An Empirical Investigation to Assign Weights to the Possible Responses in a True-False Examination When Guess or Do Not Guess Directions Are Given

Of the many forms of objective type tests, the true-false has been, perhaps more than any other, the subject of much discussion. In the true-false, yes-no or other forms of two-choice question, on the average a student with no knowledge of the subject will mark correctly, by chance, approximately half of the answers. This defect has long been recognized, and many writers have offered methods for eliminating this erroneous factor. Some investigators have also endeavored to learn whether it is better to direct pupils to guess when they are not sure, or are ignorant of the answer; or to warn them against such a procedure. To date there has not been complete agreement on the extent to which guessing should be penalized, or the directions that should be given. It is the purpose of this study to investigate these two problems, i.e. I. What weights to assign to the various responses in a true-false examination when directions are (a) Do not guess. (b) Guess. II. To determine the effects of Guess and Do Not Guess directions on (a) Reliability (b) Validity We shall investigate as a minor problem the relative difficulty of true and false statements by noting it pupils, tend to answer incorrectly more true statements or more false statements. The results obtained in this study will be compared with the findings of previous investigations