Improving the quality of computer scored tests in agronomic courses

Abstract

Student evaluation can be greatly facilitated by using computer scored and analyzed multiple choice tests. Test parameters can aid the instructor in the interpretation of student scores, and item analyses can provide him with valuable feedback for improving both teaching and writing. The caliber of testing influences the caliber of learning, and at college level it is more appropriate to evaluate complex learning outcomes than it is simply to test recall of information. Therefore we are justified in making an effort to evaluate attainment of higher levels of knowledge on multiple choice tests and to de-emphasize rote memorization. This paper describes ways of increasing the quality and usefulness of computer scored tests by examples from an introductory soil science course. The authors use test and item analysis; define the variable to be tested; and make a table of specifications to assure that the test reflects a balance, characteristic of the course, of both subject matter and level of knowledge. Most of their examples show how to write better test items, but they also give examples of ways to reduce the guessing factor and how to increase the value of the distractor. Some techniques for writing test items for high-level knowledge are: 1) using terms in problems rather than defining the terms, 2) presenting novel rather than familiar situations, 3) using problems that require applying knowledge, and 4) asking for interpretations of data requiring both knowledge and reasoning.