scale reliability

n. A measure of the consistency with which a device measures a particular variable. The most important measure of scale reliability is a test-retest correlation coefficient, which estimates the temporal stability of a test. Kuder-Richardson 20, coefficient alpha, and split-half reliability coefficients are all measures of the internal consistency of a scale, which is important when unidimensionality is an issue or when the dimension being measured is not expected to have temporal stability, as in a mood scale. The alternate forms reliability measure is rarely used as it is difficult to ensure that alternate forms are actually equivalent.