A Technique for Developing Suppression Tests

EVEN though the theory of suppression tests has received some discussion (i, 2, 3, 4, 5) in the literature, such tests are almost nonexistent. This brief article proposes to point out a technique for economically developing serviceable suppression tests by changing the theoretical frame of reference from sup-pression tests to suppression items. An effective suppression test is one which measures only the non-valid variance which is present in an otherwise valid test (predictor). It will show a high correlation with the test for which it is a suppressor, but a low or no correlation with the criterion. When a battery of tests is used to predict a criterion, this type of test will enter into a composite prediction with a negative beta weight. In other words, higher scores on a sup-pression test lower the composite scores of those individuals. This serves to partial out the non-valid variance from an