Algorithms for the Construction of Parallel Tests by Zero-One Programming. Project Psychometric Aspects of Item Banking No. 7. Research Report 86-7.

Autor/in	Boekkooi-Timminga, Ellen
Institution	Twente Univ., Enschede (Netherlands). Dept. of Education.
Titel	Algorithms for the Construction of Parallel Tests by Zero-One Programming. Project Psychometric Aspects of Item Banking No. 7. Research Report 86-7.
Quelle	(1986), (35 Seiten) PDF als Volltext kostenfreie Datei Verfügbarkeit
Sprache	englisch
Dokumenttyp	gedruckt; online; Monographie
Schlagwörter	Algorithms; Computer Assisted Testing; Foreign Countries; Item Banks; Latent Trait Theory; Test Construction; Test Items + Suchen Sie Ihr Suchwort? Algorithm; Algorithmus; Ausland; Latent-Trait-Modell; Testaufbau; Test content; Testaufgabe
Abstract	Nine methods for automated test construction are described. All are based on the concepts of information from item response theory. Two general kinds of methods for the construction of parallel tests are presented: (1) sequential test design; and (2) simultaneous test design. Sequential design implies that the tests are constructed one after the other; simultaneous test design considers the construction of all forms at the same time. The first two sequential methods (Methods 1 and 2) discussed do not assume any manipulation on the collection of items. One is based on constructing two tests following the same test item specifications, and the other assigns items in the second test to items in the first test. Four additional sequential methods (Methods 3 through 6) assume that the collection of items is divided into two comparable parts. These methods differ from the first two only in the specific groups of items from which the desired tests are determined. Three methods (Methods 7 through 9) are also presented from simultaneous test construction. Method 8 is based on the method of E. Boekkooi-Timminga (1986). The other two methods start from the principle of assignment. All of the methods are based on the use of the optimization technique zero-one programming. Illustrations of each of the nine methods indicate that the first two methods suggested may not be suitable for constructing parallel tests and that Method 8 most warrants further investigation. Three tables present test design methods and data from test selection. (SLD)
Anmerkungen	Mediatheek, Faculteit Toegepaste Onderwijskunde, Universiteit Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands.
Erfasst von	ERIC (Education Resources Information Center), Washington, DC