t tests for a mean and the paired t-test
Studying this course? Keep your progress in one place.
Key facts
- With unknown, : symmetrical with heavier tails than .
- One-sample test: with .
- Confidence interval: , using the upper point.
- Paired data: find the differences, then do a one-sample -test on them, with = number of pairs.
- The population (or the differences) must be Normal.
The t distribution
Replace by and the statistic follows a distribution with degrees of freedom.
When and is unknown, replace by the sample standard deviation . Then follows a distribution with degrees of freedom.
It is symmetrical about but has heavier tails, more so for small , and approaches as grows. The population must be Normal. Example: the upper point of is .
A sample of size 12 is used to test a mean with unknown. How many degrees of freedom does have?
One-sample t test
Find and compare it with the critical value of .
To test , find and (for summary data, ), then . Compare with : one-tailed for or , two-tailed with the level split for .
- 1
Find and
from the data or summary statistics
- 2
Calculate
- 3
Compare
critical value of
- 4
Conclude
in context
Worked example
, , . Test against at .
- 1
.
- 2
.
- 3
.
- 4
The two-tailed critical value of is , and .
In the statistic , what is ?
Interval with unknown variance
Use , taking the upper point of .
A confidence interval for uses in place of , where is the upper point for a interval. For look up the upper point, not the point.
As before, a interval matches a two-tailed test at the level.
- Confidence interval
Worked example
, , . Find a confidence interval for .
- 1
, and the upper point of is .
- 2
.
Which tail probability do you look up for a confidence interval?
The paired t-test
Find the difference for each pair, then run a one-sample -test on the differences.
Use a paired test when each observation in one sample is linked to one in the other, such as the same people before and after. Calculate for each pair and treat the differences as one sample from . To test , use with , where is the number of pairs.
The differences must be Normal. State the direction of and write to match.
| Paired | Independent samples | |
|---|---|---|
| Data | Each value linked to one in the other sample | No link between the samples |
| Test | One-sample -test on the differences | Two-sample test of means |
| Degrees of freedom | pairs | from both sample sizes |
Paired
- Data:
- Each value linked to one in the other sample
- Test:
- One-sample -test on the differences
- Degrees of freedom:
- pairs
Independent samples
- Data:
- No link between the samples
- Test:
- Two-sample test of means
- Degrees of freedom:
- from both sample sizes
Worked example
Scores before and after give differences . Test against at .
- 1
and .
- 2
.
- 3
The upper point of is , and .
Eight people are weighed before and after a diet. How many degrees of freedom does the paired -test have?
Try an exam question
A random sample of batteries from a Normal population has mean life hours and sample standard deviation hours. Test, at the level, whether the mean life is greater than hours. The upper point of is .
[4 marks]
- [1], .
- [1].
- [1]Compares with the critical value ().
- [1]: do not reject ; insufficient evidence that the mean life exceeds hours.
That's the notes covered.
Carry on to the next subtopic.
Try a question
Under , which distribution does have?
1 mark, marked against the mark scheme
Under , which distribution does have?
1 mark, marked against the mark scheme