UGC-NET (June 2025) · India · National Testing Agency
Original practice questions written to the published syllabus · not past papersUGC-NET (June 2025) is a PhD, fellowship and lectureship exam run by National Testing Agency, for candidates in India. CBT (since Dec 2018). Paper I 50 questions / 100 marks; Paper II 100 questions / 200 marks; 180 minutes total; +2, no negative marking.
Source: ugcnet.nta.ac.in.
Checked against the exam body's published information. Formats and fees change; confirm the current edition there before you book.
Conceptual questions on multivariate techniques, regression, digital marketing terms and metrics.
Using data, GIS, ERP and digital channels in agribusiness: the data-science workflow, basic machine learning and ICT for development.
Our 52 UGC-NET questions in this subject: 13 easy, 27 medium, 12 hard, across 4 units.
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 UGC-NET and are not shown on any other exam's page.
A demand model has errors of 4, −2, 6 and −8 tonnes on four test weeks. What is the root mean squared error (RMSE)?
A. 5.48
RMSE = √[(16 + 4 + 36 + 64)/4] = √(120/4) = √30 = 5.48 tonnes. 5.00 is the MAE (average of absolute errors 20/4), which RMSE always equals or exceeds. 0.00 is the mean error after positive and negative errors cancel. 30.00 is the MSE before taking the root.
In a logistic regression of technology adoption, the coefficient on 'member of a farmer producer organisation' (1 = yes) is 0.693. How is it interpreted?
C. Members have twice the odds of adopting compared with non-members
In logistic regression e^b is the odds ratio: e^0.693 ≈ 2.0, so membership doubles the odds. A coefficient is neither a change in probability in percentage points nor a probability itself, and it is not a share of variance explained.
A crop-disease classifier is tested on 200 fields: 40 true positives, 10 false positives, 20 false negatives and 130 true negatives. What is its recall (sensitivity)?
B. 0.67
Recall = TP/(TP + FN) = 40/(40 + 20) = 0.67. 0.80 is precision, TP/(TP + FP) = 40/50. 0.85 is accuracy, (40 + 130)/200. 0.93 is specificity, 130/140.
A crop-disease classifier is tested on 200 fields: 40 true positives, 10 false positives, 20 false negatives and 130 true negatives. What is its F1 score?
A. 0.73
Precision = 40/50 = 0.80; recall = 40/60 = 0.667. F1 = 2 × 0.80 × 0.667/(0.80 + 0.667) = 1.067/1.467 = 0.73. 0.53 is the product of precision and recall (0.80 × 0.667), not their harmonic mean. 0.85 is accuracy and 0.67 is recall alone.
To test a price-forecasting model for onions on six years of weekly data, which evaluation design is correct?
A. Fit on the first five years and test on the sixth, moving forward in time
Forecasts must be evaluated forward in time: fit on earlier periods and test on later ones. Random weeks let the model use future information next to each test week, testing on the past reverses time, and testing on the fitting data measures fit, not forecasting ability.
When is logistic regression usually preferred over linear discriminant analysis for a two-group problem?
B. When predictors include dummy variables and are clearly not multivariate normal
Logistic regression does not require normal predictors or equal covariances, so it suits mixed and categorical predictors. When LDA's assumptions hold, LDA is more efficient, and it extends naturally to several groups with a low-dimensional view; those conditions favour LDA, not logistic regression.
In discriminant output, why do analysts often interpret the structure matrix (loadings) rather than the standardised coefficients when predictors are correlated?
B. Correlated predictors make coefficients unstable or flip sign; loadings are plain correlations
When predictors are correlated, standardised coefficients share credit unpredictably and can flip sign; loadings, the correlations of each predictor with the function, are steadier for naming the function. Standardised coefficients are unit-free (that is why they are standardised), group size does not decide which output is valid, and loadings do not sum to one.
A food brand spent 50,000 on ads that generated 180,000 of attributed revenue at a 20% gross margin. Which statement is correct?
A. ROAS is 3.6, yet the ads lost money because gross profit of 36,000 is below the spend
ROAS = revenue/ad spend = 180,000/50,000 = 3.6. Gross profit = 20% × 180,000 = 36,000, which is less than the 50,000 spend, so the campaign lost 14,000 despite ROAS above 1. 130,000 treats revenue as profit; 0.72 uses profit, which is not how ROAS is defined; 0.28 inverts the ratio.
A team wants to detect an increase in conversion from 5% to 6% with a two-sided 5% test and 80% power. Using n = (1.96 + 0.84)² × [p1(1 − p1) + p2(1 − p2)]/(p2 − p1)², about how many visitors are needed per version?
D. 8,200
(1.96 + 0.84)² = 7.84. p1(1 − p1) + p2(1 − p2) = 0.0475 + 0.0564 = 0.1039. (p2 − p1)² = 0.0001. n = 7.84 × 0.1039/0.0001 ≈ 8,146, about 8,200 per version (8,155 with z = 0.8416). 816 and 82 slip the squared difference (using 0.001 or 0.01), and 81,600 slips a factor of ten the other way. Small lifts on low baselines need thousands of users per arm.
A seed company's crop-advisory videos are shared widely by farmers and covered by a regional newspaper. In the owned–paid–earned framework, this coverage is which kind of media?
D. Earned media
Shares, reviews and press coverage that others give the brand are earned media. The company's own channel is owned media, bought advertising is paid media, and programmatic refers to automated ad buying, a form of paid media.
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.
Read the manual chapters for these first and do every worked question you can find on them.
Aim for one sitting per unit, then mixed questions across them.
Come back to these after a first pass, untimed, and read every explanation, including for the options you rejected.
Do these early in revision so they are fast on the day.
Last updated · How Laisha writes and checks its pages · Report a mistake
US$10 a month · one price worldwide · cancel any time
| Free | Smarter | |
|---|---|---|
| Grow guides, pest pages, articles | Read all | Read all |
| Ask Laisha | 5 a day | Up to 1,000 a month |
| Exam practice | The sample on each page | Every question, timed papers |
| Crop calendar and pest alerts | — | For your own fields |
| Calculators and tools | — | All |
| Uploads and reports | — | 100 a month, real files |
An AI operating system for agriculture: free guides, manuals and tools written for every country.