Examining Two Strategies to Link Mixed-Format Tests Using Multiple-Choice Anchors NEAT
- Author(s):
- Walker, Michael E.; Kim, Sooyeon
- Publication Year:
- 2010
- Report Number:
- RR-10-18
- Source:
- ETS Research Report
- Document Type:
- Report
- Page Count:
- 25
- Subject/Key Words:
- Mixed-Format Test, Constructed Response, Equating, Scaling, Non-Equivalent-Groups Anchor Test (NEAT) Design
Abstract
This study examined the use of an all multiple-choice (MC) anchor for linking mixed format tests containing both MC and constructed-response (CR) items, in a nonequivalent groups design. An MC-only anchor could effectively link two such test forms if either (a) the MC and CR portions of the test measured the same construct, so that the MC anchor adequately represented the entire test, or (b) the relationship between the MC portion and the total test remained constant across the new and reference linking groups. The study also evaluated whether linking mixed-format tests through MC-only anchors would be more effective than a two-stage strategy in which MC portions were equated through MC anchors and then composite scores were scaled to the MC scores. Anchor linking and two-stage linking yielded identical (or nearly so) results for both linear and nonlinear chained linking methods. With post-stratification linking methods the two-stage strategy resulted in smaller bias. The paper discusses some advantages of both approaches.
Read More
- Request Copy (specify title and report number, if any)
- https://doi.org/10.1002/j.2333-8504.2010.tb02225.x