Hypothesis Tests for Differences in Matched Pairs (College Board AP® Statistics)
Study Guide
Written by: Naomi C
Reviewed by: Dan Finlay
Paired t-test
What is a paired t-test?
A paired t-test is used to test whether or not the population means of two pieces of data that are linked are equal by examining the differences between paired data
A paired t-test is easily confused with a two-sample t-test
but the data for a two-sample t-test must come from two independent populations
e.g. the mean lifespan of two different brands of battery
whereas the data for a paired t-test is linked and comes from one population
e.g. the result of a blood test for a group of people before a medication is given and the result after it is given
To try to prove your case, you take two measures from a recent random sample of size from a single population and run a one-sample test on the differences between the two values
e.g. you take a random sample of 30 students and record their test scores in both Science and Math, then run a one-sample t-test on the differences between each data pair
What are the hypotheses for a paired t-test?
The null hypothesis, , is the assumption that mean of the differences between the two measures is 0
e.g. The mean of the differences between the test score for Science and the test score for Math is 0,
It is assumed to be correct, unless evidence proves otherwise
The alternative hypothesis, , is what you think the mean of the differences is relative to 0, i.e. if it is positive or negative
e.g. The mean of the differences between the test score for Science and the test score for Math is less than 0,
Remember that a t-test could be one-tailed or two-tailed, this will affect your alternative hypothesis
Examiner Tips and Tricks
When writing out your hypotheses, always fully define the symbol used for the population parameters in context, e.g. '... where is the mean of the differences between the test score for Science and the test score for Math, and '.
What are the conditions for a paired t-test?
When performing a paired t-test, you must show that it meets the following conditions:
The two measures should come from the same items within the population
Items in the sample (or experiment) must satisfy the independence condition
by verifying that data is collected by random sampling
or random assignment (in an experiment)
and, if sampling without replacement, showing that the sample size is less than 10% of the population size
The distribution of the differences is approximately normally distributed
The shape needs to be approximately symmetric
There should be no outliers
If the distribution of the differences is very skewed, you can only do a t-test if
How do I calculate the standardized test statistic (t-value)?
The t-value, for the paired t-test is given by:
where is the mean of the difference between each pair of data in the sample, is the standard deviation of the differences in the sample, and is the sample size
Examiner Tips and Tricks
The formula for the standardized test statistic is given in the exam, , along with tables of parameters and standard errors.
You will need to apply this correctly to get the t-value. In the numerator, the statistic is the mean of the pair differences in the sample and the parameter is the mean of the pair differences in the population, which is 0 for the null hypothesis.
How do I calculate the p-value?
Work out the t-value
Find the appropriate number of degrees of freedom ('dof')
For a paired t-test,
Using the t-distribution table given to you:
find the row that corresponds to the dof
identify the t-value in the row that is closest to the calculated value
write down the value in the corresponding column header
this is the p-value
Note that the p-value from the t-table is for one tail
How do I conclude a hypothesis test?
Conclusions to a hypothesis test need to show two things:
a decision about the null hypothesis
an interpretation of this decision in the context of the question
To make the decision, compare the p-value to the significance level,
If then the null hypothesis should be rejected
If then the null hypothesis should not be rejected
In a two-tailed test, double the p-value and compare this to
This is because there are two regions that are as extreme, or more extreme, than the one observed in the sample
Examiner Tips and Tricks
Remember that the test should be interpreted within the context of the question.
Use the same language in your conclusion that is used in the problem, e.g. 'The data provides sufficient evidence that the mean of the differences between the test score for Science and the test score for Math is less than 0, so students perform better in Math than Science'.
What are the steps for performing a paired t-test on a calculator?
When using a calculator to conduct a paired t-test, you must still write down all steps of the hypothesis testing process:
State the null and alternative hypotheses and clearly define your parameters
Describe the test being used and show that the situation meets the conditions required
Calculate the t-value and the degrees of freedom
Calculate the p-value using your calculator
select a one-sample t-test and enter the relevant summary statistics or data to generate the p-value for the differences
Compare the p-value to the significance level
Write down the conclusion to the test and interpret it in the context of the problem
Examiner Tips and Tricks
Even if you perform the paired t-test on your calculator, it is still important to show all of your working to demonstrate full understanding. Therefore you should still show workings for calculating the t-value and the degrees of freedom.
Worked Example
A research laboratory are conducting trials to investigate the effect of a new medication on blood pressure. It is believed that the medication will effectively lower blood pressure for people that currently have high blood pressure. 9 volunteers with a history of high blood pressure are selected at random from a pool of 100 and their systolic blood pressure is measured. They are then given a dose of the new medication and their systolic blood pressure is again measured 30 minutes after receiving it. You may assume that the differences in systolic blood pressure before and after the medication follow a normal distribution. The results of the trial are shown in the table below.
Volunteer | Systolic blood pressure before (mmgH) | Systolic blood pressure after (mmgH) |
---|---|---|
1 | 156 | 151 |
2 | 162 | 158 |
3 | 149 | 147 |
4 | 168 | 165 |
5 | 153 | 152 |
6 | 165 | 161 |
7 | 159 | 156 |
8 | 151 | 151 |
9 | 164 | 158 |
Perform an appropriate hypothesis test at the 1% significance level. Is the medication effective at reducing high blood pressure?
Method 1: Using the t-table
State the type of test being used and verify the conditions for the test
The correct inference procedure is a paired t-test with
Both pieces of data were taken from each volunteer
The independence condition is satisfied, as
the sample of 9 volunteers was selected at random
and the sample size, 9, is less than 10% of the population, 100 (9 < 10)
It is assumed that the differences in systolic blood pressure before and after the medication follow a normal distribution
The sample size is small (, which is less than 30) and the population standard deviation is unknown, so the t-distribution can be used
Define the population parameter,
Let be the mean of the differences between the initial blood pressure of a volunteer and their blood pressure after taking medication, where
Write the null and alternative hypotheses
This will be a one-tailed test as it is believed that the medication will effectively lower blood pressure, if blood pressure is lowered then blood pressure before - blood pressure after will be greater than 0
Add an additional column to the table to summarize the differences of the blood pressure results
Volunteer | Systolic blood pressure before (mmgH) | Systolic blood pressure after (mmgH) | Difference in blood pressure (mmgH) |
---|---|---|---|
1 | 156 | 151 | 5 |
2 | 162 | 158 | 4 |
3 | 149 | 147 | 2 |
4 | 168 | 165 | 3 |
5 | 153 | 152 | 1 |
6 | 165 | 161 | 4 |
7 | 159 | 156 | 3 |
8 | 151 | 151 | 0 |
9 | 164 | 158 | 6 |
Use your calculator, or work out by hand, the mean and the standard deviation of the differences
Calculate the standardized test statistic
Calculate the number of degrees of freedom
degrees of freedom = 9 - 1 = 8
Find the p-value from the t-tables
Find the row corresponding to 8 degrees of freedom and identify the t-value that is closest to the absolute calculated t-value (4.911)
closest t-value = 5.041
corresponding p-value is
Compare the probability to the significance level and state the conclusion of the test
is rejected
Interpret the result in the context of the question
We have sufficient evidence to support the belief that the medication will reduce blood pressure in those that currently have high blood pressure
Method 2: Using a calculator
State the type of test being used and verify the conditions for the test
The correct inference procedure is a paired t-test with
Both pieces of data were taken from each volunteer
The independence condition is satisfied, as
the sample of 9 volunteers was selected at random
and the sample size, 9, is less than 10% of the population, 100 (9 < 10)
It is assumed that the differences in systolic blood pressure before and after the medication follow a normal distribution
The sample size is small (, which is less than 30) and the population standard deviation is unknown, so the t-distribution can be used
Define the population parameter,
Let be the mean of the differences between the initial blood pressure of a volunteer and their blood pressure after taking medication, where
Write the null and alternative hypotheses
This will be a one-tailed test as it is believed that the medication will effectively lower blood pressure, if blood pressure is lowered then blood pressure before - blood pressure after will be greater than 0
Add an additional column to the table to summarize the differences of the blood pressure results
Volunteer | Systolic blood pressure before (mmgH) | Systolic blood pressure after (mmgH) | Difference in blood pressure (mmgH) |
---|---|---|---|
1 | 156 | 151 | 5 |
2 | 162 | 158 | 4 |
3 | 149 | 147 | 2 |
4 | 168 | 165 | 3 |
5 | 153 | 152 | 1 |
6 | 165 | 161 | 4 |
7 | 159 | 156 | 3 |
8 | 151 | 151 | 0 |
9 | 164 | 158 | 6 |
Use your calculator, or work out by hand, the mean and the standard deviation of the differences
Calculate the standardized test statistic
Calculate the number of degrees of freedom
degrees of freedom = 9 - 1 = 8
Create a table in your calculator of the difference in blood pressure column
Perform a one-sample t-test using the data, remember that the population mean of the differences, , and the alternative hypothesis,
Calculate the -value
Compare the probability to the significance level and state the conclusion of the test
is rejected
Interpret the result in the context of the question
We have sufficient evidence to support the belief that the medication will reduce blood pressure in those that currently have high blood pressure
Last updated:
You've read 0 of your 5 free study guides this week
Sign up now. It’s free!
Did this page help you?