Quick Answer
Hypothesis testing starts from a null assumed true until evidence rejects it. The significance level α caps the Type-I error (rejecting a true null); Type-II error is failing to reject a false null. The p-value measures the probability of data at least this extreme given the null — never the probability that the null is true.
Visual Summary
- H₀ presumed — evidence must embarrass it to reject.
- Type I — rejecting a true null; capped by α.
- Type II — keeping a false null; power = 1 − β.
- p-value — probability of the DATA given H₀.
The Two Errors
Type-I error (α): rejecting a true null — its ceiling is the significance level chosen before the test (commonly 0.05). Type-II error (β): accepting a false null; its complement is POWER — the test’s ability to detect a real effect, which grows with sample size and effect size. Reducing α raises β for fixed n; only larger samples improve both.
The Test Machinery
Choose the statistic by data type and size: z for large samples or known σ; t for small samples with unknown σ; chi-square for categorical frequencies (goodness of fit, independence); F for variance ratios (ANOVA). Compute the statistic, compare against the critical value or convert to a p-value: reject H₀ when p < α. Directional hypotheses use one-tailed tests; the two-tailed test splits α across both tails.
Common Misreadings
The p-value is NOT the probability that H₀ is true, and failing to reject does not prove H₀. Statistical significance is not practical significance — with huge samples, trivial differences become significant.
Exam Angle
NET passages stage a decision and ask which error occurred, or misread the p-value. Master the two-error grid. Practise in the UGC NET Commerce mock.
FAQ
Does failing to reject prove the null?
No — it only means the evidence was insufficient against it.
What raises test power?
Larger samples, bigger true effects, and one-tailed directional tests.
Related reading
- Sampling Designs in Research Methodology: Probability and Non-Probability Types
- CA Foundation: The Four-Lane Revision Map
Exam angle
How exams use this: CA/Foundation papers mix recall, calculation and application in one sitting — train all three under +1/−0.25. Test it under exam conditions →
Quick revision
- H₀ presumed: — evidence must embarrass it to reject.
- Type I: — rejecting a true null; capped by α.
- Type II: — keeping a false null; power = 1 − β.
- p-value: — probability of the DATA given H₀.
- CA Foundation: The Four-Lane Revision Map
Have a doubt on this topic?







