ICAR AICE-JRF/SRF Linear Mixed Models, Variance Components & Spatial Analysis of Field Trials: 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
78written to this exam in this subject
10syllabus 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.
Checked against the exam body's published information. Formats and fees change; confirm the current edition there before you book.
Linear Mixed Models, Variance Components & Spatial Analysis of Field Trials in ICAR AICE-JRF/SRF
Random, fixed and mixed effects, nested and crossed classifications, expected mean squares and estimation of variance components; genetic variance and heritability as ratios of components.
Models with more than one source of random variation: ANOVA, ML and REML estimation of variance components, BLUE and BLUP, Henderson's mixed-model equations, and spatial models that correct field trials for fertility trends and neighbour correlation.
Our 78 ICAR AICE-JRF/SRF questions in this subject: 25 easy, 31 medium, 22 hard, across 10 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 · medium
When should genotype effects usually be fixed rather than random?
When a few named varieties are compared for release
When the aim is to estimate heritability of a breeding population
When thousands of lines are screened for selection
When genotypes are a sample from a population
Answer and working
A. When a few named varieties are compared for release
Precise comparison of specific varieties calls for BLUEs; selection among many sampled lines calls for BLUPs.
Question 2 · medium
Who introduced REML for variance component estimation?
Hartley and Rao (1967)
Kenward and Roger (1997)
Henderson (1953)
Patterson and Thompson, 1971
Answer and working
D. Patterson and Thompson, 1971
Hartley and Rao developed ML for mixed models; Henderson the ANOVA methods and MME; Kenward and Roger the small-sample correction.
Question 3 · medium
On what does REML maximise the likelihood?
The fixed-effect estimates only
The block totals of the experiment, ignoring individual plots
The squared residuals divided by n
A set of error contrasts free of the fixed effects
Answer and working
D. A set of error contrasts free of the fixed effects
Error contrasts (linear combinations with zero expectation) carry all information on variance components.
Question 4 · medium
In a balanced one-way random model, how does ML differ from REML for the between-group component?
ML ignores the within-group variance completely in this model
ML is always zero in this model
ML gives exactly the ANOVA estimator
ML divides the between-group SS by a instead of a - 1
Answer and working
D. ML divides the between-group SS by a instead of a - 1
This reflects ML's failure to account for estimating the mean, so ML is smaller than REML.
Question 5 · hard
What does the Kenward-Roger method adjust that Satterthwaite's does not?
The fixed-effect estimates themselves
The standard errors of fixed effects, for uncertainty in variance estimates
The residual sum of squares, recomputed without the random effects
The number of random effects
Answer and working
B. The standard errors of fixed effects, for uncertainty in variance estimates
KR inflates the estimated covariance of fixed effects and then uses a Satterthwaite-type df for the F test.
Question 6 · hard
Measurements are taken at days 14, 21, 35 and 63. Which structure handles the uneven spacing?
Toeplitz by measurement order
Ordinary AR(1) defined by the order of measurement
Compound symmetry
Spatial power: correlation rho^(time difference)
Answer and working
D. Spatial power: correlation rho^(time difference)
Structures defined on actual time differences respect unequal gaps; AR(1) by order treats all gaps as equal.
Question 7 · medium
What is the nugget in a spatial model?
The largest residual in the trial, used as an outlier flag
Independent measurement error at zero separation
The correlation between rows
The global row trend
Answer and working
B. Independent measurement error at zero separation
The nugget is the variance at very small distances, from measurement error and micro-variation.
Question 8 · hard
How is the amount of smoothing chosen in a P-spline spatial model fitted as a mixed model?
By estimating the penalty as a variance component by REML
By setting it equal to the number of rows in the field layout
By the analyst choosing it by eye
By using the largest possible penalty
Answer and working
A. By estimating the penalty as a variance component by REML
Writing the penalty as a random effect lets REML choose the smoothness.
Question 9 · easy
How many covariance parameters does compound symmetry have?
t(t + 1)/2
2
t
1
Answer and working
B. 2
One common variance and one common covariance.
Question 10 · easy
Which two directions does an AR1 x AR1 model use?
Genotypes and environments
Years and sites
Field rows and field columns
Replicates and blocks
Answer and working
C. Field rows and field columns
The separable process models correlation along rows and along columns of the plot grid.
A study plan for Linear Mixed Models, Variance Components & Spatial Analysis of Field Trials 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.
First, the heavy units (50% of the 78 questions): Covariance structures for repeated measures (11, 14%); Henderson's mixed-model equations; BLUE and BLUP (11, 14%); ANOVA (Henderson's methods) estimation of variance components (9, 12%); Fixed, random and mixed models revisited (8, 10%).
Read the manual chapters for these first and do every worked question you can find on them.
Then the middle band: Maximum likelihood and REML (8, 10%); Spatial analysis of field trials: nearest-neighbour and AR1 x AR1 models (8, 10%); Model selection with likelihood-ratio tests and information criteria (7, 9%); Unbalanced data and Satterthwaite and Kenward-Roger degrees of freedom (7, 9%).
Aim for one sitting per unit, then mixed questions across them.
Then the lighter units: Trend and spline adjustment (e.g. SpATS approach) (6, 8%); Software: lme4, nlme, ASReml, SAS PROC MIXED (3, 4%).
Learn the definitions, classifications and standard numbers; these units are more often recalled than worked.
Leave time for the hard ones: Model selection with likelihood-ratio tests and information criteria (3 hard of 7), Unbalanced data and Satterthwaite and Kenward-Roger degrees of freedom (3 hard of 7), Maximum likelihood and REML (3 hard of 8).
Come back to these after a first pass, untimed, and read every explanation, including for the options you rejected.
Bank the easy marks: Software: lme4, nlme, ASReml, SAS PROC MIXED (2 easy of 3); Unbalanced data and Satterthwaite and Kenward-Roger degrees of freedom (3 easy of 7).
Do these early in revision so they are fast on the day.
Finish timed. Sit the subject as a timed set without notes, mark it, and send each wrong answer back to its unit above.