| Abstract Research in cognitive science has shown that vocalization (speaking) can be a useful rehearsal strategy for enhancing memory. But the relationship between vocalization (singing) and memory for music is still unconfirmed. This study will therefore examine the degree to which singing can improve memory for melodies, thereby also potentially improving error detection accuracy. In choosing melodies for this study, it was important to first consider their level of difficulty. Based on existing perceptual research, a model of melodic complexity was constructed that accounts for the effects of interval size, contour, interval direction, chromaticism, and meter. Subjects then participated in a recall task that included melodies of varying difficulty. After hearing each melody, subjects sang back as much as they could remember, repeating this process until two correct consecutive vocalizations were produced. After each melody was memorized, subjects examined a notated version and located one pitch error and one contour error. On average, participants needed eight repetitions to memorize each melody, with no significant differences being found between the various levels of difficulty. This contradicts the hypothesis that a positive correlation would exist between number of repetitions and degree of melodic complexity. For the error detection task, the prediction that accuracy would be higher for contour errors than for pitch errors was confirmed. The final portion of this study involves a non-vocalization version of the same task. The data reported on above is being used to determine the number of times participants will hear a melody before performing the same error detection task, this time without vocalizing between each repetition. Forthcoming data will provide additional information about whether or not singing is indeed a strategy that can improve both memory capacity and error detection accuracy. |
