Effect of the closed-response format on modified rhyme test scores

An experiment was performed to determine whether high scores typically obtained with the Modified Rhyme Test (MRT) in the evaluation of speech-communication systems can be attributed to the closed-response format of the test. The 300-word MRT vocabulary was recorded by an adult male speaker and administered to two groups of listeners at two signal-to-noise ratios. The first group came for five test sessions during each of which they heard the six MRT lists at each test condition and utilized an open-response (write-down) format. The second group came for one test session and heard the six MRT lists at each test condition; they utilized the regular closed-response (multiple-choice) format. Test scores obtained from the listeners utilizing the open-response format and having considerable exposure to the test vocabulary were not appreciably different from test scores obtained from the listeners utilizing the regular closed-response format having no previous exposure to the test vocabulary. Additional tests administered to two other groups of listeners showed that MRT scores do not rise as listening experience is acquired.