LAISHA
Open Laisha

GATE ST Categorical Data Analysis & Generalized Linear Models: practice questions

GATE - Statistics (ST) · India · IITs and IISc on rotation (2027: IIT Madras)

Original practice questions written to the published syllabus · not past papers
10worked here, free
74written to this exam in this subject
11syllabus units

About GATE ST

GATE - Statistics (ST) is a postgraduate admission exam run by IITs and IISc on rotation (2027: IIT Madras), for candidates in India. Computer-based, 3 hours, 100 marks (15 General Aptitude + 85 subject). MCQ, MSQ and NAT; negative marking on MCQ only; on-screen virtual calculator.

Source: gate2026.iitg.ac.in.

Checked against the exam body's published information. Formats and fees change; confirm the current edition there before you book.

Categorical Data Analysis & Generalized Linear Models in GATE ST

Models for counts, proportions and categories: contingency tables, the exponential-family GLM with its link functions, logistic and Poisson regression, log-linear models, overdispersion, ordinal outcomes and generalised linear mixed models, applied to germination, mortality, pest counts and adoption data.

Our 74 GATE ST questions in this subject: 16 easy, 45 medium, 13 hard, across 11 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 GATE ST and are not shown on any other exam's page.

Question 1 · easy

What is the deviance of a normal linear model with identity link?

  1. The total sum of squares
  2. The regression sum of squares
  3. The number of residual degrees of freedom
  4. The residual sum of squares
Answer and working

D. The residual sum of squares

For the normal, D = Σ(y − μ̂)² (times 1/σ² for the scaled deviance), so analysis of deviance generalises analysis of variance.

Question 2 · easy

A logistic model for germination is logit(p) = −1.2 + 0.8 × (seed treated). What is the odds ratio for treatment?

  1. 2.23
  2. 0.80
  3. 0.301
  4. 1.73
Answer and working

A. 2.23

exp(0.8) = 2.23. 1.73 is the ratio of probabilities (relative risk); e^(−1.2) is the baseline odds.

Question 3 · medium

In a logistic regression, at what fitted probability is the effect of a unit increase in xⱼ on p largest, and how large is it?

  1. At p = 0 or 1, about βⱼ
  2. At p = 0.5, exactly βⱼ/2
  3. At p = 0.5, about βⱼ/4
  4. At every p, exactly βⱼ
Answer and working

C. At p = 0.5, about βⱼ/4

dp/dxⱼ = βⱼp(1 − p), maximised at p = 0.5 where it equals βⱼ/4, the 'divide by 4' rule.

Question 4 · medium

Germination, emergence and establishment form sequential stages. Which ordinal model compares each stage with all later ones?

  1. The adjacent-category model
  2. The baseline-category logit
  3. The cumulative probit with equal thresholds
  4. The continuation-ratio model
Answer and working

D. The continuation-ratio model

Continuation-ratio logits model P(Y = j | Y ≥ j), natural for sequential processes; adjacent-category logits compare neighbouring categories.

Question 5 · medium

In a multinomial logit of crop choice with pigeonpea as reference, irrigation access has coefficient 0.7 in the maize equation. What does it mean?

  1. Irrigation raises the probability of maize by 70%
  2. Irrigation raises the probability of maize by 0.7
  3. Irrigated farmers have 2.0 times the odds of maize rather than pigeonpea
  4. Irrigated farmers are 2.0 times as likely to choose maize as any other crop
Answer and working

C. Irrigated farmers have 2.0 times the odds of maize rather than pigeonpea

Each equation compares a category with the reference: exp(0.7) = 2.0 is an odds ratio for maize versus pigeonpea, not a probability or a ratio against all other crops.

Question 6 · hard

Why has the Vuong test for zero-inflated versus ordinary count models been criticised?

  1. It requires normally distributed counts
  2. The null π = 0 lies on a boundary
  3. It always favours the simpler model
  4. It can only compare Poisson with binomial models
Answer and working

B. The null π = 0 lies on a boundary

The ordinary model is the zero-inflated one with π = 0, a boundary of the parameter space, so the non-nested Vuong framework does not apply cleanly.

Question 7 · medium

Which estimation method for GLMMs is fast but known to be biased with binary data and small clusters?

  1. MCMC with long chains
  2. The Laplace approximation with many observations per cluster
  3. Penalised quasi-likelihood
  4. Adaptive Gauss-Hermite quadrature
Answer and working

C. Penalised quasi-likelihood

PQL underestimates variance components and coefficients for binary data with few observations per cluster; quadrature is accurate when feasible.

Question 8 · medium

What is the most informative check that a count model handles zeros and large counts correctly?

  1. Plotting raw counts against plot number
  2. Checking that the residual deviance is below 1
  3. Comparing observed and predicted frequencies of each count value
  4. Testing normality of the raw counts
Answer and working

C. Comparing observed and predicted frequencies of each count value

Tabulating observed versus fitted frequencies of 0, 1, 2, ... directly shows excess zeros or a heavy tail that summary statistics hide.

Question 9 · medium

The squares of which residuals sum to Pearson's X² for a GLM?

  1. Raw residuals
  2. Deviance residuals
  3. Pearson residuals
  4. Randomised quantile residuals
Answer and working

C. Pearson residuals

Σ[(y − μ̂)/√V(μ̂)]² = X²; deviance residuals' squares sum to the deviance.

Question 10 · hard

Diseased and healthy plants are sampled separately (case-control) and their exposure to a practice recorded. What can a logistic regression of disease on exposure estimate?

  1. Only the relative risk
  2. Both the odds ratio and the absolute risk
  3. The odds ratio, though not the absolute risk
  4. Neither, because sampling depended on the outcome
Answer and working

C. The odds ratio, though not the absolute risk

Outcome-dependent sampling distorts the intercept but leaves slope coefficients (log odds ratios) estimable; absolute risks need the population prevalence.

A study plan for Categorical Data Analysis & Generalized Linear Models in GATE ST

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 (42% of the 74 questions): Link functions, deviance and iteratively reweighted least squares (8, 11%); Ordinal and multinomial logistic models (8, 11%); Poisson regression for counts (8, 11%); Contingency tables: association, odds ratios and tests (7, 9%).

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

  2. Then the middle band: Exponential family and the generalized linear model (7, 9%); Model checking: residuals and goodness of fit (7, 9%); Zero-inflated and hurdle models (7, 9%); Generalized linear mixed models (6, 8%).

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

  3. Then the lighter units: Overdispersion: quasi-likelihood and negative binomial models (6, 8%); Log-linear models for multi-way tables (5, 7%); Logistic regression for binary and binomial data (5, 7%).

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

  4. Leave time for the hard ones: Generalized linear mixed models (2 hard of 6), Exponential family and the generalized linear model (2 hard of 7), Model checking: residuals and goodness of fit (2 hard of 7).

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

  5. Bank the easy marks: Exponential family and the generalized linear model (3 easy of 7); Overdispersion: quasi-likelihood and negative binomial models (2 easy of 6).

    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 — GATE ST Categorical Data Analysis & Generalized Linear Models

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.