Principles and Applications of Automated Test Assembly
Presented by Wim J. van der Linden, University of Twente
The topic of optimal test assembly, introduced by Birnbaum in his contribution to Lord and Novick (1968), has been a prolific topic of research, particularly since the attempts to formulate optimal test assembly problems as problems of mixed integer programming (MIP) in the mid 1980s became successful.
Problems that can be solved routinely using MIP are
- IRT-based test assembly with targets for the test information function
- classical test assembly with optimization of test reliability or validity
- assembly of tests to have the same observed-score distributiona as a previous test (=observed-score pre-equating)
- item matching
- simultaneous assembly of multiple tests (such as sets of parallel tests, pretest-posttests, and multistage testing systems), as well as multidimensional tests
The same methodology can be used to assemble tests from pools of items organized as sets around common stimuli, select adaptive tests to elaborate systems of content specifications, and design and manage item pools. In his presentation, Mr. van der Linden introduced the basic principles of automated test assembly and showed a selection of its current applications to everyday testing problems.
Literature
Birnbaum, A. (1968). Some latent trait models and their use in inferring an examinee's ability. In F. M. Lord & M. R. Novick, Statistical theories of mental test scores (pp. 397-479). Reading, MA: Addison-Wesley.
van der Linden, W.J. (2005). Linear models for optimal test design. New York: Springer.
Want to learn more? Click here to learn more about this topic through our framework.
