QBUS5001 Chap.8 One-Population Hypothesis Testing
One-Population Hypothesis Testing
This chapter organises a hypothesis test into population claim, null and alternative statements, assumptions, statistic, evidence and contextual conclusion. It translates greater than, less than and different from into tails fixed before calculation. Type I error, Type II error and power are interpreted as business consequences. The final language is always reject or do not reject, never accept the null.
The null supplies the reference model used to judge the observed statistic, while the alternative fixes whether evidence lies in a left tail, right tail or both tails. A test statistic measures the sample result in standard-error units from the null value. The p-value is the probability, calculated under the null model, of a result at least as inconsistent with that model in the specified direction.
Alpha is a decision threshold chosen before seeing the evidence. A Type I error rejects a true null, while a Type II error fails to reject a false null; power is the chance of detecting a specified alternative. Statistical significance does not report business magnitude, so the conclusion must pair the test decision with the observed effect and its practical context.
What this chapter covers
- 01
Null and alternative hypotheses for a parameter
- 02
One-tailed and two-tailed claims
- 03
Type I error, Type II error and power
- 04
Test statistics and sampling distributions
- 05
P-values and rejection regions
- 06
Mean and proportion tests with contextual conclusions
Worked example · free
Test a population mean against a target
- +1Set H0: mu equals 25 and H1: mu is greater than 25.
- +1Compute standard error 5 divided by square root 49 equals about 0.7143.
- +1Calculate z=(26.8-25)/0.7143 equals about 2.52.
- +1Find the right-tail p-value of approximately 0.0059.
- +1Reject H0 at the 5% significance level.
- +1State sufficient evidence that the population mean exceeds 25 minutes and report the 1.8-minute estimated excess.
Key terms
- Null Hypothesis
- The equality-bearing population claim evaluated by the test.
- Alternative Hypothesis
- The directional or difference claim supported when evidence is sufficiently inconsistent with the null.
- Significance Level
- The chosen long-run probability of rejecting a true null under the test design.
- Type I Error
- Rejecting the null hypothesis when it is true.
- Statistical Power
- The probability of rejecting the null at a specified alternative where it is false.
One-Population Hypothesis Testing FAQ
How do I know which hypothesis gets equality?
Place equality in the null hypothesis. Translate the business claim into greater than, less than or different from for the alternative. Write the parameter and group or process context before substituting any sample values.
What does a p-value tell me?
Assuming the null and model conditions, it is the probability of a test statistic at least as extreme as observed in the alternative's direction. It is not the probability that the null is true and does not measure effect size.
Can I say the null is accepted?
No. Use reject or do not reject. A failure to reject may reflect a true null, a small effect, high variability or limited sample size. The correct statement is that evidence is insufficient for the alternative at the chosen level.
How are significance and business importance different?
Statistical significance compares an estimate with its sampling uncertainty. Business importance concerns the magnitude and consequence of the effect. Report the estimated difference and an interval or operational benchmark beside the test decision.
Exam move
Use a six-line template for every test: parameter, hypotheses, assumptions, statistic, evidence and conclusion. Practise translating verbal claims before seeing data so the tail cannot be chosen afterward. Write Type I and Type II errors in the scenario's language. Check the sign of the statistic against the sample difference and attach practical magnitude to every significant result. Separate planning from calculation.
Read only the claim and write the population parameter, null, alternative and tail before revealing the sample values. Then list the conditions and identify which standard error applies. After calculating the statistic, sketch its location on the reference distribution and shade the p-value region so a tail reversal becomes visible.
Practise the same scenario at two alpha levels and explain why the evidence did not change even though the decision might. Translate both error types into concrete business consequences and decide which cost matters more for the setting. Finish by writing a conclusion without the words prove, accept or definitely. Add the estimated effect size so a small p-value is not mistaken for a large operational difference.
Build one power comparison by holding alpha and the meaningful alternative fixed while increasing sample size. Explain why the sampling distribution narrows and why the alternative is detected more often. Then note the cost of data collection. This connects the formal error probabilities to an actual design choice instead of treating power as a detached formula or software output.
Working through One-Population Hypothesis Testing in QBUS5001? Sia is AskSia’s AI Statistics tutor — ask any QBUS5001 One-Population Hypothesis Testing question and get a clear, step-by-step explanation grounded in how QBUS5001 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.