# Matched pairs

Description of activity and Excel instructions: Matched pairs

Dataset: seed.xls

In matched pairs designs we work with the differences between the observed values for each pair. Calculate the difference for each pair in the dataset:

- In cell A3, type "=A2-B2"
- Highlight cell A3 and drag the handle downward to copy the formula for all of the data pairs. (The cell reference will automatically adjust.)

## Always look at the data

Create a histogram as a summary of the data.

- In an empty column to the right of the data, set up the upper bound for each class (Note that the data include negative and positive values.
- Create the frequency distribution next to the classes.
- Create a histogram based on the frequencies.

Calculate the mean and standard deviation of the difference.

- Type "=AVERAGE(C2:C12)".
- Type "=STDEV(C2:C12)".

Return to the OLI Matched pairs page to complete the interpretive exercises on this portion of the analyses.

## Calculate the t-test and p-value

Calc does not report the t-test value. The formula for this test is:

- [math]t = \frac {\overline{y_d} - 0}{\frac{s_d}{\sqrt{n}}}[/math]

To calculate the t-test in Calc:

- Enter the formula "=(AVERAGE(C2:C12))/(STDEV(C2:C12)/SQRT(11))", where C2:C12 is the range of differences

Calc provides a formula for calculating the p-value which can be used with a few types of t-tests. Type=1 is for use in this situation: matched pairs. Note that Mode=1 indicates a one-sided test:

- To calculate the p-value, enter the formula "=TTEST([range y1],[range y2],1,1)".

Return to the OLI Matched pairs page to complete the interpretive exercises.