Abstract
Cognitive ability tests are widely used in organizations around the world because they have high predictive validity in selection contexts. Although these tests typically measure several subdomains, testing is usually carried out for a single subdomain at a time. This can be ineffective when the subdomains assessed are highly correlated. This article illustrates the applicability of using multidimensional computerized adaptive testing (MCAT) for a cognitive ability test used in organizational contexts. MCAT can increase the validity of subdomain ability scores by effectively administering and scoring items based on the relations between the subdomains. This can be particularly useful in settings where there is a need to provide detailed feedback regarding subdomain scores, but there is limited time for assessment. Results indicate that MCAT leads to improved test precision, shorter tests, and increased selection utility compared to more traditional testing methods.
Original language | English |
---|---|
Pages (from-to) | 123-139 |
Journal | International journal of testing |
Volume | 13 |
Issue number | 2 |
DOIs | |
Publication status | Published - 2013 |
Keywords
- METIS-300479
- IR-88532