A new Method of Weighting Scores in a True-False Test

Abstract

The experiment herein reported involved the giving of a mid-quarter examination in a course in Extra-Curricular Activities to a class of 62 graduate and undergraduate students at the University of Minnesota during the summer session of 1935. The test was composed of 75 true false items and was so constructed as to call for the weighting of his responses by the examinee himself according to the degree of assurance with which he made those responses. Thus the examination was calcu lated to test not only knowledge as such but also the certainty with which that knowledge was held. The instructions to the student were as follows : Mark a statement + if you judge it to be true and O if false. You may claim credit up to 4 points for each of your responses if you wish. Before your responses encircle 4, 3, or 2, depending on the credit you want. If your answer is wrong, the penalty will be double the amount of credit you claim. (It is advisable to claim 4 credits if you are sure your answer is correct.) ' ' If you claim no special credit as described above, you should, never theless, answer all questions, even if you have to guess. All such answers will be scored in the ordinary way, right-minus-wrong.? ' Item No. 52 of the test is quoted below to illustrate how each of the 75 questions appeared in the test : 52. 4-3-2Investigations of actual practice indicate that school administrators and teachers have had little to say relative to the content of assembly programs. ' ' Because the penalty for a wrong answer was double the credit claimed, the examinee was actually in a position of having to offer odds of 2 to 1 that his answers were correct. It was thought that such large odds would be a better test of the degree of conviction with which the examinee made his responses than the ordinary true-false test situation, and would dis courage more guessing than is done by the latter. As the instructions further indicate every question was to be answered, even if no special credit was claimed. This was required in order to per mit a comparison of scores obtained in the ordinary way (right-minus wrong) with the weighting method. Thus the test was first scored by 290