High-stakes tests are often used as agents for change, but attempts to introduce change in the classroom are not as effective as their designers wished they would be (Wall, 1997). The English for specific purpose (ESP) tests in Iran are such a case. The primary function of ESP tests in Iran is to select candidates for institutions of higher education (master and doctoral degrees). They are also designed specifically to promote changes in teaching ESP/ English for academic purposes at Iranian universities. This study was set out to examine washback effects of these tests on teaching ESP at Iranian universities. In doing so, data were collected through a questionnaire and classroom observations. The questionnaire was administered to 45 subject specialists teaching at Iranian universities. 10 other ESP teachers were also observed. The data were analyzed through descriptive and inferential statistics. Surprisingly enough, the results indicate that the ESP tests have fallen short of the goal. That is, these tests do not lead to innovation in teaching ESP, not do they influence teachers' teaching activities and ESP contents.