Comparison of Three Matching Item Formats

Abstract

Three matching item test formats were compared empirically. The random order test format produced performance significantly superior to both the homogeneity of meaning test format and the one overall group test format. It was concluded: that the format of matching tests influences performance; that random ordering of matching items produces highest performance; that instructors should determine that both they and their students have denotative knowledge of the meaning of the so-called key words used in essay questions; and that matching assignments should not involve too large a number of stimuli and responses.