ERIC Number: EJ1161905
Record Type: Journal
Publication Date: 2017
Abstractor: As Provided
Comparison of IRT and CTT Using Secondary School Reading Comprehension Assessments
Coggins, Joanne V.; Kim, Jwa K.; Briggs, Laura C.
Research in the Schools, v24 n1 p80-93 Spr 2017
The Gates-MacGinitie Reading Comprehension Test, fourth edition (GMRT-4) and the ACT Reading Tests (ACT-R) were administered to 423 high school students in order to explore the similarities and dissimilarities of data produced through classical test theory (CTT) and item response theory (IRT) analysis. Despite the many advantages of IRT demonstrated in simulation studies comparing CTT and IRT, it is still necessary to study the practical application of IRT to the measurement of reading comprehension assessment using actual test data in order to determine whether IRT's significant advantages persist. Accordingly, the following research questions were addressed: (a) To what degree does IRT offer a significant advantage over CTT when evaluating student reading comprehension ability? and (b) Which IRT model provides the best fit for the data? Results revealed that IRT's ability to provide more accurate information about item-level properties combined with independent trait calibration minimizing measurement error afforded significant advantages over CTT, thereby strengthening test design. It was determined that the 3-parameter logistic (3-PL) model provided better fit than did the 1-parameter (1-PL) and 2-paramenter (2-PL) models.
Descriptors: Item Response Theory, Test Theory, Reading Comprehension, Reading Tests, Comparative Analysis, College Entrance Examinations, High School Students, Goodness of Fit, Accuracy, Error of Measurement
Mid-South Educational Research Association (MSERA). Web site: http://www.msera.org/publications-rits.html
Publication Type: Journal Articles; Reports - Research
Education Level: Higher Education
Authoring Institution: N/A
Identifiers - Assessments and Surveys: Gates MacGinitie Reading Tests; ACT Assessment