LAISHA
Open Laisha

ICAR AICE-JRF/SRF Statistical Inference I: Estimation: practice questions

ICAR AICE-JRF/SRF - Agricultural Statistics, Computer Application, Bioinformatics (PhD admission and fellowship) · India · National Testing Agency for ICAR

Original practice questions written to the published syllabus · not past papers
10worked here, free
51written to this exam in this subject
11syllabus units

About ICAR AICE-JRF/SRF

ICAR AICE-JRF/SRF - Agricultural Statistics, Computer Application, Bioinformatics (PhD admission and fellowship) is a PhD, fellowship and lectureship exam run by National Testing Agency for ICAR, for candidates in India. Computer-based test, 2 hours, 120 MCQs, +4 / -1 (Part A general, Part B core group, Part C subject). 2026 exam held with AIEEA-PG in July 2026.

Source: icar.org.in.

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

Statistical Inference I: Estimation in ICAR AICE-JRF/SRF

Statements and applications of the Cramér-Rao, Rao-Blackwell and Lehmann-Scheffé theorems; sufficiency by factorisation; MLE and moment estimators of standard distributions; interval estimation.

How to estimate well and know it: sufficiency and completeness, unbiasedness and minimum variance, the Cramer-Rao bound, the Rao-Blackwell and Lehmann-Scheffe theorems, the methods of moments, maximum likelihood and minimum chi-square, and confidence intervals.

Our 51 ICAR AICE-JRF/SRF questions in this subject: 11 easy, 20 medium, 20 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 ICAR AICE-JRF/SRF and are not shown on any other exam's page.

Question 1 · hard

X is a single Poisson observation with mean λ. Which is the unique unbiased estimator of e^(−2λ)?

  1. e^(−2X)
  2. (−1)^X
  3. 1{X = 0}
  4. (1/2)^X
Answer and working

B. (−1)^X

E(c^X) = e^(λ(c − 1)); setting c − 1 = −2 gives c = −1, so E[(−1)^X] = e^(−2λ). It is unbiased yet absurd, taking the value −1 for odd X, a warning that unbiasedness alone is not a sufficient criterion. 1{X = 0} is unbiased for e^(−λ); (1/2)^X has mean e^(−λ/2).

Question 2 · medium

X₁, ..., Xₙ are iid uniform on (θ, θ + 1). What is the minimal sufficient statistic for θ?

  1. X̄
  2. (X₍₁₎, X₍ₙ₎)
  3. (X̄, X₍ₙ₎) as a pair
  4. X₍₁₎ alone
Answer and working

B. (X₍₁₎, X₍ₙ₎)

L(θ) = 1{X₍ₙ₎ − 1 ≤ θ ≤ X₍₁₎}, which depends on both extremes; the likelihood-ratio criterion shows that the pair is minimal. A two-dimensional minimal sufficient statistic for a one-dimensional parameter: the family is not of exponential type.

Question 3 · hard

For a random sample from the Cauchy distribution with unknown location θ and scale 1, what is the minimal sufficient statistic?

  1. The sample median
  2. The full set of order statistics
  3. The pair of extreme values X₍₁₎ and X₍ₙ₎
  4. The sample mean
Answer and working

B. The full set of order statistics

No reduction below the order statistics is possible for the Cauchy location family: the ratio of likelihoods at two samples is free of θ only if the samples are permutations of each other. The median is a good estimator but not sufficient.

Question 4 · medium

X₁, ..., Xₙ are iid with density θx^(θ−1) on (0, 1), θ > 0. Which statistic is sufficient for θ?

  1. Σ log Xᵢ
  2. ΣXᵢ and ΣXᵢ²
  3. X₍ₙ₎
  4. ΣXᵢ
Answer and working

A. Σ log Xᵢ

L = θⁿ(ΠXᵢ)^(θ−1) = θⁿ exp[(θ − 1)Σ log Xᵢ], so ΠXᵢ, equivalently Σ log Xᵢ, is sufficient. This is a beta(θ, 1) exponential family with T(x) = log x.

Question 5 · hard

X₁, ..., Xₙ are iid exponential with rate λ. Which statement is correct?

  1. (n − 1)/ΣXᵢ is the UMVUE of λ and attains the bound λ²/n for every n > 2
  2. 1/X̄ is unbiased for λ and attains the bound λ²/n
  3. No unbiased estimator of λ attains the bound λ²/n; the UMVUE has variance λ²/(n − 2)
  4. X̄ is unbiased for λ and attains the bound λ²/n
Answer and working

C. No unbiased estimator of λ attains the bound λ²/n; the UMVUE has variance λ²/(n − 2)

The score n/λ − ΣXᵢ is linear in X̄, whose mean is 1/λ; so only 1/λ (the mean) has an attaining estimator. For λ, the UMVUE (n − 1)/ΣX has variance λ²/(n − 2) > λ²/n. 1/X̄ is biased, and X̄ estimates the mean, not the rate.

Question 6 · hard

X₁, ..., Xₙ are iid uniform on (θ − 1/2, θ + 1/2). Which statement about the MLE of θ is correct?

  1. It is unique and equals the sample maximum X₍ₙ₎
  2. It is not unique: any value in [X₍ₙ₎ − 1/2, X₍₁₎ + 1/2] maximises L
  3. It does not exist, because the likelihood is unbounded near the sample extremes
  4. It is unique and equals the sample mean X̄
Answer and working

B. It is not unique: any value in [X₍ₙ₎ − 1/2, X₍₁₎ + 1/2] maximises L

L(θ) = 1 when X₍ₙ₎ − 1/2 ≤ θ ≤ X₍₁₎ + 1/2 and 0 otherwise: a flat maximum over an interval. The midrange is one natural choice within it.

Question 7 · easy

What is the difference between the minimum chi-square and the modified minimum chi-square estimators?

  1. Minimum χ² is biased; the modified version is the unique unbiased estimator
  2. Minimum χ² uses grouped data; the modified version uses the raw, ungrouped observations
  3. Minimum χ² divides each squared deviation by the expected frequency; the modified version divides by the observed frequency instead
  4. Minimum χ² divides each squared deviation by the observed frequency; the modified version divides by the expected frequency
Answer and working

C. Minimum χ² divides each squared deviation by the expected frequency; the modified version divides by the observed frequency instead

Pearson's χ² = Σ(O − E)²/E; Neyman's modified χ² = Σ(O − E)²/O, which often gives linear estimating equations but fails when some O = 0. Both work on grouped data.

Question 8 · hard

Which large-sample property do minimum chi-square, modified minimum chi-square and grouped-data maximum-likelihood estimators share?

  1. They are all best asymptotically normal, asymptotically equivalent to each other
  2. They all attain the Cramér-Rao bound exactly, for every sample size and every model
  3. They are all exactly unbiased in samples of every size
  4. They all give identical estimates whatever the sample size
Answer and working

A. They are all best asymptotically normal, asymptotically equivalent to each other

All three are consistent, asymptotically normal and asymptotically efficient for grouped data (Neyman's BAN class). In finite samples they differ and are generally biased.

Question 9 · easy

What is a pivotal quantity?

  1. An estimator whose variance attains the Cramér-Rao bound for every θ
  2. A sufficient statistic that is also unbiased for θ
  3. A function of the data and θ whose distribution does not depend on θ at all
  4. A statistic whose distribution does not depend on θ
Answer and working

C. A function of the data and θ whose distribution does not depend on θ at all

A pivot involves θ (so it can be inverted) but has a θ-free distribution, e.g. (X̄ − μ)/(s/√n) ~ t. A statistic with a θ-free distribution and no θ in it is ancillary.

Question 10 · medium

How is the confidence coefficient of an interval procedure defined?

  1. The average over θ of the coverage probability P_θ(L ≤ θ ≤ U)
  2. The posterior probability that θ lies in the observed interval
  3. The infimum over θ of the coverage probability P_θ(L ≤ θ ≤ U), the worst case
  4. The coverage probability at the true value of θ only
Answer and working

C. The infimum over θ of the coverage probability P_θ(L ≤ θ ≤ U), the worst case

The confidence coefficient is the guaranteed (worst-case) coverage. For discrete data coverage varies with θ, and exact methods such as Clopper-Pearson keep the infimum at least at the nominal level.

A study plan for Statistical Inference I: Estimation in ICAR AICE-JRF/SRF

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 (49% of the 51 questions): Completeness, ancillarity and Basu's theorem (7, 14%); Iterative maximum likelihood: Newton-Raphson and Fisher scoring (7, 14%); Maximum likelihood: properties, invariance and asymptotic normality (6, 12%); Cramer-Rao inequality and Fisher information (5, 10%).

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

  2. Then the middle band: Minimum variance unbiased estimation; Rao-Blackwell and Lehmann-Scheffe theorems (5, 10%); Sufficiency, the factorisation theorem and minimal sufficiency (5, 10%); Minimum chi-square and other methods (4, 8%); Point estimation: unbiasedness, consistency and efficiency (4, 8%).

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

  3. Then the lighter units: Interval estimation: pivotal quantities and coverage (3, 6%); Large-sample confidence intervals (3, 6%); Method of moments (2, 4%).

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

  4. Leave time for the hard ones: Large-sample confidence intervals (2 hard of 3), Cramer-Rao inequality and Fisher information (3 hard of 5), Completeness, ancillarity and Basu's theorem (4 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: Method of moments (1 easy of 2); Minimum variance unbiased estimation; Rao-Blackwell and Lehmann-Scheffe theorems (2 easy of 5).

    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 — ICAR AICE-JRF/SRF Statistical Inference I: Estimation

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.