LAISHA
Open Laisha

ASRB Agricultural Research Service Multivariate 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
27written to this exam in this subject
12syllabus 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.

Multivariate Analysis in ASRB Agricultural Research Service

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 27 ASRB Agricultural Research Service questions in this subject: 0 easy, 25 medium, 2 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 Agricultural Research Service and are not shown on any other exam's page.

Question 1 · medium

Two seed treatments are compared on germination, vigour and root length, measured on the same seed lots. How should the multivariate comparison be made?

  1. MANOVA with lots ignored
  2. Two-sample T² treating the lots as independent
  3. T² on within-lot differences
  4. Three paired t-tests at α = 0.05 each
Answer and working

C. T² on within-lot differences

Paired multivariate data are analysed as differences: T² = n d̄′S_d⁻¹d̄ with (p, n − p) df after the F transformation. Treating paired data as independent ignores the lot effect.

Question 2 · medium

Which MANOVA statistic is often recommended as the most robust to violations of assumptions such as unequal covariance matrices?

  1. Box's M
  2. Pillai's trace
  3. Hotelling's one-sample T²
  4. Roy's largest root
Answer and working

B. Pillai's trace

Pillai's trace is the most robust of the four; Roy's root is most powerful when means differ along one direction but least robust. Box's M tests equality of covariance matrices.

Question 3 · medium

In a scree plot, which components are usually retained?

  1. Those after the eigenvalues level off
  2. All components with positive eigenvalues
  3. Those before the elbow
  4. Only components with eigenvalues below 1
Answer and working

C. Those before the elbow

The 'elbow' separates large, informative eigenvalues from a flat tail of small ones attributed to noise.

Question 4 · medium

Fifteen survey items on farmers' attitudes are reduced to three latent constructs such as risk aversion and openness to innovation. Which method suits this aim?

  1. Canonical correlation analysis
  2. Hotelling's T²
  3. k-means clustering of the items
  4. Factor analysis
Answer and working

D. Factor analysis

Explaining correlations among many items by a few latent constructs is the factor-analysis model; confirmatory FA or structural equation models test a hypothesised structure.

Question 5 · medium

A first canonical correlation of 0.85 is found between soil and crop sets. Why might it overstate how much the sets share?

  1. It is a squared quantity, so its square root should be reported
  2. Redundancy indices, not ρ₁, measure shared variance
  3. Canonical correlations are always biased towards zero
  4. It must be divided by the number of variables in each set
Answer and working

B. Redundancy indices, not ρ₁, measure shared variance

A high ρ₁ can reflect one strongly related pair while most variance in each set is unrelated; redundancy indices measure the proportion of one set's variance explained by the other's canonical variates.

Question 6 · medium

Why are structure coefficients usually interpreted instead of raw canonical coefficients?

  1. Raw coefficients depend on the sample size only
  2. Raw weights are unstable under within-set collinearity
  3. Structure coefficients are unbiased estimates of the ρᵢ
  4. Raw coefficients are always zero for some variables
Answer and working

B. Raw weights are unstable under within-set collinearity

With multicollinearity, raw weights can have unexpected signs and large sampling variability; correlations between variables and canonical variates (structure coefficients) are more stable.

Question 7 · medium

A linear discriminant score for seed lots has mean 12 for variety A and 8 for variety B. With equal priors and costs, a new lot scores 9.4. How is it classified?

  1. It cannot be classified without the dispersion matrix
  2. As variety B, since 9.4 is below the midpoint 10
  3. As variety A, since 9.4 is above the B mean 8
  4. As variety A, since the score is positive
Answer and working

B. As variety B, since 9.4 is below the midpoint 10

The cut-off is the midpoint (12 + 8)/2 = 10; 9.4 < 10, so the lot is assigned to B. The dispersion matrix is already built into the discriminant score.

Question 8 · medium

Diseased and healthy plants are to be classified from reflectance at several wavelengths plus a categorical cultivar code. Which method avoids assuming multivariate normal predictors?

  1. Logistic regression
  2. Factor analysis
  3. Fisher's LDA with pooled S
  4. Hotelling's T²
Answer and working

A. Logistic regression

Logistic regression models P(diseased | x) directly and handles categorical predictors; LDA is more efficient when normality holds but is sensitive when it does not.

Question 9 · medium

Molecular-marker profiles (presence or absence of bands) are used to cluster accessions. Which dissimilarity basis is appropriate?

  1. A similarity coefficient such as Jaccard, converted to distance
  2. Euclidean distance on standardised band intensities
  3. Mahalanobis distance on the 0/1 codes
  4. Correlation between accessions over time
Answer and working

A. A similarity coefficient such as Jaccard, converted to distance

For binary data, matching or Jaccard coefficients (which ignore shared absences) are standard; the resulting distances are clustered, typically by UPGMA.

Question 10 · medium

Farm-household survey data contain many categorical variables (crop type, irrigation source, tenure, credit access). Which method maps all their categories together?

  1. Hotelling's T²
  2. Multiple correspondence analysis
  3. Principal component analysis on the category codes
  4. Canonical correlation between two codes
Answer and working

B. Multiple correspondence analysis

MCA extends correspondence analysis to several categorical variables via the indicator or Burt matrix; PCA on arbitrary category codes treats nominal codes as numbers.

A study plan for Multivariate 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 (48% of the 27 questions): Estimation of the mean vector and dispersion matrix (4, 15%); Discriminant analysis and classification (3, 11%); Multidimensional scaling and correspondence analysis (3, 11%); Principal component analysis (3, 11%).

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

  2. Then the middle band: Canonical correlation analysis (2, 7%); Cluster analysis: hierarchical and k-means (2, 7%); Factor analysis (2, 7%); Hotelling's T-squared and its applications (2, 7%).

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

  3. Then the lighter units: Mahalanobis D-squared and genetic divergence (2, 7%); Multivariate analysis of variance (MANOVA) (2, 7%); Multivariate normal distribution: properties, marginals and conditionals (1, 4%); Wishart distribution (1, 4%).

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

  4. Leave time for the hard ones: Factor analysis (1 hard of 2), Estimation of the mean vector and dispersion matrix (1 hard of 4).

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

  5. 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 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.