Notwithstanding the pioneer work of Spearman (1927 (and numerous studies by Thurstone on problems of intelligence, most testing programs seem to postulate only one type of intelligence. Division of the total score may be made into arbitrary categories for use in advising students but the program itself does not seem to conform to accepted theory on the subject. There has been little change in the administration of the program. This is partly due to the elaborate nature of test materials used. To save labor in developing a test it has been customary to build a long one to cover a given area most comprehensively. In many fields it has been found that less material carefully selected will do the job as well or better, thus affording notable economy in time and effort. Further, there is a limit to the measurement of human capacities by the longer methods used in the past. In order to sample human traits adequately it is necessary to reduce the length of conventional tests without doing violence to the reliability of scores. This problem was studied by Rostron and Lauer (1939) and the ground work laid for a testing program of diversified nature possible of administration within limits of time normally available in crowded orientation programs. The present study is an extension of the above paper designed to facilitate the development of parallel forms of tests. It is presented as a method of approach when limited time and energy are available to spend on the construction and application of measuring instruments.
Proceedings of the Iowa Academy of Science
©1945 Iowa Academy of Science, Inc.
Dodds, Mary; Day, Barbara; and Lauer, A. R.
"A Method for Evaluation and Equation of Test Forms,"
Proceedings of the Iowa Academy of Science, 52(1), 245-247.
Available at: https://scholarworks.uni.edu/pias/vol52/iss1/30