Abstract
Some of the problems and hazards associated with the use of standardized reading test scores in the formation and re-formation of ability groups for reading instruction are discussed in this paper. The notion of “homogeneous grouping” is briefly examined, and some procedures that are commonly employed for the grouping of pupils on the basis of test scores are looked at together with a consideration of test error. Problems associated with the determination of pupil “growth”, and the re-location of pupils in reading groups on the basis of test score change, are examined in some detail. Particular attention is given to two of the major defects of “gain scores” (correlation between pretest and gain scores, and gain score unreliability), the influence of “time of testing” on test scores, and the conclusions that might be drawn about the effectiveness of grouping and instructional procedures.