Using longitudinal data to estimate reliability in the presence of correlated errors of measurement

Test-retest correlations can lead to biased reliability estimates when there is instability of the true scores in the interval between tests and/or when the measurement errors are correlated. Using three occasion data on the Test of Standard Written English and es-say ratings, an analysis is demonstrated which separates true score instability and correlated errors. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT 1980, 40 HEISE (1969) has shown how to estimate reliability from test-retest correlations taking into account the instability (i.e., change) in true scores over time, based on the assumption that reliability is constant over time. Wiley and Wiley (1970) used a similar approach to reliabil-ity estimation based on the assumption that error variances are con-stant over time. Both procedures are based on the hypothesis that the true scores over time form a &dquo;simplex pattern&dquo; which means that when a true score obtained at one time is predicted from an individ-ual’s true score on the just prior occasion, other earlier true scores will not improve that prediction. There is now a considerable body of liter