The measurement of change in quasi-experimental educational research was discussed. Problems related to measuring change exist to varying degrees in all research designs; these issues are less troublesome in experimental studies because the investigator can manipulate the interest variables and observe their effects on other variables. Measuring change is more difficult in quasi-experimental studies because the investigator lacks the freedom to manipulate the variables. This study focused on issues of change associated with the latter design. Specifically, this study utilized the non-equivalent control group design, in which the results of one or two pretests are available prior to the investigation. Four models, or solutions to the fan spread hypothesis, in which variables can be used to predict change, were compared: (1) gains in standard scores; (2) single covariable analysis of covariance with estimated true scores; (3) gain scores adjusted for differential growth rates; and (4) multiple fallible covariable analysis of covariance. Methods (1), (2), and (3) were found to be suitable when the pretest-post test correlation of individual growth was unity. When this correlation was imperfect, methods (1) and (3) estimated the desired effect. (GDC)