Power calculation is a statistical method used to determine the sample size needed to detect an effect of a given size with a certain degree of confidence. It helps researchers ensure that their study is neither underpowered, which risks missing a true effect, nor overpowered, which can waste resources.