Hypothesis tests for a mean
Studying this course? Keep your progress in one place.
Key facts
- ; is , (one-tailed) or (two-tailed). Hypotheses are about , never .
- Variance known: under .
- Critical values: (one-tailed 5%), (two-tailed 5%), (one-tailed 1%), (two-tailed 1%).
- The Central Limit Theorem lets you use the test for large from any population.
- If is unknown and is large, use in place of .
The structure of a test
State the hypotheses, find the distribution under , calculate the statistic, then conclude in context.
A hypothesis test decides whether sample data give enough evidence against a claim about a population parameter. The null hypothesis is the claim; the alternative hypothesis says what you are looking for. The significance level is the probability of rejecting when it is true.
Hypotheses are always about the population mean , never the sample mean.
- 1
State the hypotheses
, and the level
- 2
Give the distribution
Of assuming
- 3
Calculate the statistic
Compare with the critical value
- 4
Conclude
In context
Which is a correct null hypothesis for a test about a mean?
Known variance
Under , ; compare it with the critical value.
If with known, then under , and the test statistic is .
Critical values: one-tailed is and is ; two-tailed is and is .
Worked example
Lifetimes . , , , . Test at the level.
- 1
Standard error: .
- 2
.
- 3
Critical value is and .
A two-tailed test at the 1% level uses critical values:
Critical regions and p-values
Reject if the statistic is in the critical region, or if the -value is below the significance level.
The critical region is the set of values of the test statistic for which is rejected. It can be written for : for a two-tailed test with , , , reject if or .
The -value is the probability, assuming , of a result at least as extreme as observed. For a two-tailed test double the tail: gives .
Critical region
- Find critical (or ) values
- Reject if the statistic is inside
p-value
- Probability of a result at least as extreme
- Reject if p < significance level
For a two-tailed test, has tail probability 0.0122. Compare with the 5% level:
Non-Normal populations
For large the Central Limit Theorem makes approximately Normal.
The Central Limit Theorem states that if has mean and variance , then for large (roughly ) , whatever the shape of . The same test statistic and confidence intervals can then be used, approximately.
Example: skewed waiting times, , , , , . Then : insufficient evidence at .
- Large n (about 30 or more)
Waiting times are skewed with . Can you use a Normal test for the mean?
Unknown variance
With a large sample, use the unbiased estimate in place of .
When is unknown, use the unbiased estimate . If is large, treat as .
Example: , , , , . Then , and , so reject . For small samples this approximation is not reliable, so say that is large.
- 1
Mean
x̄ = 4930 ÷ 100 = 49.3
- 2
Unbiased variance
s² = (246 613 − 100 × 49.3²) ÷ 99 = 36
- 3
Test statistic
z = 1.3 ÷ (6 ÷ 10) = 2.17
- 4
Compare
2.17 > 1.96, so reject H₀ at the 5% level
- Unknown σ, large n
is unknown and . What do you do?
Try an exam question
Bottles are filled with a mean volume of ml and a known standard deviation of ml. A random sample of bottles has mean volume ml. Test, at the level, whether the mean volume has changed.
[5 marks]
- [1], .
- [1]Under , .
- [1].
- [1]Critical values are , and lies inside them.
- [1]Do not reject : insufficient evidence that the mean volume has changed from ml.
That's the notes covered.
Carry on to the next subtopic.
Try a question
Which pair of hypotheses should the consumer group use?
1 mark, marked against the mark scheme
Which pair of hypotheses should the consumer group use?
1 mark, marked against the mark scheme