ICAR AICE-JRF/SRF (PhD) · India · NTA for ICAR
Original practice questions written to the published syllabus · not past papersThe statistics a veterinarian needs to read a paper, design a trial or analyse herd records: descriptive statistics, tests of significance, regression and experimental design.
In the exams: ICAR AIEEA-PG 14 (statistics unit); ASRB NET
ICAR AICE-JRF/SRF (PhD)
Research methodology, experimental design (RBD, Latin square, factorial, missing plot), sample-size reasoning, multiple comparisons, regression and the interpretation of P values and confidence intervals.
How our questions for ICAR AICE-JRF/SRF spread across the units of Biostatistics, Computer Application & Research Methodology. Difficulty: 1 easy, 10 medium, 20 hard.
| Sample-size calculation | 5 | |
| Correlation and regression | 4 | |
| Experimental designs: CRD and RBD | 4 | |
| Research methodology | 4 | |
| Analysis of variance | 3 | |
| Computer and spreadsheet basics | 3 | |
| Probability distributions | 2 | |
| Tests of significance: the t-test | 2 | |
| The chi-square test | 2 | |
| Measures of dispersion | 1 | |
| Sampling methods | 1 |
Choose an answer, then open the working. Each explanation says why the right answer is right and why the others are not.
Two diagnostic tests are applied to the same 200 sheep. Which test compares the proportions positive by the two methods?
A. McNemar's test using the discordant pairs
Results on the same animals are paired, so McNemar's test applies: χ² = (b − c)²/(b + c), using only sheep positive on one test and negative on the other. An ordinary chi-square treats the two sets of results as independent samples.
Further reading: Thrusfield M, Christley R. Veterinary Epidemiology, 4th ed.
Tick counts per animal are grouped into 6 classes and a Poisson distribution, with its mean estimated from the same data, is fitted. How many df does the goodness-of-fit chi-square test have?
D. 4
df = classes − 1 − parameters estimated = 6 − 1 − 1 = 4. 5 forgets the df lost for estimating the Poisson mean; 3 subtracts two parameters, as for a fitted normal.
Further reading: Snedecor GW, Cochran WG. Statistical Methods, 8th ed.
A one-way ANOVA of rat liver weights gives treatment SS 54 and error SS 6. What proportion of the total variation is explained by the treatments (eta squared)?
B. 0.90
Total SS = 54 + 6 = 60; η² = 54/60 = 0.90. 0.10 is the unexplained share (6/60); 9.00 is the ratio of treatment to error SS and 0.11 its inverse (6/54).
Further reading: Petrie A, Watson P. Statistics for Veterinary and Animal Science, 3rd ed.
Farms with more cattle report both more milk sold and more cases of mastitis, giving a strong correlation between milk sold and mastitis cases. What is the best interpretation?
C. The correlation is likely spurious, both depending on herd size
Both variables rise with a third variable, herd size, so their correlation does not show that one causes the other. Analysing per-cow rates or adjusting for herd size would remove the confounding.
Further reading: Dohoo I, Martin W, Stryhn H. Veterinary Epidemiologic Research, 2nd ed.
In a logistic regression of calf mortality on colostrum feeding, the coefficient for 'no colostrum within 6 hours' is 0.693. What does this mean?
D. Odds of death are doubled
In logistic regression, exp(coefficient) is the odds ratio: e^0.693 = 2.0, so the odds of death are twice as high. The coefficient is on the log-odds scale, not a probability or a percentage change in risk.
Further reading: Dohoo I, Martin W, Stryhn H. Veterinary Epidemiologic Research, 2nd ed.
A 2³ factorial experiment is laid out in a randomised block design with 3 blocks. What are the error degrees of freedom?
A. 14
8 treatment combinations: treatment df 7, block df 2, error df 7 × 2 = 14, total 23. 16 is the CRD error df 8 × 2; 7 is the treatment df; 21 subtracts only the block df from the total (23 − 2).
Further reading: Gomez KA, Gomez AA. Statistical Procedures for Agricultural Research, 2nd ed.
A simple-random-sample size of 246 cattle is needed. The survey will instead sample 10 cattle per herd, and the intra-herd correlation is 0.2. Using design effect DE = 1 + (m − 1)ρ, how many cattle are needed?
A. 689
DE = 1 + 9 × 0.2 = 2.8; n = 246 × 2.8 = 688.8, rounded up to 689. 492 assumes DE = 2; 295 uses DE = 1.2 (ρ without the m − 1 factor; 295.2); 2460 multiplies by m.
Further reading: Dohoo I, Martin W, Stryhn H. Veterinary Epidemiologic Research, 2nd ed.
A vaccine trial expects 30% morbidity in controls and 15% in vaccinated calves. Using n = (zα/2 + zβ)²[p1q1 + p2q2]/(p1 − p2)² per group, α = 0.05 two-sided and 80% power (z values 1.96 and 0.842), how many calves per group are needed?
C. 118
(2.802)² = 7.849; p1q1 + p2q2 = 0.21 + 0.1275 = 0.3375; n = 7.849 × 0.3375/0.0225 = 117.7, so 118 per group. 58 ignores power; 236 is the total for both groups and 59 wrongly halves the per-group figure. Pooled-variance or continuity-corrected formulas give somewhat larger numbers.
Further reading: Thrusfield M, Christley R. Veterinary Epidemiology, 4th ed.
In a clinical trial of an analgesic in dogs, neither the owners nor the clinicians scoring pain know which dog received the drug or the placebo. What is the purpose of this?
B. To prevent bias in assessing the outcome
Blinding protects against measurement (observer) and performance bias, especially for subjective outcomes such as pain scores. Comparability at baseline comes from randomisation, not blinding.
Further reading: Petrie A, Watson P. Statistics for Veterinary and Animal Science, 3rd ed.
A trial in one university herd of high-yielding cows shows a clear benefit of a ration. Smallholders ask whether it applies to their low-yielding indigenous cows. This question concerns:
C. External validity of the trial
External validity (generalisability) asks whether results apply beyond the study population. Internal validity asks whether the result is true for the animals studied.
Further reading: Petrie A, Watson P. Statistics for Veterinary and Animal Science, 3rd ed.
Ten here, 31 in the app. The Laisha app has every practice question we have written to ICAR AICE-JRF/SRF in this subject, marked as you go, with timed papers, a jump-to-unanswered list and a certificate when you pass at 70 per cent.
Free for ten days, then one subscription of US$10 a month. We would rather say so here than have you find it later.
An AI operating system for agriculture: free guides, manuals and tools written for every country.