Distributions and Statistical Inference
CSIR NET Mathematical Sciences Unit 4: standard distributions, sampling theory, estimation, MP/UMP and LR tests, chi-square, nonparametric and Bayesian methods.
- 6
- Topics
- 113
- Min read
- 01 Standard Discrete and Continuous DistributionsStart here CSIR NET statistics: binomial, Poisson, geometric, negative binomial, hypergeometric, normal, exponential, gamma, beta, Cauchy; moments, MGFs and links.
- 02 Sampling Distributions, Standard Errors and Order Statistics CSIR NET statistics: chi-square, t and F laws, normal sampling theory, standard errors, CLT and delta method, sample quantiles and order statistics.
- 03 Point Estimation and Confidence Intervals CSIR NET statistics: unbiasedness, consistency, sufficiency, completeness, Cramér–Rao, Rao–Blackwell, UMVUE, MLE and moments, and confidence intervals.
- 04 Hypothesis Testing: MP, UMP and Likelihood Ratio Tests CSIR NET statistics: errors, size and power, Neyman–Pearson lemma, randomised MP tests, MLR and Karlin–Rubin UMP tests, likelihood ratio tests, Wilks.
- 05 Chi-Square Goodness of Fit and Large-Sample Tests CSIR NET statistics: z-tests for means, proportions and correlations; Pearson chi-square goodness of fit, contingency tables, degrees of freedom and pitfalls.
- 06 Nonparametric Tests, Rank Correlation and Bayesian Inference CSIR NET statistics: sign, Wilcoxon, Mann–Whitney, runs, Kruskal–Wallis and KS tests; Spearman and Kendall correlation; conjugate priors, Bayes estimators.
- Distributions and Statistical Inference Formula Sheet and Short Notes Revise The whole chapter on one page.
Chapter Overview
This chapter is the core of classical statistics in Unit 4 of the CSIR NET Mathematical Sciences syllabus: the standard probability laws, the distributions of statistics computed from samples, and the two inferential tasks built on them, estimation and testing. It closes with the methods that drop distributional assumptions (rank and sign tests) and with the Bayesian alternative to the whole framework.
These ideas underlie much of applied statistics: a regression or ANOVA question is a \( t \) or \( F \) test underneath, and a sampling-theory question is a standard-error computation. The notes assume a first MSc course in statistical inference (at the level of Casella and Berger’s Statistical Inference or Rohatgi and Saleh’s An Introduction to Probability and Statistics). They do not re-teach that course. They collect the exact statements, the standard computations, and the counter-examples (the uniform family, the Cauchy distribution, discrete tests that need randomisation) that decide multiple-correct Part C questions.
What You’ll Learn
By the end of this chapter you will be able to:
- Recall the mean, variance and MGF of every standard discrete and continuous law, and move between them through sums, limits and transformations
- Identify chi-square, \( t \) and \( F \) statistics, use the normal sampling theorem, and compute standard errors and delta-method variances
- Work with order statistics: densities, uniform and exponential special cases, a distribution-free interval for the median
- Find MLEs, moment estimators, sufficient and complete statistics, UMVUEs and Cramér–Rao bounds, and build confidence intervals from pivots
- Construct most powerful tests by the Neyman–Pearson lemma (with randomisation), UMP tests by Karlin–Rubin, and likelihood ratio tests with Wilks’ theorem
- Run large-sample \( z \)-tests, chi-square goodness-of-fit and contingency-table tests
- Apply sign, signed-rank, Mann–Whitney, runs, Kruskal–Wallis and Kolmogorov–Smirnov tests; compute Spearman’s and Kendall’s rank correlations
- Update conjugate priors, and find Bayes estimators and credible intervals
Prerequisites
- The probability chapter of Unit 4: random variables, expectation, moments, convergence and the CLT. Its first page, Descriptive Statistics and Exploratory Data Analysis, sets up the sample summaries used here.
- Counting from Combinatorics and Counting, for binomial-type laws and rank statistics.
- Quadratic Forms and orthogonal matrices, for the geometry behind \( \chi^2 \) degrees of freedom.
- Improper integrals from Riemann and Lebesgue Integration, for the gamma and beta functions.
Learning Path
graph TD
A["1. Standard Discrete and Continuous Distributions"] --> B["2. Sampling Distributions, Standard Errors and Order Statistics"]
B --> C["3. Point Estimation and Confidence Intervals"]
C --> D["4. Hypothesis Testing: MP, UMP and LR Tests"]
B --> E["5. Chi-Square Goodness of Fit and Large-Sample Tests"]
D --> E
D --> F["6. Nonparametric Tests, Rank Correlation and Bayesian Inference"]
C --> F
E --> G["Quick Revision"]
F --> GTopics in This Chapter
| # | Topic | What it covers | Time |
|---|---|---|---|
| 1 | Standard Discrete and Continuous Distributions | Bernoulli to hypergeometric; uniform, normal, exponential, gamma, beta, Cauchy, lognormal; MGFs, additivity, limits, transformations | 35 min |
| 2 | Sampling Distributions, Standard Errors and Order Statistics | \( \chi^2 \), \( t \), \( F \); normal sampling theorem; standard errors; CLT, Slutsky, delta method; sample quantiles; order statistics | 35 min |
| 3 | Point Estimation and Confidence Intervals | unbiasedness, consistency, efficiency; sufficiency, completeness; Rao–Blackwell, Lehmann–Scheffé; Cramér–Rao; moments and MLE; pivots and intervals | 40 min |
| 4 | Hypothesis Testing: MP, UMP and Likelihood Ratio Tests | size, power, p-values; Neyman–Pearson lemma and randomisation; MLR and Karlin–Rubin; when no UMP test exists; LRT and Wilks | 40 min |
| 5 | Chi-Square Goodness of Fit and Large-Sample Tests | \( z \)-tests for means, proportions, correlations; exact small-sample tests; Pearson’s \( \chi^2 \); contingency tables, \( 2 \times 2 \) shortcut | 30 min |
| 6 | Nonparametric Tests, Rank Correlation and Bayesian Inference | sign, signed-rank, Mann–Whitney, runs, Kruskal–Wallis, KS; Spearman and Kendall; priors, conjugacy, Bayes estimators, credible intervals | 40 min |
| Quick Revision | every formula and trap of the chapter on one page | 10 min |
Official syllabus lines and where they are covered
| Topic line (editorial split of Unit 4) | Covered in |
|---|---|
| Standard discrete and continuous distributions | Standard Discrete and Continuous Distributions |
| Sampling distributions, standard errors, asymptotics, order statistics | Sampling Distributions, Standard Errors and Order Statistics |
| Estimation methods and properties, confidence intervals | Point Estimation and Confidence Intervals |
| MP/UMP and likelihood ratio tests | Hypothesis Testing: MP, UMP and Likelihood Ratio Tests |
| Chi-square goodness of fit, large-sample tests | Chi-Square Goodness of Fit and Large-Sample Tests |
| Nonparametric tests, rank correlation | Nonparametric Tests, Rank Correlation and Bayesian Inference (first half) |
| Elementary Bayesian inference | Nonparametric Tests, Rank Correlation and Bayesian Inference (second half) |
How CSIR NET Asks It
Per the NTA information bulletin for December 2026, the Mathematical Sciences paper (code 704) is marked as follows:
| Part B | Part C | |
|---|---|---|
| Questions given | 40 | 60 |
| Questions to attempt | 25 | 20 |
| Marks per correct answer | 3 | 4.75 |
| Marks per wrong answer | −0.75 | 0 |
| Maximum | 75 | 95 |
NTA publishes no topic-wise weightage, so we make no claim about how many questions come from this chapter. The format tells you how to prepare:
- Part B (single correct) rewards fast, exact computation: a parameter from two moments, \( P(X > s + t \mid X > s) \), the law of a ratio of normal variables, \( E\,X_{(r)} \) for a uniform sample, an MLE or UMVUE, a Cramér–Rao bound, a critical value or the size of a given region, a \( \chi^2 \) value or its degrees of freedom, Spearman’s \( r_s \), a conjugate posterior mean.
- Part C (one or more correct) rewards knowing where the theorems stop. Stems fix a model ("\( X_1, \dots, X_n \) i.i.d. \( U(0, \theta) \)", “one observation from a Poisson”, “a normal sample with both parameters unknown”) and list four claims about unbiasedness, sufficiency, efficiency, UMP existence, independence or asymptotic laws. Each is settled by a short computation or a standard counter-example. Every topic page ends with Level 3 practice in exactly this format, with every option justified.
For the marking scheme and the other units, see the Mathematical Sciences overview. Part A has its own light statistics in Histograms, Distributions, Scatter and Box Plots and Experimental Design, Sampling and Data. Found an error? See About for how to report it.
Start Learning
Begin with the first topic: