Traditional item analysis and reliability estimation procedures cannot be used in evaluating the effectiveness of a test consisting in whole or in part of optional items. This paper considers alternative methods for determining the difficulty, discrimination, and reliability of optional items on essay and other types of tests. Although not so precise as traditional procedures, these approaches can yield useful information for assessing the functioning of optional items and item composites. A computer program, available in BASICA and FORTRAN, has been written to compute the measures described in this paper. In addition, illustrative analysis of two tests with optional items are provided.
Get full access to this article
View all access options for this article.
References
1.
Aiken, L. R. (1988a). A program for computing the reliability and maximum reliability of a weighted composite. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT, 48, 703-706.
2.
Aiken, L. R. (1988b). Small sample difference tests of goodness of fit and independence. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT, 48, 905-912.
