LAISHA
Open Laisha

ASRB Agricultural Research Service Statistical Inference II: Testing of Hypotheses & Sequential Analysis: practice questions

ASRB Agricultural Research Service (ARS) - Agricultural Statistics scientist · India · Agricultural Scientists Recruitment Board (ICAR)

Original practice questions written to the published syllabus · not past papers
10worked here, free
25written to this exam in this subject
9syllabus units

About ASRB Agricultural Research Service

ASRB Agricultural Research Service (ARS) - Agricultural Statistics scientist is a government recruitment exam run by Agricultural Scientists Recruitment Board (ICAR), for candidates in India. Combined NET/ARS/SMS/STO recruitment: preliminary (objective), mains (descriptive, discipline paper) and interview for ARS.

Official site: asrb.org.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 ASRB Agricultural Research Service

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 25 ASRB Agricultural Research Service questions in this subject: 5 easy, 14 medium, 6 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 ASRB Agricultural Research Service and are not shown on any other exam's page.

Question 1 · medium

A small trial finds no significant yield difference between a new and a standard variety (p = 0.40). What is the correct conclusion?

  1. The trial proves that any difference is smaller than 0.40 t/ha
  2. The new variety is inferior, since the test failed to show it is better
  3. The data do not show a difference; they do not prove the varieties are equal
  4. The varieties are equal in yield, at the 60% confidence level
Answer and working

C. The data do not show a difference; they do not prove the varieties are equal

Failing to reject H₀ is not evidence for H₀, especially in a small, low-powered trial. To claim equivalence, set up an equivalence (TOST) test or show that the confidence interval lies within an agreed margin.

Question 2 · easy

A 4 × 3 contingency table of crop variety against disease grade is tested for independence. How many degrees of freedom does the test have?

  1. 7
  2. 6
  3. 12
  4. 11
Answer and working

B. 6

(r − 1)(c − 1) = 3 × 2 = 6. 12 is the number of cells; 11 is cells minus one.

Question 3 · hard

A likelihood-ratio test of H₀: σ²_g = 0 (no genetic variance) against σ²_g > 0 gives a naive χ²₁ p-value of 0.08. What is the appropriate p-value?

  1. 0.28, the square root of 0.08, because the variance is a squared quantity on the standard-deviation scale
  2. 0.04, because the null value is on the boundary and the limit is a 50:50 mixture of 0 and χ²₁
  3. 0.16, because the one-sided alternative doubles the two-sided p-value
  4. 0.08, because Wilks' theorem applies whenever the models are nested
Answer and working

B. 0.04, because the null value is on the boundary and the limit is a 50:50 mixture of 0 and χ²₁

With the null value on the edge of the parameter space, −2 log Λ is asymptotically ½χ²₀ + ½χ²₁, so the p-value is half the naive χ²₁ value: 0.04. Ignoring this makes variance-component tests conservative.

Question 4 · hard

In a logistic regression with a very large true effect, the Wald statistic for that coefficient becomes smaller as the effect grows. What is this called and what is the remedy?

  1. The Stein phenomenon; shrink the coefficient towards zero
  2. The Hauck-Donner effect; use the likelihood-ratio test instead
  3. The Lindley paradox; use a Bayesian test with a flat prior
  4. The Neyman-Scott problem; use REML for that coefficient
Answer and working

B. The Hauck-Donner effect; use the likelihood-ratio test instead

Near separation the standard error grows faster than the estimate, so the Wald z falls (Hauck and Donner, 1977). The LRT does not suffer from this.

Question 5 · easy

Ten independent tests of true null hypotheses are each carried out at α = 0.05. What is the probability of at least one false rejection?

  1. 0.5987
  2. 0.5000
  3. 0.4013
  4. 0.0500
Answer and working

C. 0.4013

1 − 0.95¹⁰ = 0.4013. 0.5000 is 10 × 0.05, the expected number of false rejections, not a probability; 0.5987 is the probability of none.

Question 6 · easy

Twenty varieties are each compared with a standard, and the family-wise error rate is to be held at 0.05 by the Bonferroni method. At what level is each comparison tested?

  1. 0.05
  2. 0.0026
  3. 0.0025
  4. 0.01
Answer and working

C. 0.0025

α/m = 0.05/20 = 0.0025. 0.0026 is Šidák's 1 − 0.95^(1/20) ≈ 0.00256, a slightly less conservative alternative.

Question 7 · easy

In a variety trial, eight new entries are each to be compared only with one check variety, with control of the family-wise error rate. Which procedure is designed for this?

  1. Duncan's multiple range test
  2. Scheffé's method
  3. Dunnett's test
  4. Tukey's HSD test
Answer and working

C. Dunnett's test

Dunnett's procedure handles many-to-one comparisons with a control and is more powerful there than Tukey's all-pairs method or Scheffé's all-contrasts method.

Question 8 · easy

A marker screen tests 200 markers, all with no true effect, each at α = 0.05. How many false positives are expected?

  1. 0.05
  2. 1
  3. 10
  4. 5
Answer and working

C. 10

Expected false rejections = mα = 200 × 0.05 = 10, whatever the dependence among the tests.

Question 9 · medium

A breeder screens 5,000 markers to pick candidates for follow-up, then confirms a few in an independent trial for a release decision. Which error rates suit the two stages?

  1. Bonferroni at both stages, since it is always the most powerful method
  2. No correction at either stage, since the markers are independent
  3. FDR for the screen, FWER for the confirmatory trial
  4. FWER control for the screen, FDR control for the confirmatory trial
Answer and working

C. FDR for the screen, FWER for the confirmatory trial

Exploratory screening tolerates some false leads in return for power, so FDR suits it; a release claim needs strict control of any false claim, so FWER. Bonferroni is not the most powerful FWER method (Holm dominates it).

Question 10 · hard

A seed treatment should raise germination from 80% to 90%. With two-sided α = 0.05 and power 0.80, about how many seeds per group are needed by the normal approximation?

  1. 100
  2. 393
  3. 199
  4. 49
Answer and working

C. 199

n = [1.96√(2 × 0.85 × 0.15) + 0.8416√(0.16 + 0.09)]²/0.1² = 199.0, so about 199 per group. 49 is only the size for estimating one proportion within ±0.1 and ignores power.

A study plan for Statistical Inference II: Testing of Hypotheses & Sequential Analysis in ASRB Agricultural Research Service

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 25 questions): Multiple testing: family-wise error rate and false discovery rate (6, 24%); Power and sample-size calculation for common tests (5, 20%); Likelihood ratio tests and Wilks' theorem (3, 12%).

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

  2. Then the middle band: Sequential probability ratio test: OC and ASN functions (3, 12%); Tests based on Z, t, chi-square and F distributions (3, 12%); Hypotheses, the two types of error, size and power (2, 8%).

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

  3. Then the lighter units: Neyman-Pearson lemma and most powerful tests (1, 4%); Relation between tests and confidence intervals (1, 4%); Wald and score (Rao) 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: Wald and score (Rao) tests (1 hard of 1), Likelihood ratio tests and Wilks' theorem (2 hard of 3), Sequential probability ratio test: OC and ASN functions (2 hard of 3).

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

  5. Bank the easy marks: Multiple testing: family-wise error rate and false discovery rate (4 easy of 6); Likelihood ratio tests and Wilks' theorem (1 easy of 3).

    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 — ASRB Agricultural Research Service Statistical Inference II: Testing of Hypotheses & Seque

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.