The differences between multiple-choice, simulated, and concurrent tests of software-skills proficiency are discussed. For three basic human-resource functions, the advantages of concurrent tests (i.e., those that use the actual application software) include true performance-based assessment, unconstrained response alternatives, and increased job relatedness. The Judd Tests, a recently developed line of software skills proficiency tests that offer assessment in five DOS-platform applications, are described to illustrate the advantages of concurrent testing. These advantages are: (1) concurrent tests are performance based, (2) concurrent tests allow examinees to use all functions normally available in the software application, (3) concurrent tests do not constrain examinees to certain paths of responding, and (4) concurrent tests are defensibly job-related for positions requiring use of computer software. (Author/SLD)