Some considerations in the measurement of change

When conducting a study to assess the "gains in ability" or the "shifts in attitudes" of a group of subjects, the researcher must be aware of the presence of measurement error and regression toward the mean and must attempt to minimize and allow for these phenomena. Also of concern are whether the study can be generalized to a larger population and what statistical techniques are appropriate for the conditions of the study and the type of data collected. Educational and psychologic measurement is subject to particular types of error and must be analyzed and interpreted accordingly. The authors discuss and make recommendations regarding these most important considerations.