In this article, after describing one approach for teaching critical thinking (CT) that was in place at Baker University from 1990 to 2008, the author describes their experience assessing CT using three standardized exams and shows why the choice of a standardized CT test can be problematic and the results misleading. These results can be misleading because the effect size gains (defined as the mean gain expressed as a percentage of a standard deviation) on CT tests can vary significantly, even under ideal conditions where the course, text, and teachers remain relatively the same throughout a lengthy assessment period. In addition, the data indicate a wide disparity of student gains relative to their instructors. As a consequence, the assessment results may vary greatly depending on which test is chosen or whose scores are reported. At a time when some suggest that an institution's test results be made public, this comparison of three common tests of CT should be helpful. Although there has been research employing each of the three standardized CT tests used at Baker, no study has focused on the question of whether these three widely used tests (the Ennis-Weir Critical Thinking Essay Test, the California Critical Thinking Skills Test, and the Cornell Critical Thinking Test, Level Z) are similar in reporting student performance. Without such knowledge, one might think that one's favored approach to teaching CT was yielding terrific results, when in fact if one used another assessment instrument the results would be quite different. Therefore, clear guidance on choosing a test that best fits how one understands CT and which skills are to be tested is essential. (Contains 1 table.)