Come see me at the annual convention of the Association for Educational Communications and Technology in Jacksonville, Florida in November where I’ll be presenting my dissertation results in a featured research session titled Facilitating Variable-Length Computerized Classification Testing In Massively Open Online Contexts Via Automatic Racing Calibration Heuristics on Friday, Nov 7 from 9:15 to 10:15am (2nd Level, Grand 7).
Here’s the abstract for the session…
Computer Adaptive Testing (CAT) provides an accurate and efficient method of assessment that is particularly useful in massively open online contexts. However, use of CAT in such contexts is hindered by arduous item calibration requirements. This study uses data from thousands of examinees to evaluate an innovative approach to item calibration — Automatic Racing Calibration Heuristics (ARCH). Results demonstrate that ARCH matched established methods in test accuracy while, in some cases, significantly shortening test lengths.
Hope to see you there!