A Study of the Extent to Which Positive or Negative Phrasing Affects Answers in a True-False Test

THIS PROBLEM involved the determination of whether or not the phrasing of the statement?either positively or negatively?has any effect upon the an swer given in a true-false test. In order to ascertain whether a difference in responses is typical, the au thor gave a number of true-false tests. There were four main groups which he wished to test: (a) those in which the test questions were all key ed true, (b) those in which the test questions were all keyed false, (c) those in w hic h all the statements were positive, and (d) those in which all the s ta te m ents were negative, in each of two general divisions of subject matter ? science and non-science. However, in view of the limited number of test?es, and the fact that one atypical testee's responses might make a statistical difference in the results, the author decided to combine all four groups?that is, each test?e would be in each of four groups. In order to accomplish this, each test was set up with a number of positive statements keyed true and an equal number of positive statements keyed false, an equal number of negative statements keyed true, and an equal number of negative statements keyed false. In order to attempt to cancel out the effects of specific determiners (except the positive or negative phrasing of the state ment), two sets of each test were given, each the opposite of the other. Where question number 1 was a positive statement keyed true in one test, question 1 was the same question, stated negatively and keyed false, in the othe r . The subject matter was based upon subject matter taught in class and would be used in regular tests. It was divided into two groups, that of non-science, and that of science material. Thus, for the non-science material we have as examples; Positive-true statement, Religious teachings were the most important