Knowledge and professional skills’ tests are widely used in selection for government as well as private sector jobs in many countries. In order to assess the qualities of such tools, this research proposes to use an innovative combination of Classical Test Theory (CTT) and Item Response Theory (IRT) approaches. This study examines the structure, reliability, and validity estimates of the Accounting Knowledge and Skill Test (AKST) developed as part of a selection process for government employees in the Kingdom of Saudi Arabia. The analysis was performed on 1,915 applicants. By Combining the CTT and IRT approaches, a better identification of good and bad items was achieved as compared with following any single approach exclusively. The implications of using this combined method to assess the quality of such tests are discussed. Key issues related to psychometric evaluation of ability tests in selection context are also highlighted to encourage further research to collect more validity evidences.