MTH645 Probability Theory & Sampling Guide | Virtual University Pakistan

The MTH645 handout from Virtual University of Pakistan opens with probability axioms that form the bedrock of statistical inference. You'll encounter classical, relative frequency, and subjective probability approaches — but VU exams heavily favor the axiomatic approach with Kolmogorov's three axioms. In the exam you will see questions asking you to prove basic properties like P(A ∪ B) = P(A) + P(B) - P(A ∩ B) using only these axioms. The 2026 handout emphasizes conditional probability and independence with concrete examples like medical testing scenarios — exactly the type that appears in Question 2 of every midterm. Don't just memorize formulas; understand why P(A|B) = P(A∩B)/P(B) makes intuitive sense when you think about reduced sample spaces.

Bayes' theorem is the crown jewel of MTH645 — Virtual University of Pakistan examiners love it because it tests both conceptual understanding and algebraic manipulation. The 2026 handout presents the classic disease screening example: a test with 99% sensitivity and 99% specificity for a disease affecting 0.1% of population. In the exam you will calculate positive predictive value and be shocked by the low result — this exact numerical problem appears in 80% of VU finals. Master the tree diagram approach; it's faster than formula plugging and reduces arithmetic errors under exam pressure. The handout also covers Bayesian updating with multiple tests — a favorite for 10-mark long questions in final term papers.

The MTH645 handout dedicates three lectures to distributions — Binomial, Poisson, Normal, and Exponential. Virtual University of Pakistan's exam pattern shows clear preferences: Binomial for fixed trials with success/failure, Poisson for rare events over time/space, Normal for everything else thanks to Central Limit Theorem. In the exam you will see questions like 'If X ~ Binomial(50, 0.2), approximate P(X ≤ 8) using normal approximation' — this tests both distribution knowledge and continuity correction. The 2026 handout emphasizes when approximations are valid (np>5, nq>5) because VU graders deduct marks for blind approximation. Memorize the Poisson approximation to Binomial conditions too — they're easy marks if you know the rule of thumb.

Sampling theory forms the bridge to inference — Virtual University of Pakistan's MTH645 handout covers simple random, stratified, systematic, and cluster sampling with Pakistani context examples like agricultural yield surveys. The sampling distribution of sample mean is crucial; the 2026 handout derives its mean and variance for both finite and infinite populations. In the exam you will see finite population correction factor questions — they appear when sample size exceeds 5% of population. The Central Limit Theorem gets heavy coverage because it justifies normal-based inference. VU finals often ask: 'Explain why CLT allows us to use normal distribution for sample means from non-normal populations' — answer with 'sample size ≥ 30' and 'i.i.d. samples' for full marks.

The MTH645 handout transitions to statistical inference with point estimators properties — unbiasedness, consistency, efficiency. Virtual University of Pakistan examiners test these concepts through multiple choice questions in midterms. Interval estimation gets more weight: confidence intervals for mean (known/unknown variance), proportion, variance, and difference of means. The 2026 handout uses t-distribution for unknown variance with n<30 — know your degrees of freedom. In the exam you will construct 95% CI for population mean given sample data — this 5-mark question appears in every final term. Interpretation matters too: 'We are 95% confident the true mean lies in this interval' not 'probability is 0.95'. VU graders are strict on frequentist interpretation.

Hypothesis testing is the heart of MTH645 — Virtual University of Pakistan's 2026 handout presents the 7-step framework that every VU student must memorize: state hypotheses, choose significance level, select test statistic, determine critical region, calculate test statistic, make decision, interpret result. The exam pattern shows consistent testing of one-sample mean (z/t), proportion, variance, and two-sample comparisons. In the exam you will see both critical value and p-value approaches — know both. Type I/II error trade-offs appear as conceptual questions. The handout includes Pakistani context examples like testing if new fertilizer increases wheat yield — these make abstract concepts concrete. Practice the 7 steps until they're automatic; time pressure makes students skip steps and lose marks.

The final modules of MTH645 handout cover ANOVA for comparing multiple means and Chi-square tests for independence and goodness-of-fit. Virtual University of Pakistan includes these because they're essential for research methodology courses later. One-way ANOVA table completion questions appear in 60% of finals — know SSBetween, SSWithin, MST, MSE, F-statistic calculation. Chi-square contingency table tests use Pakistani census data examples in the 2026 handout. In the exam you will check assumptions: expected frequencies ≥ 5 for Chi-square, normality and equal variances for ANOVA. These assumption checks are easy marks many students miss. The handout also mentions non-parametric alternatives — know when to use Kruskal-Wallis instead of ANOVA.

Success in MTH645 requires strategic use of Virtual University of Pakistan resources. The LMS hosts complete handouts, video lectures, and past papers — the 2026 semester handout you have is your primary reference. Start with past papers from last 3 years; VU repeats questions with numerical changes. Form a study group to explain concepts — teaching Bayes' theorem to peers reveals gaps. The graded discussion boards (GDB) often preview exam topics; participate actively. Assignments test application; don't just copy solutions — understand each step. Midterm covers probability and distributions; final is comprehensive with emphasis on sampling and inference. Time management: allocate 1.5 minutes per mark. Most importantly, sleep before exam — probability questions need clear thinking.