Reasons for administering tests fall into two categories--decision-making and promoting learning. The two bases of tests are learning objectives and the level of learning at which training is developed. Test development involves a number of steps. The best way to tie objectives to test items is through the use of a table of specifications, which organizes test development, specifies level of learning, determines relative importance of subject matter, and promotes uniformity across tests. Another part of test planning is the determination of test item formats to be used. Test item development follows these steps: write, review, rewrite, pilot. After test items are developed, they should be put into a test item bank. The next step is preparation for the organization and administration of the examination. Written tests can be improved only one way--by increasing the reliability and validity of the test. One activity that is very beneficial is the process of test item analysis. It includes three major activities--item difficulty, item discrimination, and alternative analysis. A test is a measuring device, and the numbers must be interpreted to have meaning as an evaluation. Part of the interpretation of test data is the application of an appropriate reference. Three major reference group classifications are content-referenced, group-referenced, and criterion-referenced. (YLB)