A Comparison of Four Methods of Selecting Items for Computer Assisted Testing
NAVAL PERSONNEL AND TRAINING RESEARCH LAB SAN DIEGO CA
Pagination or Media Count:
Four methods were used to select items for shortened versions of the Navy General Classification Test GCT and the Navy Mechanical Aptitude Test MECH. Two of these methods were used to produce short linear tests, and two to produce short branching tests. Item response data banks were used for simulated administration of all short tests. Obtained scores were then correlated with full length test scores. Navy recruits item responses were used throughout. Additional recruit samples were used for administering short tests in paper-and pencil form and short branching tests via computer terminal. Correlations with previously administered full length tests were obtained.