testretestreliability

Test-retest reliability, or the stability of a measurement over time, assesses whether an instrument yields consistent scores when administered to the same respondents under similar conditions on two or more occasions. The underlying assumption is that the trait or ability being measured remains relatively unchanged between administrations.

To evaluate it, researchers administer the same test to the same participants on two occasions separated by an appropriate interval. They then examine the relationship between the two sets of scores using statistical indices that reflect agreement or consistency. Pearson’s r is commonly used for interval data, while Spearman’s rho is used for ordinal data. Intraclass correlation coefficients (ICCs) are also widely used, particularly when multiple measurements or raters are involved, as they can account for both agreement and consistency and for different sources of variation.

Interpretation of results depends on the context; higher reliability indicates less random measurement error and more stable scores across time. However, high test-retest reliability does not by itself establish validity—the extent to which the instrument measures the intended construct. Reliability is a prerequisite for validity but not a guarantee. The choice of time interval, administration conditions, and participant characteristics can influence estimates of reliability. Short intervals may induce recall or practice effects, while long intervals may allow genuine change in the trait being measured.

Applications include scale development, clinical assessments, and longitudinal research. Researchers should report the chosen statistics, interval length, sample characteristics, and any potential sources of change or bias that may affect stability.