ERIC Number: EJ1193601
Record Type: Journal
Publication Date: 2018
Pages: 6
Abstractor: As Provided
ISBN: N/A
ISSN: EISSN-2375-5636
EISSN: N/A
Available Date: N/A
Item Response Theory-Based Methods for Estimating Classification Accuracy and Consistency
Diao, Hongyu; Sireci, Stephen G.
Journal of Applied Testing Technology, v19 n1 p20-25 2018
Whenever classification decisions are made on educational tests, such as pass/fail, or basic, proficient, or advanced, the consistency and accuracy of those decisions should be estimated and reported. Methods for estimating the reliability of classification decisions made on the basis of educational tests are well-established (e.g., Rudner, 2001; Rudner, 2005; Lee, 2010). However, they are not covered in most measurement textbooks and so they are not widely known. Moreover, few practitioners are aware of freely available software that can be used to implement current methods for evaluating decision consistency and decision accuracy that are appropriate for contemporary educational assessments. In this article, we describe current methods for estimating decision consistency and decision accuracy and provide descriptions of "freeware" software that can estimate these statistics. Similarities and differences across these software packages are discussed. We focus on methods based on item response theory, which are particularly well-suited to most 21st century assessments.
Descriptors: Classification, Item Response Theory, Accuracy, Reliability, Computer Software, Educational Assessment, Criterion Referenced Tests, Computation
Association of Test Publishers. 601 Pennsylvania Avenue NW, South Building Suite 900, Washington DC 20004. Tel: 866-240-7909; Fax: 717-755-8962; e-mail: wgh.atp@att.net; Web site: http://www.testpublishers.org
Publication Type: Journal Articles; Reports - Descriptive
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Grant or Contract Numbers: N/A
Author Affiliations: N/A