This monograph reports research, related to the multiple-choice test item, which is conducted by psychometricians and educational technologists in Japan. Sato's number of hypothetical equivalent alternatives is introduced. The author proposes a new index, k*, which can be used, among other things, for invalidating three-parameter models for the multiple-choice item. Shiba's research on the measurement of vocabulary, which is based upon latent trait theory, includes an eventual tailored test on vocabulary, utilizing information obtained from distractors as well as correct answers. With this research in mind, the author has developed basic ideas about a new family of models for the multiple-choice item. These are based upon both the information given by distractors, and the correct answer and the noise resulting from random guessing. (Author)