This study has attempted to investigate the administered written tests in the language department of Islamic Azad University of Hamedan, Iran from validity, practicality and reliability points of view. To this end two steps were taken. First, examining 112 tests, we knew that the face validity of 50 tests had been threatened, 9 tests lacked content validity, and 2 tests were impractical. Second, randomly the sores of 103 tests (8 multiple-choice and 95 essay-type), related to different academic courses, were fed into SPSS software to estimate their reliability statistically using Cronbach's alpha. The results showed that 7 multiple-choice tests had adequate reliability, 39 out of 95 essay-type tests had acceptable reliability, and totally there was a rather acceptable correlation coefficient among the scores of 95 essay-type tests. Finally, some suggestions were offered to improve the future tests.