The problem of the comparability of change scores is investigated. Change quotients and residual change scores are evaluated as alternative approaches and methods for estimating the true change and true score residual, the reliability of change scores and residuals, and procedures for constructing confidence intervals for residuals are explored. It is concluded that: (1) residuals can be used to compare the performance of individuals or groups while holding the initial status variables mathematically constant if the data meets the assumptions of a multivariate normal distribution; (2) group means should be used to compute residuals for comparing groups with the same sample size; (3) parallel-forms reliability of raw-score residuals and other estimated true scores are not necessarily equal to the index reliability; (4) for statistical analyses of the determinants of change, partial correlations or multiple regression analysis should be used with final status as the criterion and initial status as one of the covariates; and (5) errors in the predictor or initial status variables can change even the sign of partial correlation or multiple regression coefficients. It is recommended that test-retest estimates of reliability be used to correct coefficients for attenuation. A bibliography and statistical data are included. (AE)