Open AccessAN EXPLORATORY STUDY OF PROGRAMMED TESTS
Author(s) -
Cleary T. Anne,
Linn Robert L.,
Rock Donald A.
Publication year - 1966
Publication title -
ets research bulletin series
Language(s) - English
Resource type - Journals
eISSN - 2333-8504
pISSN - 0424-6144
DOI - 10.1002/j.2333-8504.1966.tb00541.x
Subject(s) - reliability (semiconductor) , test (biology) , statistics , branching (polymer chemistry) , section (typography) , subject (documents) , programmed instruction , computer science , mathematics , arithmetic , mathematics education , chemistry , paleontology , power (physics) , physics , quantum mechanics , library science , biology , operating system , organic chemistry
One approach to the reduction of testing time is through the use of “programmed tests,” which are tests based on a sequential system of branching in which a subject is directed to items appropriate to his ability level. The programmed tests considered in the present paper comprised two major sections: (1) A routing section which contains the branching necessary to direct the subject to the appropriate items, and (2) a measurement section which contains a short test with the item difficulties concentrated at the appropriate level for the subject. Four methods were investigated on an existing item data for — 885 students. By using approximately 40 items in the programmed tests it was possible to obtain correlations with the total test score on 190 items that were quite close to the reliability of the 190‐item test. However, essentially the same results were achieved for two shortened conventional tests consisting of the 40 or 42 items with the highest point‐biserial correlations with the total test score.