Sara L. Hagen,Valley City State University

The study of creative behavior has received increased interest during the last 45 years, following the keynote address given by J.P. Guilford in 1950 to the American Psychological Association. The purpose of recent research, including this project, has been to help teachers understand what they may expect from their students in the classroom and to increase their ability to assess musical growth (Reinhardt, 1995). Continued research may provide teachers with better strategies for providing environments which are the most effective in giving students the opportunities to create original music. Hickey (1995) suggested that by creating such an environment, teachers may be able to teach their students a "deeper musical understanding and appreciation" (p. 11). She also suggested that the computer might be the most practical means with which to provide an environment rich in varied tasks from specific to open-ended. The present study used computer technology to encourage fifth-graders to create musical compositions using synthesized percussion timbres. The role of audiation and its possible effects in the process of composing was also explored.

The purpose of this study was to examine the role of audiation on children’s compositions using synthesized percussion timbres to determine if there are any relationships between audiation and the ability to compose using only percussion sounds.

Five questions were addressed in the current study: (1) Is there a relationship between audiation as determined by the Gordon IMMA test and the process by which children compose music? (2) Is there a relationship between audiation and the musical characteristics of children’s compositions? (3) Is there a relationship between children’s compositional processes and the musical characteristics of the resulting compositions (products)? (4) Is there a relationship between certain process behaviors (e.g., exploration, silence, repetition, or development) or musical characteristics in the resulting products (e.g., tempo or metric cohesiveness, repeated timbral or rhythmic patterns, developed timbral or rhythmic patterns) and students’ ability to replicate their compositions? (5) Do children’s self-reports provide any additional insights into the process of composition?

The variables were scores from the tonal and rhythmic subtests of Gordon's Immediate Measures of Music Audiation (IMMA), scores from the process and product definitions as listed in Figure 1, and the ability to replicate. One process variable was created for the present study, the relationship between the process and the product. This variable was rated by the process judges to determine how much of the process was reflected in the final product, or the degree to which the ideas that subjects used and practiced during the composing process were incorporated into their final compositions.

Exploration: The music in a 5-second interval sounds unlike music played in earlier 5-second intervals. No specific references to music played earlier can be heard.

Development:The music in a 5-second interval sounds similar to, yet different from, music played in an earlier 5-second interval. Clear reference to music played earlier can be heard in the rhythm, timbre combination, or both.

Repetition:The music in a 5-second interval sounds the same as music played in an earlier 5-second interval.

Silence: No music is heard in a 5-second interval.

Timbral pattern: 2 to 7 timbres that form a distinct and recognizable pattern which is perceived as a unified whole.

Repeated timbral pattern :A timbral pattern that is identical to a previously occurring timbral pattern. (Y = pattern exists, N = no pattern exists)

Developed timbral pattern : A timbral pattern that is similar to, yet different from, a previously occurring pattern. (Y = pattern exists, N = no pattern exists)

Rhythmic pattern : 2 to 7 durations that form a distinct durational pattern that is perceived as a unified whole.

Repeated rhythmic pattern : A rhythmic pattern that is identical to a previously occurring rhythmic pattern.

(Y = pattern exists, N = no pattern exists)

Developed rhythmic pattern: A rhythmic pattern that is similar to, yet different from, a previously occurring pattern. (Y = pattern exists, N = no pattern exists)

From Journal of Research in Music Education, Summer. (c) Copyright 1994 by Music Educators National Conference. Reproduced with permission for modification.

The process judges for both the pilot study and the main study were the same, both music instructors at a local university. The interjudge reliabilities in the main study were somewhat lower than in the pilot study. The process judges evaluated the behavioral processes used in composing. Each 5-minute period was divided into sixty 5-second intervals and was scored as one of the process behaviors (exploration, silence, development, and repetition). The number of intervals of each process behavior was totaled and the means of the judges’ scores for the four processes were calculated. At the the end of the 5-minute preparation time, the judges rated the first recording of the composition for overall cohesiveness and the degree to which the process related to the product using Likert-type scales from 1-7 with 7 being the most cohesive and the rating for which the process was most related to the product.

The judges for the composed products were two local elementary music teachers who evaluated the composed products by rating cohesiveness and elements used within the first-recorded pieces. Using a 7-point Likert-type scale, they rated the tempo and metric stability of the products. They also identified the meter, if one existed. Repeated or developed patterns were identified with "yes" or "no" answers.

The pilot study interjudge reliabilities were all above r = .80, which was acceptable and comparable to other similar studies (Kratus, 1994, Reinhardt, 1995). In the main study, the interjudge reliability for the process judges fell to r = .59. Interjudge reliability correlation coefficients were especially low between exploration (r = .34) and development (r = .36). These two variables proved to be the most difficult to judge, perhaps due to the wide variety of sounds. Judges’ fatigue may also have been a contributing factor in the lower reliability score. The judges’ reliability for rating the relationship between the process and the product was only moderate as well (r = .57).

The interjudge reliability for the product judges for the subjects’ ability to replicate was (r = .85). The interjudge reliability scores for tempo stability and metric stability were also relatively high in the present study (r = .79 and r = .77, respectively). In addition, reliabilities were moderate to high for the use of repeated and developed rhythmic patterns (68% and 87%, respectively).

The subjects were 38 fifth-grade students with a mean age of 11.3 years in two self-contained classrooms of the same elementary school. These subjects normally received formal music instruction which included a varied curriculum from a music specialist for 100 minutes per week. This age group was chosen for a number of reasons: (a) Most fifth-graders are capable of making meaningful music; (b) Fifth-grade students exhibit varying stages in their ability to compose with an orientation to process, product , or intentionality (Reinhardt , 1995); and (c) Fifth-graders are capable of replicating their compositional products to a greater extent than are younger children (Hedden, 1992; Hedden, Hoffman, & Mims, 1990; Kratus, 1986, 1989; Reinhardt, 1995).

A pilot study with one group of fourth-graders was conducted. After consideration of the results, time allowed for composing and the number of timbres made available were increased. The judges were also given the opportunity to practice. Two sets of two judges were used, one set for process behaviors and the other for the composed products. After the pilot study, minor changes were made in the evaluation instruments for ease of use and efficiency.

The sound source for the task was a Korg X3 Workstation synthesizer. The ten percussion timbres located on ten concurrent white keys were marked with stickers using the General MIDI program Dance Kit 135. The individual timbres were open conga (E4), hi timbal (F4), agogo (G4), cabasa (A4), whistle S (B4), whistle L (C5), guiro (D5), woodblock2 (E5), woodblock3 (F5), and open cuica (G5). Finally, the synthesizer was connected with an Opcode MIDI Translator II to a Macintosh Quadra 650 computer running a researcher-designed program using HyperCard 2.2 software.

The Gordon IMMA test with both subsets was given to all of the students in one session one week before the individual sessions. At the start of each individual session, the subjects were given time to explore the available sounds. Subjects were also given the opportunity to try out the HyperCard stack by moving from card to card and manipulating the record/play function on the "tape recorder". After this practice period, the subjects were told that they would have 5 minutes to make up and practice a piece that "no one had ever heard before" using the same ten notes used during the exploration period. They would be told when their time was half-way through. Students then would have 2 minutes to complete their composition and record it on the third card. They were also told that if they did not like their first attempt, they could record their composition as many times as needed within the two minutes. Finally, students were directed to record their piece again. (Examples of collected data were shared.)

During the process of composing, subjects spent about a third of their time with exploration and silence, and the remaining two-thirds of their time with repetition and development. Moreover, even though the synthesized percussion timbres were novel to the students, they still spent less time with exploration than they did with repetition and development. These results are consistent with the results of other studies (Hedden, Hoffman, & Mims, 1990; Kratus, 1989, 1991, 1994).

Two correlation coefficients were significant between audiation (IMMA scores) and the process by which children compose music–inverse relationships between rhythmic audiation and exploration (r = -.35) and the composite audiation scores and exploration (r = -.33). These results suggest that the greater the ability to audiate, the less time is needed for exploration. Perhaps because students with better audiation skills can recall the musical materials more readily, they spend more time with repetition and development behaviors. These results corroborate Kratus’ (1994) results.

The present study found no significant correlations among audiation and the processes of development and silence. These findings were in contrast to Kratus’ (1994) in which he found significant correlation coefficients between audiation rhythm and composite scores and silence. Students may have enjoyed the synthesized timbres so much that there was very little silence during the timed interval. Students’ ability to develop their motivic ideas may have been affected by the use of synthesized percussion timbres as well. Low interjudge reliabilities in the present study (r = .36 for development) may also have been a factor in these results.

No significant correlation coefficients were found between audiation and the musical characteristics of children’s compositions. These results may be related to the fact that the compositions in the present study were largely exploratory in nature, as determined by Reinhardt’s (1995) criteria for categorizing replicability scores. Reinhardt's (1995) rating tool (with slight modifications)was used to categorize the creative products from the present study using the ability to replicate scores. Results were as follows: (a) 63% were considered to be exploratory; (b) 16% were considered to be process-oriented; (x) 8% were considered to be product-oriented; and (d) 13% were considered to be composed (replicable). Finding no significant correlations between audiation and the products composed would coincide with the inverse relationship found between exploration and audiation. Exploratory behaviors do not necessarily result in a product with an established tempo or metric stability. Additionally, although repetition and development may be occasional aspects of an exploratory piece, they probably occur more by accident than by design. Conversely, if repetition and development behaviors seem to dominate the process, perhaps the children’s lack of ability to recall could account for their lack of ability to replicate their pieces. Sixty-six percent of children’s time in the present study was spent in repetition and development to create compositions of which 63% were rated as exploratory. These results are again in contrast to Kratus (1994), who found significant correlation coefficients among tonal, rhythm, and composite audiation scores and tonal cohesiveness and among rhythm and composite audiation and metric cohesiveness. In addition, he found significant correlation coefficients among tonal and composite audiation scores and the use of a developed rhythmic pattern.

Question three pertained to the examination of relationships among the processes of composition and of the resulting products. First, subjects’ use of compositional processes was significantly correlated with the cohesiveness of their final products in all but repetition. The use of development was positively correlated to the creation of a cohesive piece. The relationship between the process and the product was related to the resulting composition in the use of interesting rhythm and timbre combinations and structure as well. These results suggest that the way in which students approach composition does affect their final products, in that the development of ideas strengthens the composition. Perhaps those students who manipulated the sounds for the purpose of developing their ideas were more thoughtful about musical decision-making. Next, exploration was inversely related to tempo and metric stability. Silence was also found to have an inverse relationship with the final cohesiveness of the product. Silence may not have been used as a time to reflect upon choices, but may have indicated that the subject was not on task.

Examining the products further, nearly half used a repeated rhythmic motive and more than one-fourth used a repeated timbral pattern. However, the percentages for the incorporation of both developmental patterns of rhythm and timbre were low. These results suggest that fifth graders are capable of creating a motive and using it repeatedly. Nevertheless, they may lack sufficient maturation, performance abilities, or training needed to manipulate and develop their motives.

The ability to replicate (question four) does have a relationship with the cohesiveness of the final product. This makes sense for two reasons. The ability to replicate means that the students can provide unity. A certain amount of unity is necessary in creating a cohesive composition. In addition, high correlation coefficients were found for the ability to replicate and tempo stability as well as for the ability to replicate and metric stability. This suggests that the presence of steady pulse and a definite meter is related to the ability to replicate a piece. (Only 7 of the 38 compositions were rated a 5 or higher on a 7-point Likert-type scale for replication.)

Question five examined students' responses to several questions following the task. Many students (58%) reported that they were able to make their compositions sound the way they were intended. Some of the reasons given for "no" and "kind of" responses were that: (a) They couldn’t remember what they wanted to do, (b) They hit "wrong" notes, (c) They liked the first piece better and couldn’t change the second recording, (d) They needed time to discover where the sounds were located, and (e) They lacked performance capability. Although almost all of the students responded that they enjoyed the experience, many seemed to have difficulty verbalizing about the composition experience. Perhaps they did not have the musical terminology with which to discuss their musical intentions.

How a child approached the process of composition did affect what the child composed. Audiation appears to have an effect on the ability to compose. Because the ability to audiate stabilizes around the age of 9, the more experiences children have to imagine sound inwardly before that time would appear to be important. Development and repetition seem to have the greatest effect on a cohesive and musical final composition. As these skills seem to improve with practice (Reinhardt,1995), teachers should offer repeated opportunities for students to practice improvising and composing with various melodic and nonmelodic materials. Teachers may improve their students’ composition skills most by showing them techniques for integrating repetition and development. Finally, three characteristics were found to be related to the ability to replicate. They were: (a) a steady tempo, (b) a definite meter, and (c) the use of repeated rhythmic patterns. Creative activities which teach students how to incorporate these characteristics into their compositions may provide students with a vocabulary of skills to use for the creation of cohesive compositions.