Key points are not available for this paper at this time.
A regression formula is derived for estimating a student's true gain from his initial and final test scores. A formula is given for the reliability of estimates thus obtained. A numerical example is given, illustrating some marked inadequacies of the simple, "common-sense" procedure that estimates gain by subtracting initial from final score. A convenient graphic procedure is presented for grouping all examinees tested according to the size of their estimated true gains.
Frederic M. Lord (Sat,) studied this question.