LAISHA
Open Laisha

ASRB NET Multivariate Analysis: practice questions

ASRB NET - Agricultural Statistics (and Computer Application / Bioinformatics) · India · Agricultural Scientists Recruitment Board (ICAR)

Original practice questions written to the published syllabus · not past papers
10worked here, free
58written to this exam in this subject
12syllabus units

About ASRB NET

ASRB NET - Agricultural Statistics (and Computer Application / Bioinformatics) is a professional certification run by Agricultural Scientists Recruitment Board (ICAR), for candidates in India. NET: one objective paper of about 150 questions, 2 hours, negative marking, online, Master's level (pattern as reported for other disciplines).

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.

Multivariate Analysis in ASRB NET

Mahalanobis D², Tocher clustering, inter- and intra-cluster distances and choice of divergent parents.

Analysing many correlated variables together: the multivariate normal and its estimation, Wishart and Hotelling's T-squared, MANOVA, and the reduction and classification tools agricultural scientists use daily (principal components, factor analysis, discriminant and cluster analysis, Mahalanobis D-squared).

Our 58 ASRB NET questions in this subject: 15 easy, 39 medium, 4 hard, across 12 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 NET and are not shown on any other exam's page.

Question 1 · medium

Yield (kg/ha), plant height (cm) and protein (%) are analysed together. Why is the correlation matrix often preferred to the covariance matrix?

  1. It needs fewer observations
  2. It removes the effect of units
  3. It is unbiased while S is biased
  4. It always has larger eigenvalues
Answer and working

B. It removes the effect of units

Covariance-based methods are dominated by variables with large numerical variances, which depend on units; standardising puts traits on an equal footing. Unbiasedness and sample-size requirements are not the reason.

Question 2 · medium

A genetic correlation estimated from genotype and error mean-product matrices in a small trial comes out as 1.15. What is the most likely explanation?

  1. Sampling error in a small trial
  2. Genetic correlations can exceed 1 when heritabilities are high
  3. The error mean products were divided by the wrong df
  4. The two traits are controlled by exactly the same genes
Answer and working

A. Sampling error in a small trial

The difference of two Wishart-type matrices need not be positive definite in small samples, so estimated genetic correlations can fall outside [−1, 1]. True correlations cannot; constrained REML or Bayesian estimation avoids such values.

Question 3 · medium

Plant height is measured on the same plants at 4 dates under two treatments. Which method tests whether the mean growth curves are parallel?

  1. Four separate t-tests, one for each date
  2. A paired t-test on the final date only
  3. Profile analysis with T²
  4. A chi-square test of the four means
Answer and working

C. Profile analysis with T²

Parallelism means the vectors of successive differences have equal means in the two groups, tested by a two-sample T² on those p − 1 contrasts; separate t-tests ignore correlation and multiplicity.

Question 4 · medium

Which test checks the MANOVA assumption of equal covariance matrices across groups, and what is its weakness?

  1. Levene's test, which applies only to one variable
  2. Bartlett's chi-square for Λ, which needs equal n
  3. Box's M, which is sensitive to non-normality
  4. Mauchly's test, which requires p = 1
Answer and working

C. Box's M, which is sensitive to non-normality

Box's M generalises Bartlett's test of equal variances; like it, it reacts strongly to non-normality, so a significant M may reflect heavy tails rather than unequal covariances.

Question 5 · medium

Why is one MANOVA on five correlated quality traits preferred to five separate ANOVAs at α = 0.05?

  1. It controls overall error and uses the trait correlations
  2. It needs no assumption of normality or equal dispersion
  3. It always finds more significant traits than separate ANOVAs
  4. It gives the same result with fewer plots
Answer and working

A. It controls overall error and uses the trait correlations

Separate ANOVAs inflate the family-wise error rate and ignore correlations; MANOVA tests the vector of means jointly and can detect differences in combinations of traits. It does assume multivariate normality and equal covariance matrices.

Question 6 · easy

A correlation-matrix PCA of five traits gives eigenvalues 2.6, 1.2, 0.6, 0.4 and 0.2. What proportion of total variance do the first two components explain?

  1. 19%
  2. 52%
  3. 86.4%
  4. 76%
Answer and working

D. 76%

For a correlation matrix the eigenvalues sum to p = 5, so (2.6 + 1.2)/5 = 76%. 52% counts only the first component; 86.4% divides by the first three eigenvalues instead of the total.

Question 7 · easy

For a PCA of the correlation matrix, what does the Kaiser criterion retain?

  1. Components with eigenvalue greater than 1
  2. Components explaining at least 5% each
  3. Components with negative loadings
  4. The first two components always
Answer and working

A. Components with eigenvalue greater than 1

Each standardised variable contributes variance 1, so a component with λ > 1 explains more than one original variable's worth. It is a rule of thumb; scree plots and parallel analysis are alternatives.

Question 8 · medium

Which method applies principal component analysis to the genotype-by-environment interaction matrix after removing main effects?

  1. Tocher's clustering
  2. Fisher's discriminant function
  3. The AMMI model
  4. Box's M test
Answer and working

C. The AMMI model

AMMI (additive main effects and multiplicative interaction) fits additive genotype and environment effects by ANOVA and the interaction by PCA (singular value decomposition); GGE biplots apply it to genotype plus G × E.

Question 9 · easy

A standardised variable has loadings 0.8 and 0.3 on two orthogonal factors. What is its communality?

  1. 0.27
  2. 0.64
  3. 0.73
  4. 1.10
Answer and working

C. 0.73

h² = 0.8² + 0.3² = 0.64 + 0.09 = 0.73; its uniqueness is 1 − 0.73 = 0.27. Adding loadings without squaring gives 1.10.

Question 10 · medium

In Rao's D² analysis, which clustering method is classically used to group genotypes?

  1. Tocher's method
  2. Single linkage
  3. k-means
  4. Ward's method
Answer and working

A. Tocher's method

Tocher's method adds genotypes to a cluster while the average intra-cluster distance stays below a threshold (the largest of the minimum distances), as described by Rao (1952).

A study plan for Multivariate Analysis in ASRB 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 (40% of the 58 questions): Hotelling's T-squared and its applications (6, 10%); Mahalanobis D-squared and genetic divergence (6, 10%); Multivariate analysis of variance (MANOVA) (6, 10%); Cluster analysis: hierarchical and k-means (5, 9%).

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

  2. Then the middle band: Discriminant analysis and classification (5, 9%); Estimation of the mean vector and dispersion matrix (5, 9%); Factor analysis (5, 9%); Multidimensional scaling and correspondence analysis (5, 9%).

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

  3. Then the lighter units: Multivariate normal distribution: properties, marginals and conditionals (5, 9%); Principal component analysis (5, 9%); Canonical correlation analysis (3, 5%); Wishart distribution (2, 3%).

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

  4. Leave time for the hard ones: Estimation of the mean vector and dispersion matrix (1 hard of 5), Factor analysis (1 hard of 5), Multivariate normal distribution: properties, marginals and conditionals (1 hard of 5).

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

  5. Bank the easy marks: Wishart distribution (1 easy of 2); Cluster analysis: hierarchical and k-means (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 — ASRB NET Multivariate 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.