Definition
A design in which repeated measurements are taken on the same subject/experimental unit at different points in time or under different Experimental Conditions. Since each subject is tested under each condition, such a design is very efficient because treatment comparisons take place within subjects, essentially eliminating subject-to-subject variation. This results in increased power, i.e., a greater chance of rejecting the null hypothesis if it is actually false. Also called cross-over design.
Examples
The paired t-test is the simplest case of a repeated measures design.
A single factor experiment with repeated measures is basically a randomized block design where the subjects are considered to be the blocks.