Multitrait Situation Reaction Tests

• In Event: Process Data Modeling and Application in Educational Assessments

Abstract

Situation Reaction Tests contain items presenting alternative actions related to some context, to be ranked preferentially. Such ranks can be compared by unfolding the number of similar decisions made. The number of unique scoring patterns is the permutation space. Each respondent’s ranking, compared to each scoring pattern, yields score sets for each item. Associating these with related criterion trait measures yields item keys for estimating those traits. Averaging over items, scored using such keys, is very efficient; one test with multiple criterion keys, creating a Multi-Trait Situation Reaction Test. Using Monte Carlo methods, we demonstrate a simulated test scoring method.

Authors

• Warren E. Lacefield, Academic Software, Inc.

• Brooks Applegate, Western Michigan University