Computerized Adaptive Testing A Primer

mind-and-cognition

• high scorers usually miss only hard items and low scorers usually correctly answer only easy ones — IRT proficiency estimates capture most of what the data say about comparing examinees

  VLA-3D83E897-RC-240

• CAT is based on the assumption that items within tests measure the same dimension of individual differences; it is then a matter of indifference which items are presented

  VLA-3D83E897-RC-216

• errors can be minimized by investigating the dimensionality of item pools before finalizing the scales to be approximated as unidimensional

  VLA-3D83E897-RC-095

education-and-formation

• development of admissions testing at American universities parallels military testing, begun in earnest at the start of the twentieth century with the founding of the College Board

  VLA-3D83E897-RC-021

• our previous lack of attention to operational issues was a grievous oversight — we focus on examinee access, item pool usage and security, and the economic realities of CAT

  VLA-3D83E897-RC-277

work-as-meaning

• a job analysis coupled with experimental administration of tests requiring the same abilities as the job, and a validity study correlating test performance with later job success, yielded instruments

  VLA-3D83E897-RC-020