LAISHA
Open Laisha

CSIR-UGC NET Statistical Inference II: Testing of Hypotheses & Sequential Analysis: practice questions

CSIR-UGC NET - Mathematical Sciences (statistics option) · India · National Testing Agency for CSIR

Original practice questions written to the published syllabus · not past papers
10worked here, free
27written to this exam in this subject
9syllabus units

About CSIR-UGC NET

CSIR-UGC NET - Mathematical Sciences (statistics option) is a PhD, fellowship and lectureship exam run by National Testing Agency for CSIR, for candidates in India. Computer-based, 3 hours, Parts A (general aptitude), B and C (subject); statisticians answer from the statistics unit.

Official site: csirnet.nta.ac.in.

We have not been able to confirm these details against the exam body's own specification, so fees, dates and pass marks are left out on purpose. Check the current format, eligibility, fees and dates on the official site.

Statistical Inference II: Testing of Hypotheses & Sequential Analysis in CSIR-UGC NET

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 27 CSIR-UGC NET questions in this subject: 2 easy, 10 medium, 15 hard, across 9 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 CSIR-UGC NET and are not shown on any other exam's page.

Question 1 · hard

For H₀: p = 0.5 against H₁: p = 0.8 with 10 Bernoulli trials, the MP test rejects for large X. P(X ≥ 9 | H₀) = 11/1024 and P(X = 8 | H₀) = 45/1024. For an exact size of 0.05, with what probability should H₀ be rejected when X = 8?

  1. 1.138
  2. 0.893
  3. 0.050
  4. 0.107
Answer and working

B. 0.893

Size = P(X ≥ 9) + γP(X = 8) = 0.05 gives γ = (0.05 − 11/1024)/(45/1024) = 0.893. Dividing 0.05 by P(X = 8) alone ignores the rejections already made at X ≥ 9.

Question 2 · medium

A family {f(x; θ)} has a monotone likelihood ratio in T(x) when:

  1. f(x; θ₂)/f(x; θ₁) is non-decreasing in T(x) for every θ₂ > θ₁
  2. f(x; θ) is non-decreasing in θ for every fixed sample point x
  3. T(x) is sufficient and complete for θ in the given family
  4. The MLE of θ is a non-decreasing function of the sample size
Answer and working

A. f(x; θ₂)/f(x; θ₁) is non-decreasing in T(x) for every θ₂ > θ₁

MLR concerns the likelihood ratio between two parameter values as a function of the statistic. Sufficiency and completeness are separate properties (though exponential families have both).

Question 3 · medium

X₁, ..., Xₙ are iid N(μ, σ²) with σ² unknown. For H₀: μ ≤ μ₀ against μ > μ₀, which statement is correct?

  1. The one-sided t-test is UMP among all tests whose level is α
  2. The z-test with σ replaced by s is UMP among all level-α tests
  3. No UMP test exists; the one-sided t-test is UMP among unbiased tests
  4. No test of level α exists, because σ² is a nuisance parameter
Answer and working

C. No UMP test exists; the one-sided t-test is UMP among unbiased tests

The nuisance parameter σ² prevents a UMP test, but the one-sided t-test is UMPU (and UMP among invariant tests). Replacing σ by s without the t distribution does not give exact level α.

Question 4 · hard

In what sense is Fisher's exact test for a 2 × 2 table optimal?

  1. It has exactly size α for every table, with no randomisation needed
  2. Its randomised version is UMPU for the odds ratio, by conditioning on the margins
  3. It is the likelihood-ratio test, so Wilks' theorem gives its exact size
  4. It is UMP among all tests of independence, conditionally and unconditionally, for any table size
Answer and working

B. Its randomised version is UMPU for the odds ratio, by conditioning on the margins

The margins are sufficient for the nuisance parameters; conditioning on them gives a hypergeometric null distribution, and the randomised conditional test is UMPU. The non-randomised version is conservative.

Question 5 · hard

For H₀: σ² = σ₀² against σ² ≠ σ₀² in a normal sample with μ unknown, how do the UMPU cut-offs for (n − 1)s²/σ₀² compare with the usual equal-tail points?

  1. They differ only when n is larger than 30
  2. They differ, because the chi-square distribution is skewed
  3. They coincide, because every UMPU test has equal tails
  4. They coincide, because s² is unbiased for σ²
Answer and working

B. They differ, because the chi-square distribution is skewed

The UMPU conditions (size α and zero slope of power at σ₀²) do not give equal tails for a skewed null distribution; for small n the difference is noticeable. Equal tails are used in practice for convenience.

Question 6 · medium

Pearson's chi-square statistic Σ(O − E)²/E for a multinomial goodness-of-fit test belongs to which family of tests?

  1. Score tests
  2. Wald tests
  3. Sequential probability ratio tests
  4. Exact conditional tests
Answer and working

A. Score tests

Pearson's X² is the score statistic for the multinomial, using expected (null) frequencies in the denominator. Neyman's modified statistic with observed frequencies is a Wald-type statistic; G² is the LRT.

Question 7 · hard

Ten Poisson counts total 30. What are the score and Wald χ² statistics for H₀: λ = 2?

  1. Score 3.333, Wald 5.000
  2. Score 5.000, Wald 5.000
  3. Score 5.000, Wald 3.333
  4. Score 5.000, Wald 1.826
Answer and working

C. Score 5.000, Wald 3.333

Score: U(2) = 30/2 − 10 = 5, I(2) = n/λ₀ = 5, S = 25/5 = 5.000. Wald: (λ̂ − λ₀)² × n/λ̂ = 1 × 10/3 = 3.333. The LRT (4.328) lies between them; all three refer to χ²₁.

Question 8 · medium

Which confidence interval for a binomial proportion is obtained by inverting the score test?

  1. The arcsine interval
  2. The Wilson interval
  3. The Clopper-Pearson interval
  4. The Wald interval
Answer and working

B. The Wilson interval

Solving |p̂ − p₀| ≤ z√(p₀(1 − p₀)/n) for p₀ gives the Wilson interval. The Wald interval inverts the Wald test; Clopper-Pearson inverts the exact binomial tests.

Question 9 · hard

Ten ordered p-values are 0.001, 0.004, 0.012, 0.020, 0.035, 0.060, 0.11, 0.25, 0.48 and 0.71. How many hypotheses does the Benjamini-Hochberg procedure reject at FDR q = 0.05?

  1. 4
  2. 2
  3. 3
  4. 5
Answer and working

A. 4

Find the largest k with p₍ₖ₎ ≤ kq/m = 0.005k: 0.020 ≤ 0.020 at k = 4, while 0.035 > 0.025 at k = 5 and later values fail too. So k = 4. Bonferroni (0.005) rejects 2.

Question 10 · hard

What did Wald and Wolfowitz prove about the SPRT?

  1. Among tests with error rates at most α and β, it minimises the expected sample size under both H₀ and H₁
  2. It minimises the maximum sample size among all sequential tests with the same error rates
  3. It has exactly the nominal error rates α and β when Wald's boundary approximations are used for A and B
  4. It is uniformly most powerful among fixed-sample tests of composite hypotheses
Answer and working

A. Among tests with error rates at most α and β, it minimises the expected sample size under both H₀ and H₁

The optimality concerns E(N) at θ₀ and θ₁. Wald's approximations give actual error rates close to, but not exactly equal to, α and β; the SPRT's sample size is unbounded unless truncated.

A study plan for Statistical Inference II: Testing of Hypotheses & Sequential Analysis in CSIR-UGC NET

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 (56% of the 27 questions): Unbiased tests and UMPU tests for exponential families (6, 22%); Wald and score (Rao) tests (5, 19%); Monotone likelihood ratio and uniformly most powerful tests (4, 15%).

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

  2. Then the middle band: Relation between tests and confidence intervals (4, 15%); Likelihood ratio tests and Wilks' theorem (2, 7%); Multiple testing: family-wise error rate and false discovery rate (2, 7%).

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

  3. Then the lighter units: Sequential probability ratio test: OC and ASN functions (2, 7%); Hypotheses, the two types of error, size and power (1, 4%); Neyman-Pearson lemma and most powerful tests (1, 4%).

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

  4. Leave time for the hard ones: Likelihood ratio tests and Wilks' theorem (2 hard of 2), Sequential probability ratio test: OC and ASN functions (2 hard of 2), Neyman-Pearson lemma and most powerful tests (1 hard of 1).

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

  5. Bank the easy marks: Wald and score (Rao) tests (1 easy of 5); Unbiased tests and UMPU tests for exponential families (1 easy of 6).

    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 — CSIR-UGC NET Statistical Inference II: Testing of Hypotheses & Sequential Analysis

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.