In countries with centralized educational systems national tests are used as primary devices through which changes in the educational system are introduced. This article examines the impact over time of two national tests which have been in operation for a number of years: one in Arabic as a second language (ASL) and one in English as a foreign language (EFL) through questionnaires, interviews and document analysis from a sample of teachers, students and language inspectors. Results showed different washback patterns for the two tests. Slight modifications in the ASL test created no effect in classroom activities, test preparation, or the status and prestige of the subject tested, yet, the inspectors expressed satisfaction and wished to continue the administration of the test as they feared that without the tests proficiency levels would drop. Slight modifications in the EFL test, on the other hand, created major impact in terms of teaching activities, time devoted for test preparation, production of new teaching material, etc. Although negative attitudes were expressed by teachers regarding the quality of the test, they would like it to continue. Inspectors believe that the test creates a meaningful change and is powerful enough to trigger changes without a need to provide training and a new curriculum. The study shows that washback varies over time, owing to many factors such as the status of the language and the uses of the test.