Knowledge about musical melodies develops gradually from birth to adulthood. By the age of five, most children can recognize out-of-key notes in a given melody, and eight-year-olds can distinguish harmonics in Western-type melodic sequences. Adults in the Western world typically have difficulty remember melodies that do not follow the rules of Western harmonic sequences and scales, when compared to more familiar tones. Weiss and colleagues (2015) examined the relationship among several factors in the development of melodic memory, including unconscious musical knowledge and relative preferences for vocal or instrumental frequencies and timbre. They conducted two studies, one with children aged 9 to 11 and another with children aged 5 to 8, to determine the effects of vocal versus instrumental timbre on their memory of melodies. The data were interpreted through the lens of dynamic systems theory. This theory states that development occurs in a complex nonlinear process that involves many interacting variables (Smith & Thelen, 2003). As a result, it is expected that the acquisition of musical knowledge and memory will involve several factors such as age, exposure to music, and timbre.
Study 1
Method
Children aged 9 to 11 years old (n=48) participated in a game in which they listened to 16 melodies in vocal, piano, banjo, and marimba timbres. They were then tested with the same 16 melodies as well as 16 new melodies to determine if they could identify whether they were familiar or unfamiliar. Melodies were presented randomly. All of the children were healthy with normal hearing. Children were recruited by age (16 in each age group) but not by gender or years of music lessons. Using computer and headphones, the tests were conducted in a sound booth. The vocal melodies used the “la” syllable rather than words. The children were not told specifically to try to remember the melodies nor was a distinction made between vocal and instrumental.

Statistical Analysis
The data were compared using a three-way mixed design analysis of variance (ANOVA) with timbre and exposure identified as the repeating measures and age as a between-subjects variable. Each child’s memory score was calculated by subtracting false alarms (incorrect identification of old versus new) from hits (correct identification of old vs. new).

Results
The ANOVA showed that children correctly identified the familiar melodies compared to the unfamiliar melodies (F=305.30, p<0.001). There was also an interaction of age with exposure which suggested that the older children were more successful than the 9 year olds (F=4.80, p<0.013). Timbre had an effect by itself (F=4.75, p=0.004) and interacted with exposure (F=5.83, p<0.001). Nonacademic Summary Children who are 9 to 11 years of age are able to identify melodies they have heard before, and 10 to 11 year olds perform better than 9 year olds. All children were more successful at identifying melodies that were presented to them by singing than they were at identifying melodies presented using instruments. Study 2 Method A total of 80 healthy children who had intact hearing (40 5-6 years; 40 7-8 years) were recruited for this study, again without concern for gender or years of music lessons. The study sessions were conducted in a sound booth with the same melodies used in Study 1. However, only the vocal and piano versions were used. The test was shortened to accommodate children’s shorter attention spans. Because each melody was heard only once, the study could not be directly compared to the first study. This was a significant limitation (Weiss et al., 2015). Statistical Analysis The same statistical analysis (three-way ANOVA) was used in the second study, with modifications for two timbres and two age groups. Results In Study 2, there was a three-way interaction between exposure, timbre, and age group (F=5.19, p=0.025). The researchers (Weiss et al., 2015) elected to perform a two-way ANOVA for each age group. For 5 to 6 year olds, there was an effect of exposure (F=18.73, p<0.001) and an effect of timbre (F=22.15, p<0.001) but there was no interaction. Younger children identified vocal melodies as old more often than piano melodies, indicating a perseverative preference for voices. For 7 to 8 year olds there was also a two-way interaction between exposure and timbre (F=7.93, p<0.001). Although distinction between old and new was made for both vocal and piano timbres, they showed better performance for vocal melodies. Nonacademic Summary Younger children, ages 5 to 6, tended to identify vocal melodies as familiar even when they were unfamiliar. This preference for the human voice reduced their accuracy. The 7 to 8 year old children were more successful at identifying familiar vocal melodies, but their responses were affected by both exposure and timbre. In most cases they were able to distinguish old from new using memory. Discussion These two studies provided important insight into the acquisition of musical knowledge and the development of memory. Dynamic systems theory predicts that there will be main effects of independent variables as well as interactions. Application of this theory to development suggests effects of age as well. The two younger age groups – 5-6 year olds and 7-8 year olds – showed a clear difference in memory for melody (effects of exposure) and developmental bias for the human voice. This is the time period when children have acquired the basics of language and their minds are evolving into more complex problem solving mechanisms. Piaget identified it as the separation point between pre-operational and concrete operational thinking (Lourenço, 2015). With regard to memory and executive functioning, around age 7 is when children begin to store by verbal description rather than phonologically (Gathercole, 2003). The change in performance between ages 9 and 11 corresponds to the move from Piaget’s concrete operational to formal operational stage (Lourenço, 2015). Therefore, as suggested by dynamic systems theory, a variety of developmental changes move children from the early preference for human voices to the ability to differentiate complex harmonic input. Critique of Methods The researchers chose to change the protocol in the second study in order to reduce fatigue and stress for the participants. They exposed children to melodies only once instead of twice. This is a significant change that, while well-intentioned, makes it more difficult to assess the results of the two studies. As the researchers stated, this change made it impossible to directly compare study 1 with study 2. Even though they stated that no difference was found with regard to gender, in study 1 the number of girls was too small to make a comparison. Therefore, the study could have been improved significantly by selecting equal numbers of males and females. Also, whether or not the child took formal music lessons was probably important and should have been considered, or at least reported in detail with information for each child. I consider it to be unlikely that music lessons would have no effect. Since even children as young as three years of age may take lessons, future research should consider this aspect. The ANOVA was probably the appropriate statistical test for this study. Technically, it is meant for interval and ratio data, but it could be argued that the data were interval even though they were integers and limited in range. Further Research As mentioned above, future research should be more rigorous by controlling gender and excluding children who have had formal music lessons. The 9 to 11 year old study should be repeated using exactly the same procedure as the study with younger children. Additional research might examine the results from children with varying levels of music training. This might produce interesting results, especially for children who begin learning the violin at age three. Conclusion Overall, this was a well-constructed study with excellent statistical analysis. In spite of the limitations indicated above, the study produced important information about children’s memory and knowledge of melodies.

References
• Gathercole, S. (2003) The development of memory. Journal of Child Psychology and Psychiatry, vol. 39, No.1, 3-27.
• Lourenço, O. L. O. (2015). Piaget’s Legacy to Human Development. World Journal of Behavioral Science, 1, 53-65.
• Smith, L. B., & Thelen, E. (2003). Development as a dynamic system.Trends in cognitive sciences, 7(8), 343-348.
• Weiss, M. W., Schellenberg, E. G., Trehub, S. E., & Dawber, E. J. (2015). Enhanced processing of vocal melodies in childhood. Developmental psychology, 51(3), 370.