What you’ll work on
- Hypothesis testing for a population mean
Start with the idea
State the null and alternative hypotheses before examining the result. For a normal population with known σ, or an appropriate large-sample approximation, standardise the sample mean as z=(xbar−μ0)/(σ/√n). Match the tail or tails to the alternative hypothesis.
The p-value is the probability, calculated under the null model, of a result at least as extreme in the direction specified by the alternative. Reject H0 when p≤α. Otherwise say there is insufficient evidence against H0; this does not prove it true.
A worked example
A normal population has known σ=8. Test H0: μ=50 against H1: μ≠50 using n=16 and xbar=54. Use α=0.05 and Φ(2)=0.97725.
- Standard error=8/4=2
- z=(54−50)/2=2
- p=2(1−0.97725)=0.0455.
Answer: z=2; two-sided p=0.0455. Reject H0 at 5%.
Try one question
Write hypotheses for a claim that a population mean has increased above 20.
Show answer and explanation
H0: μ=20; H1: μ>20.
The alternative is an upper-tail claim about the population mean.
Inside the collection
| Resource | Length | Practice | Answers |
|---|---|---|---|
| Learn and practise | 4 pages | 4 questions | Included |
| Apply and extend | 4 pages | 4 questions | Included |
| Topic checkpoint | 3 pages | 4 questions | Included |
Read on a device and work in an exercise book, or print the A4 PDFs. These files are not auto-marked digital quizzes.