Choice between paired data or two independent samples

It is sometimes possible to answer questions about the difference between two means by collecting two alternative types of data.

Two independent samples
Measurements are made from two samples of individuals from the groups whose means are to be compared. A 2-sample t-test can be used to compare the means.
One paired sample
The 'individuals' can be re-defined as pairs of related values from the two groups and a single sample of these pairs can be collected. A paired t-test can be performed on the differences to compare the means.

If the individuals in the 2 groups can be paired so that the pairs are relatively similar, a paired design gives more accurate results.

Matched pairs in experiments

In experiments to compare two treatments, it may be possible to group the experimental units into pairs that are similar in some way. These are called matched pairs. If the two experimental units in each pair are randomly assigned to the two treatments, the data can be analysed as described in this section.

The difference between the treatments is estimated more accurately than in a completely randomised experiment.