ABSTRACT
A test is said to be reliable if it yields consistent results. It is easy to see what this means by considering an extreme example. Suppose a professor writes a midterm exam on research methods that contains only four multiple-choice items. The items are on four different important concepts that were emphasized during instruction. Thus, the exam is valid in the sense that it covers appropriate content. Students who have thoroughly mastered the course content, however, should be concerned about taking such a test because it would be very easy to misinterpret a question or to miss a key term in it and get it wrong, yielding a score of 3 out of 4, or only 75% correct. On the other hand, students who have moved through the semester in a fog-not understanding even basic concepts-should be pleased at the prospect of taking this exam. With only four items, the odds of getting a few right by guessing, and thus passing the test, are reasonably high.
