

  1. Normality of the raw plot yields
  2. Independence of every observation from the overall mean
  3. Equal numbers of plots in every trial
  4. Exchangeability of the trial effects
Answer and working

D. Exchangeability of the trial effects

If the trials are a priori indistinguishable, their parameters are exchangeable, and de Finetti's representation justifies a common population distribution. Balance and normality are not required.

Question 8 · easy

In a random-walk Metropolis algorithm with a symmetric proposal, the proposed value has posterior density 0.4 times the current value's. What is the acceptance probability?

  1. 1.0
  2. 0.6
  3. 2.5
  4. 0.4
Answer and working

D. 0.4

With symmetric q, α = min{1, π(θ*)/π(θ)} = min{1, 0.4} = 0.4. A move uphill (ratio above 1) would be accepted with probability 1.

Question 9 · medium

Why are several chains started from dispersed values instead of one long chain?

  1. A single chain cannot have a stationary distribution
  2. Disagreement between chains exposes problems
  3. Dispersed starts remove the need for burn-in
  4. Several chains always converge faster than one
Answer and working

B. Disagreement between chains exposes problems

If chains from different regions have not mixed, their summaries differ, which R̂ and trace plots expose; one chain stuck in one mode can look perfectly stable.

Question 10 · hard

Efron and Morris interpreted the James-Stein estimator in which way?

  1. As the posterior mode under a flat prior
  2. As the maximum-likelihood estimator under a Cauchy model
  3. As a minimax estimator for p = 1
  4. Empirical Bayes under a normal prior at zero
Answer and working

D. Empirical Bayes under a normal prior at zero

With θᵢ ~ N(0, A), the Bayes rule shrinks by 1/(1 + A); estimating that factor from ‖x‖² gives (p − 2)/‖x‖², the James-Stein form.

A study plan for Bayesian Inference in IIT JAM

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 69 questions): Conjugate families (9, 13%); Posterior summaries and credible intervals (8, 12%); Bayes' theorem for parameters: prior, likelihood and posterior (7, 10%); Convergence diagnostics (7, 10%).

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

  2. Then the middle band: Markov chain Monte Carlo: Gibbs sampling and Metropolis-Hastings (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, 9%).

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

  3. Then the lighter units: Empirical Bayes (4, 6%); 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: Markov chain Monte Carlo: Gibbs sampling and Metropolis-Hastings (3 hard of 7), Non-informative and weakly informative priors (3 hard of 7), Posterior predictive checks and model comparison (DIC, WAIC) (2 hard of 7).

    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 — IIT JAM 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.