|
Development of Parallelism Test on Linear Equations with One Variable |
|---|---|
| รหัสดีโอไอ | |
| Creator | Thanakorn Thepnarong |
| Title | Development of Parallelism Test on Linear Equations with One Variable |
| Contributor | Kittitas Wancham |
| Publisher | Centre for Education Innovation, Print and Online Media |
| Publication Year | 2569 |
| Journal Title | An Online Journal of Education (OJED) |
| Journal Vol. | 21 |
| Journal No. | 2 |
| Page no. | 1-20 |
| Keyword | item specifications, item parallelism, test parallelism, linear equations one variable |
| URL Website | https://so01.tci-thaijo.org/index.php/OJED |
| Website title | An Online Journal of Education (OJED) |
| ISSN | 1905-4491 (Online) |
| Abstract | Parallel tests are constructed to mitigate the issue of test item exposure. They can be utilized to assess learners' knowledge while ensuring standard equivalence across examinations. Furthermore, they prevent test leaks resulting from exposure. Given these benefits, parallel tests have been developed to measure mathematics proficiency, aiming to truly enhance students' knowledge and abilities. This research aimed to: 1) Develop and validate the quality of item specifications for multiple-choice items on linear equations with one variable. 2) Develop and validate the quality of parallel tests and parallel test items on linear equations with one variable. The informants consisted of 4 mathematics teaching experts and 1 educational measurement and evaluation expert. The research sample comprised 521 Grade 7 students obtained through multi-stage random sampling. The researcher validated the quality of the item specifications using the proportion of agreement. The quality of the tests and parallel test items was evaluated based on item parameters, internal consistency reliability, content validity, construct validity, and the RMSD index. The research findings revealed that: 1) The item specifications achieved a proportion of agreement of 1.00 across all components. The developed specifications comprised 5 components: (1) measured objectives, (2) stimulus attributes, (3) response attributes, (4) sample items, and (5) specification supplement. 2) Both versions of the tests on linear equations with one variable fit the 2-parameter logistic model (2PLM). They exhibited high internal consistency reliability via Cronbach’s alpha coefficient, as well as content validity and construct validity. In addition, parallelism was demonstrated across both the test characteristic curves (TCC) and test information functions (TIF). As for the test items, parallelism was also established across both the item characteristic curves (ICC) and item information functions (IIF) for every matched pair. |
