8010 wk5 discussion

International Journal of Music Education

2016, Vol. 34(1) 32 –48 © The Author(s) 2015

Reprints and permissions: sagepub.co.uk/journalsPermissions.nav

DOI: 10.1177/0255761415584296 ijm.sagepub.com

The effects of music composition as a classroom activity on engagement in music education and academic and music achievement: A quasi- experimental study

Michel Hogenes The Hague University, The Netherlands

Bert van Oers VU University, The Netherlands

René F.W. Diekstra The Hague University, The Netherlands

Marcin Sklad University College Roosevelt, The Netherlands

Abstract The present study aims to contribute to the understanding of the effects of music education, in particular music composition as a classroom activity for fifth- and sixth-graders. The intervention (experimental condition) focused on a three-step-model for music composition, based on the Cultural Historical Activity Theory of education, and has been compared with a teacher-centered approach mainly based on students’ reproduction of music (control condition). Results indicated that after the six-month intervention period, students in the experimental group were more engaged in music education compared to students in the control group. The research did not show a statistical difference in learning outcomes with regard to intelligence, academic achievement and music achievement, although the students of the experimental group performed better with regard to reading comprehension than their counterparts in the control group. The authors conclude that music composition as a classroom activity is feasible and useful in elementary schools.

Corresponding author: Michel Hogenes, The Hague University of Applied Sciences, Johanna Westerdijkplein 75, 2125 EN The Hague, The Netherlands. Email: [email protected]

584296 IJM0010.1177/0255761415584296International Journal of Music EducationHogenes et al. research-article2015

Research Article

Hogenes et al. 33

Keywords academic achievement, cognitive functioning, cultural historical activity theory, engagement, music achievement, music composition, music education

Introduction

For a long time, music educators have suggested that music, either in the form of music education, music practice, or exposure to music, can have a significant impact on school achievement, school attendance rates, and students’ conduct, both in elementary and secondary education (Koopman, 2005; Waller, 2007). Music education and exposure to music by listening or music-making would make children smarter and would have a positive influence on children’s motor development, social-emotional skills and even improve their chance of success in society (Bastian, 2002). The question is whether these claims are supported in available scientific studies. Besides music educa- tors and musicians, educational researchers have considered the question of what effects music education can have on child development. Some researchers claim to have found effects on cogni- tive functioning, such as an increase in concentration and academic achievement, in addition to effects in the social and emotional domain (Elliott, 1995; Gardner, 2004).

Hogenes, van Oers, and Diekstra (2015) conducted a literature review on the impact of music on child functioning. Their review shows that research literature on the impact of music on child functioning can be divided into three groups: (1) the influence of music on cognitive functioning; (2) the influence of music on social-emotional functioning; and (3) the influence of music on motor functioning. They identified 21 studies that met their inclusion criteria, such as the use of an experimental or quasi-experimental research design, and age range (3–18 years). Eighteen of the 21 studies focused on cognitive functioning (1) studies in which the influence of music is examined in relation to the academic performance of children (Bastian, 2002; Bolduc, 2009; Eastlund Gromko, 2005; Geoghegan & Mitchelmore, 1996; Rossini, 2000); (2) studies with regard to enhancement of cognitive task performance (including the so called “Mozart effect”) (Bilhartz, Bruhn, & Olson, 2000; Costa-Giomi 1999; Eastlund Gromko & Smith Poorman, 1998; Hallam, Price, & Katsarou, 2002; Hallam & Price 1998; Ivanov & Geake, 2003; Koutsoupidou & Hargreaves, 2009; McKelvie & Low, 2002; Schellenberg, 2004; Schellenberg, Nakata, Hunter, & Tamoto, 2007); and (3) intervention studies in which music has been investigated as facilitator of cognitive processes (Abikoff, Courtney, Szeibel, & Koplewicz, 1996; Furnham & Stephenson, 2007; Furnham & Strbac, 2002);. Of the 18 studies on cognitive functioning all, with three excep- tions (Costa-Giomi 1999; Eastlund Gromko & Smith Poorman, 1998; McKelvie & Low, 2002), reported positive or moderate positive effects of music education on academic achievement, pho- nemic awareness, spatial reasoning, creative thinking, and cognitive task performance. In the two studies on effects on social-emotional functioning (Bastian, 2002; Ulfarsdottir & Erwin, 1999) (one of these fell also in the category of studies on cognitive functioning) a positive effect of music education was observed on interpersonal problem solving, alternative solution thinking, consequential thinking, positive interactions in the classroom, social climate in the classroom and the school as a whole. In one of the studies on social-emotional skills, positive effects were only found in the long term. A significant difference with regard to alternative solution thinking and consequential thinking was shown at the follow-up test after 7 months. The two studies identified with regard to motor functioning showed positive effects on motor independency, jumping and dynamic balance (Palmer & Meyer, 2000; Zachopouloua, Tsapakidou, & Derric, 2004). Hogenes, Van Oers, and Diekstra concluded that music education and exposure to music appear to have a positive influence on child functioning. However, the diversity in, among others, research design,

34 International Journal of Music Education 34(1)

validity, dependent and independent variables of the reviewed studies made it difficult to draw robust conclusions.

The aim of the present study is to gather empirical evidence with regard to the effects of produc- tive music education on engagement in music education, and both music and academic achieve- ment. The rationale for studying the connections between music education and other academic achievements is primarily to assess whether music education contributes to development in gen- eral. Our main research question is: “What are the effects of music composition as a classroom activity on engagement in music education and on academic and music achievement?” The authors specifically investigated the effect on academic and music achievement of productive music edu- cation on elementary school students compared to a teacher-centered approach, mainly based on reproduction of music with regard to singing, playing instruments, and music and movement. Engagement in music education means that students are able to, and motivated to, participate in music activities.

Theoretical basis and prior research on music composition as a classroom activity

Activity theory

The theoretical framework for this study is the Cultural Historical Activity Theory (CHAT) of education (Cole, 1996; Karpov, 2005; van Oers et al., 2008). The starting point for this approach is Vygotsky’s assumption that education can promote students’ development by assisting them to appropriate relevant cultural tools that help them to become self-dependent participants in cultural practices (Vygotsky, 1978a; Vygotsky, 1981). More recent elaborations of the approach emphasize the importance of participation in specific cultural practices that serve as context for meaningful learning in participants (Lave & Wenger, 1991). At first newcomers in a cultural practice just play a peripheral role in the practice (for example, as an observer), but in due time these novices master the relevant tools with the help of more experienced members for the improvement of their abilities for participation.

Involvement in real-life cultural practices generally confronts participants with problems that arouse new needs for knowledge and abilities. By connecting the guided learning processes in the context of such practices to the personal needs of participants, the learning is going to make per- sonal sense for them, according to Leont’ev (1981), and as such contributes to the integration of the learning outcomes in a person’s identity. Moreover, we can assume that strengthening the per- sonal sense of the learning within practices will also contribute to the transformation of formal involvement in a practice into authentic personal engagement of the students with the practice (Lave & Wenger, 1991). Engagement with problems, people, and domains can have a synergistic effect (Stahl, 2006). Many dimensions can be distinguished with regard to engagement in learning. The nature of a problem given to students is critical. To get students to engage with a problem, the problem has to be meaningful for them (Leont’ev, 1981; Tolman, 1999), i.e., be functionally related to the practice. It has to involve issues that make sense to students within their interpretive perspec- tives on the world. It should also be a problem that challenges their current understanding, but is within reach of their understanding.

Music education

From the perspective of CHAT, music is a product of cultural history that always encompasses (by definition) a number of actually present or virtual co-actors. As such, music from this perspective

Hogenes et al. 35

is to be conceived as a form of distributed cognition (Cole & Engeström, 1993), produced in a col- laborative process with actual or virtual others.

As a cultural phenomenon music activity can be interpreted in terms of rule-based, goal directed, and tool-mediated actions with sounds. Such musical activity can take several forms, such as reproduction of previously composed music or production of new musical pieces (composing). According to CHAT, learning to take part in such cultural activities implies getting involved in the related cultural practices with culturally more experienced people who can guide the novice towards appropriation of actions or fundamental operations (like using music notations in the case of a music activity) that are deemed relevant by the music community involved. Hence, music education can be conceived as a cultural endeavor to get children collaboratively engaged in the musical practices of the community and assist them in appropriating the roles and related tools in order to enhance their participation in such roles, as listeners, singers, players of musical instru- ments, or composers. Most of the time, however, children’s involvement in music activities in schools are of the receptive kind (learning music composed by others). Like in other subject matter domains (e.g., reading and writing texts), we assume, however, that in music activities more pro- ductive versions of music-making (i.e. composing) may contribute in new and significant ways to children’s development. Starting out such learning activities from the children’s own musical imagination and giving them the (relative) freedom to compose their own songs, makes it a form of authentic learning. Activity theory (Leont’ev, 1981) provides us with a detailed theoretical lan- guage to describe and analyze the processes involved in such activity of composing (in terms of actions, personal sense, tools to use, rules to follow, goals to achieve, automatized operations, like audiation).

Music education in schools today includes several domains of musical behavior: singing, play- ing instruments, listening to music, music and movement, working with musical notation, and reflecting on listening and/or performance. Music composition can be added to this list, but can also be considered as part of, or as derivative of, the domains singing, playing instruments, and working with musical notation (Campbell & Scott-Kassner, 2006).

Apart from listening there is a cognitive process called audiation, necessary for the understand- ing of music. The term audiation was introduced in 1975 by music education researcher Edwin Gordon. Audiation can be considered as the most important process in making music or listening to music (Gordon, 2003). It is a high-level thought process that involves mentally hearing and comprehending music, even when no physical sound is present. It is a cognitive process by which a person gives meaning to musical sounds with the help of the brain. In essence, audiation of music is analogous to thinking in a language. The term audiation should not be confused with audition, the mere perception of sound. Audiation is also more than just a musical form of auditory imagery. Developed audiation includes the necessary understanding of music to enable the conscious pre- diction of patterns in unfamiliar music.

In the present study we will focus on the comparison between students’ involvement in musi- cal practices as a composer (music production) versus reproducer/performer of music. Composing music can be seen as an activity that is similar to writing texts (Hogenes, van Oers & Diekstra, 2014). Within CHAT, language is seen as an important tool for cognitive processes (van Oers, 2005; Vygotsky, 1978b). Moreover, it is a means for communication. Vygotsky mainly saw lan- guage as speech, in other words as a process of dialogically composing texts with communica- tive intentions (Vygotsky, 1981). Vygotsky also pointed out that the invention and mastery of written means for communication strongly improved the communicative possibilities of man- kind and even its thinking faculty (Scinto, 1986; Vygotsky, 1978b, pp. 105–119). A strong anal- ogy can be made between text composition and music composition. As in writing texts, in music composition students deal with problems of expressing main ideas, sequencing, classification,

36 International Journal of Music Education 34(1)

and categorizing, and need special technical tools in order to become proficient as composers. Musical tools, like music notation, help them to acquire these intellectual competencies (Ruthmann, 2007; Wiggins, 1990).

Although active music composition may be uncommon for most elementary school students, on the basis of our theory, we can expect that participation in composition activity with the help of an expert and peers, may enhance the student’s confidence in his or her possibilities to take part mean- ingfully in this activity (Mahn & John-Steiner, 2002), and even may stimulate his or her achieve- ment motive and engagement (Markova, 1983). One of the aims of the present study, next to the effects of composition on academic and music achievement, is to examine these assumptions.

The present study

Given the main research question concerning music composition as a classroom activity (produc- tive music education) a study was set up to examine the effects of music composition as a class- room activity on engagement in music education, music achievement, and academic achievement. In the present study two formats of music education: productive music education with composi- tion as a classroom activity as core activity (designated as experimental condition), and a teacher- centered approach mainly based on reproduction of music (designated as control condition) were compared. Active music listening, and music and movement were important elements of both interventions.

Method

Design—procedure

A randomized groups pre-test–post-test–follow-up design was used for this study. Although the students were not randomly assigned to the experimental and control group, the classes were. The two music interventions were implemented in 18 weekly lessons of 45 minutes each. The lessons were given on a weekly basis in the period September 2010 through February 2011. Pre-test data on singing, listening, intelligence, language, reading comprehension, and mathematics were col- lected during the first 2 weeks of the school year, starting in September 2010. Post-test data on the same variables were collected 6 months later, right after the intervention. Follow-up data were collected at the end of the school year, 5 months after the intervention (July 2011). All measures were group-administered to students in the intervention and control groups by the first author. Data were collected and analyzed by means of standardized tests. At post-test, besides the variables mentioned above, engagement was measured by using a questionnaire. In many Dutch elementary schools, such as De Vijver, music is mainly used as a means during social occasions, like birthday celebrations. However, music education was no part of the school program. As the students had no systematic experience in music education before the intervention, it was not possible to conduct the questionnaire as pre-test. The intervention and the measures are described below.

Participants

All participants were students attending the elementary school De Vijver, located in the City of The Hague, comprising 500 students in 21 classes. The school works on the basis of the Dalton concept (Parkhurst, 1922/2007). The Dalton concept has been developed by Parkhurst, and is based on three principles: (1) freedom (students may choose from a limited list of optional tasks. They have freedom to choose the time on which, and the tempo in which, they work on certain subjects); (2)

Hogenes et al. 37

cooperation (cooperation refers to the social character of learning and knowledge); and (3) assign- ments (students have to plan and execute tasks independently without much guidance from an adult) (Parkhurst, 1922/2007).

The experimental group consisted of 31 girls (49.2%) and 35 boys (53%) comprising one fifth- grade class and two sixth-grade classes at the elementary school De Vijver in The Hague, the capital city of the province of Zuid-Holland, The Netherlands. The control group consisted of 33 girls (50.8%) and 32 boys (49.2%) comprising two fifth-grade classes and one sixth-grade class at the same school as the students of the experimental group. The music instructor for both interventions was the first author of this article (male, 40 years old, 17 years of instructional practice, who majored in music performance [electronic organ and keyboards], and music in education). The students in the experimental group had a mean age of 9.38 years (SD = .69). The students in the control group had an overall mean age at pre-test of 8.92 years (SD = .63). The vast majority of the participants were native speakers of Dutch, with 23 participants (13 in the intervention group, 10 in the control group) having a first language other than Dutch. In both groups, language and mathematics were taught the same way. The average class size was 22 students in both conditions.

For the analysis of post-test and follow-up data the N differs between measures depending upon differences in missing data. The number of missing data varied from 7.7% to 10.6%.

Interventions

The experimental group (N = 66) was involved in a program constructed according to a CHAT approach. In this intervention students were actively engaged in the activity of composing music, in which musical notation (as a sign system) was a helpful tool for organizing the activity, provid- ing a means for effective communication about the object of the activity (in other words the music piece they were composing), and regulating specific actions to be carried out (Jones, 2011, for the role of signs in activities). Reflection played a crucial role here to keep track of the process and its (intermediate) outcomes. The music composition and music notation activities took an average of 30 minutes (2/3) per session. The experimental group worked collaboratively at music composition activities. The group work can be characterized by simultaneous interaction, positive interdepend- ence, and individual accountability (Kagan, 1994; Slavin, 1983).

In contrast to the experimental group in which music composition as a form of productive music-making was central, the control group focused on reproductive music-making. The control group (N = 65) was taught from a teacher-centered approach. In this setting composing as a class- room activity played a minor role. The students of the control group mainly sang songs composed by songwriters, played music written by composers, did a lot of music and movement (activities in which aspects of sound and form are represented by movement and dance activities). The sing- ing, playing and movement took an average of 30 minutes (2/3) per session. Both groups used all kinds of musical notation (traditional music notation, graphical and pictorial notation). In both interventions the production of music was the core activity, but the experimental intervention focused on the child as composer (music production), while the control intervention focused on the child as performer (music reproduction). One could also speak of production versus re-pro- duction. Manuals for both interventions are available from the first author of this article. The manuals contain complete lesson plans with goals, activities, and guidelines for materials and classroom layout. Before the actual experimental study was carried out a pilot study was under- taken to examine the feasibility of the proposed methods and design. In 10 weekly sessions of 45 minutes two classes of 9 and 10 year olds participated in music lessons. One group executed a sequence of lessons comparable with the intervention of the experimental group of the current study. The other group executed a sequence of lessons comparable with the intervention of the

38 International Journal of Music Education 34(1)

control group of this study. The pilot study showed the authors the importance of the first step of the three-step-model developed for music composition as a classroom activity, which will be described below: the creation of a common basis in order to start the process of music composi- tion. Although the authors spent time creating a common basis in music composition activities in the pilot study, the executed activities demonstrated a need for extensive attention to this phase in the process. A second issue that came up in the pilot study concerned the second step of the music composition process: creating ideas and writing the composition. Part of this step is the revision phase. The revision of a piece of music always focuses on the goal to make students think about their compositions and to help them to improve the draft version of their composition (Hogenes et al., 2014). Revision of music composition has been widely described (Hickey, 2012; Kaschub and Smith, 2009, 2013; Kratus, 2012; Wiggins, 1990) The authors consequently became aware of the need to revise pieces in three rounds: (1) on the ideas of the composer and the content of the piece; (2) the construction of the piece and its style; (3) the notation of the music. Observations showed that insufficient attention to one of the rounds and/or changing the order of the rounds delivered less interesting music compositions (e.g., insufficient use of aspects of sound and form), as well as reducing the student’s engagement.

Research questions and hypotheses

This study addressed the following research questions:

1. What difference can be found between the effects of a music education intervention based on music composition as a classroom activity versus a music education intervention based on a teacher-centered approach mainly comprising reproduction of music on students’ engagement in music education?

2. What difference can be found between the effects of a music education intervention based on music composition as a classroom activity versus a music education intervention based on a teacher-centered approach mainly comprising reproduction of music on intelligence, academic achievement, and music achievement?

Based on previous research (Bodovski & Farkas, 2007; Engeström, Engeström & Suntio, 2002; Marks, 2000), the authors of this article expected that students in the experimental group would be more deeply engaged in music activities in the classroom than the control group (Hypothesis 1). They would outperform on nonverbal intelligence compared with the control group (Hypothesis 2). Moreover, it was expected that students of the experimental group would perform better with regard to academic skills than the control group (Hypothesis 3). Furthermore, the authors expected the students of the experimental group to develop better musical abilities, especially listening and audiation, than the control group (Hypothesis 4). The reason for this expectation is that composi- tion as a classroom activity may demand more high-level thought processes that involve mentally hearing and comprehending music (audiation) than performing music (Gordon, 2003).

Measures

Overview. To determine the effects of the two interventions with respect to children’s academic performance, intelligence, and music performance, a battery of tests was used: CITO1 Language Test (Spelling), CITO Reading Test, CITO Mathematics Tests, Raven Standard Progressive Matri- ces (SPM), and a Musical Abilities Test for Singing and Listening (Hogenes, van Oers & Diekstra, 2010a, 2010b). The CITO tests are part of the student monitoring system of the school. An

Hogenes et al. 39

intelligence test was included because relations between music education and intelligence have been found in previous studies (Bastian, 2002). The Raven Standard Progressive Matrices is a valid nonverbal intelligence test that can be used for group testing and was therefore suitable for this study. The Musical Abilities Test for Singing and Listening was designed for this study.

For this study, a pre- and post-test design was used. The pre-test was administered two weeks before the interventions started. The pre-test consisted of the Raven Standard Progressive Matrices, and the Musical Abilities Test for Singing and Listening. Also data from the student monitoring system were part of the pre-test. The post-test took place in the week upon completion of the inter- vention. The measurements from the pre-test were identically administered during the post-test. Additionally, a questionnaire was administered to measure students’ experiences with and engage- ment in the interventions. The follow-up was conducted five months after the post-test. Measurements were the same as at the post-test with the exception of the experience and engage- ment questionnaire.

CITO, language test, spelling. For this study a CITO language test for spelling was used. The test has to be completed twice a year. The first test has to be conducted in the second half of the month of January or the first week of February, and the second test at the end of the school year. The test is part of the CITO student monitoring system. The spelling test assesses the ability of children to put words into word images. In grades 5 and 6 mainly two- and three-syllable words are tested for each year in 19 spelling categories. The tested words reflect the curriculum of the most frequently used language and spelling methods. The tested words are partly presented in the form of word and sentence dictation, and partly in the form of multiple-choice assignments (De Wijs & Krom, 2008; De Wijs, Krom & Van Berkel, 2007).

CITO, reading tests. For this study two CITO reading tests were used: technical reading and reading comprehension. Both sets of tests are completed twice a year. Both tests are part of the CITO stu- dent monitoring system.

The technical reading tests measure the accuracy and tempo of reading (Jongen, Krom & Roumans, 2009, 2010). The reading comprehension tests provide an insight into the development of reading skills of students and the differences between students. The tests cover the main objec- tives and their intermediate targets. The tests consist of multiple-choice questions with regard to texts the students have to read. Every test has two follow-up tests: an easier and a more complex one to make it possible to test the students in an adaptive way (Feenstra, 2008; Feenstra, Krom & Berkel, 2007).

CITO, mathematics test. The CITO Rekenen-Wiskunde [Arithmetic-Mathematics] tests were used to assess the children’s progression during this subject. The assignments concern three domains: (1) numbers and operations; (2) measurement, geometry, time and money; and (3) ratios, fractions, and percentages. The tests were conducted twice a year. The tests are part of the CITO student monitoring system (Jansen, Scheltens, & Kraemer, 2006a, 2006b).

Raven, Standard Progressive Matrices (SPM). The Raven SPM test comprised 60 items arranged in 5 sets (A, B, C, D, and E) of 12 items each. It was used to measure the students’ ability to form per- ceptual relations and to reason by analogy independent of language and formal schooling. The matrices consist of a series of patterns in upward progression. The first item of each set is almost natural to solve. The subsequent items build on the reasoning of the previous item and become more and more difficult. By solving the items in the given order, the method necessary to solve the following items is learned automatically. The five sets offer five opportunities to learn the

40 International Journal of Music Education 34(1)

necessary to solve the items and five progressive ways to measure the child’s intellectual ability (Raven, 2004). In the literature reliability coefficients (Cronbach’s alpha) are commonly reported to fall within the .80–.90 range (Raven, 2004). Our data deliver Cronbach’s alpha T0 α = .81, T1 α = .82, T2 α = .79.

Musical ability test, singing. During the first music lesson the song “Aap in de boom” [Monkey in the tree] was taught to the students in both the experimental and the control group. The authors of this study wanted to assess the students’ musical abilities as much as possible during the lessons accord- ing to the concept of dynamic assessment (Lidz & Elliott, 2000; Sternberg & Grigorenko, 2002; Tzuriel, 2001). This approach is based on the assumption that the level of (musical) ability can be measured by assessing how easy a person finds it to learn something new with more or less help. In small groups of maximum four students they sang this song after the lesson. During the assess- ment the song was rehearsed one time without filming. Then it was sung for a second time, the moment of assessment of learning. The singing was taped on video and marked by two assessors (Interrater reliability at T0, Melody, Kappa = .786; Rhythm, Kappa = .832; Comprehensibility, Kappa = .866; Expression, Kappa = .908).

The children’s singing was scored on four items: correct melody performance, correct rhythm performance, comprehensibility, and expression. All items were scored on a 5-point scale: 1 = the student is unable to demonstrate …; 2 = the student is beginning to demonstrate …; 3 = the student is developing …; 4 = the student consistently demonstrates …; and 5 = the student exceeds their competency on the task (Hogenes et al., 2010a). Singing was assessed before the intervention started and right after the intervention and 5 months after the post-test (pre-, post- and follow-up tests).

Musical ability test, listening. During the first music lesson the students’ listening abilities were assessed. The assessment was based on a measure developed by the authors and covering three domains: instrument discrimination/recognition, ensemble discrimination/recognition, and audiation.

The instrument discrimination/recognition part contained eight items (such as saxophone, and double bass). The ensemble discrimination/recognition parts contained four items, and the student had to fill out eight items of the audiation game (Hogenes et al., 2010b). The Cronbach’s alphas for these measures were less than satisfactory (instrument and ensemble discrimination/recognition T0 α = -.20, T1 α = .19, T2 α = .36, audiation T0 α = .57, T1 α = .43, T2 α = .07). Therefore it was decided to exclude the listening tests from the analysis.

Engagement test. At post-test an Experience and Engagement questionnaire was administered. This questionnaire was developed by the authors and contains 16 questions and statements focusing on children’s experience and engagement in music education. Examples of statements used are: “I became curious to how music is constructed by composing music myself”; “Most of the time, we could improve our compositions by looking to our compositions (revising) together with the teacher”; “There was lots of ‘space’ for my own ideas in the music classes.” All questions were assessed with Likert-type scales. The reliability of the questionnaire appeared to be highly satisfac- tory (Cronbach’s alfa α = .96).

Results

Below the results will be presented as follows. First of all the effect of composition as a classroom activity on engagement in music education, followed by the effect of composition on intelligence, academic performance (language (spelling), mathematics and reading comprehension), singing (melody, rhythm, comprehensibility and expression) will be presented.

Hogenes et al. 41

Table 2. Correlations among variables at post-test.

Variable 1 2 3 4 5 6 7 8

1. Engagement – 2. Intelligence .27** .739*** 3. Language (spelling) .36*** .39*** .816*** 4. Mathematics .23*** .64*** .51*** .890*** 5. Reading comprehension .02 .51*** ,32*** .42*** – 6. Singing (melody) −.11 .03 .10 .02 .11 .572*** 7. Singing (Rhythm) −.19* −.00 .03 −.09 .15 .55*** .658*** 8. Singing (comprehensibility

and expression) −.27** −.02 −.17 −.12 .15 .40*** .42*** .688***

Mean 0.00 36.78 130.79 70.99 2.79 3.30 3.75 6.51 SD 1.00 7.89 9.20 15.83 1.27 .67 .54 1.31

*p < .05. **p < .01. ***p < .001. The diagonal shows correlations between post-test and follow-up test.

Correlations among variables at pre-test, post-test, and follow-up tests respectively, are shown in Tables 1, 2 and 3. At pre-test, spelling correlated positively with intelligence, as did mathematics and reading comprehension. Moreover, singing (comprehensibility and expression) correlated positively with singing (rhythm). At post-test, the same positive correlations were found, except for reading comprehension. Finally, engagement correlated positively with spelling, mathematics, and reading comprehension, but correlated negatively with all aspects of singing.

Table 4 shows descriptive statistics (means and standard deviations) for all measured variables at pre-test, post-test and follow-up test in both conditions (intervention and control group). All score distributions were approximately normal and, thus, appropriate for use in parametric statisti- cal analyses.

To address the first research question concerning the effect of composition as a classroom activ- ity on engagement in music education, an analysis of covariance (ANCOVA) was performed with age used as covariate, and with condition (intervention vs. control) as the independent variable. Levene’s test indicated that the assumption of homogeneity of variances was violated (p = .002).

Table 1. Correlations among variables at pre-test.

Variable 1 2 3 4 5 6 7

1. Intelligence .745*** 2. Language (spelling) .24** .818*** 3. Mathematics .57*** .44*** .867*** 4. Reading comprehension .31** .29** .35*** .753*** 5. Singing (melody) .08 −.12 −.08 .18 .425*** 6. Singing (rhythm) −.01 .15 −.09 .05 .06 .322*** 7. Singing (comprehensibility and expression)

.07 −.06 −.02 .16 .04 .41*** .661***

Mean 34.42 126.00 62.67 3.03 1.96 2.63 4.86 SD 8.13 8.66 16.35 1.18 .700 .78 1.12

*p < .05. **p < .01. ***p < .001. The diagonal shows correlations between pre-test and post-test.

42 International Journal of Music Education 34(1)

But ANCOVA is robust to mild violation of the assumption with roughly equal groups sizes (Rogan & Keselman, 1977). For all consecutive ANCOVA models assumption of homogeneity of

Table 3. Correlations among variables at follow-up test.

Variable 1 2 3 4 5 6

1. Intelligence .779*** 2. Language (spelling) .32*** .807*** 3. Mathematics .64*** .55*** .857*** 4. Singing (melody) −.11 .14 −.09 .237** 5. Singing (Rhythm) .05 .16 −.06 .43*** .275** 6. Singing (comprehensibility

and expression) .17 .25 .003 .43*** .56*** .491***

Mean 36.17 135.41 78.09 3.18 3.64 6.83 SD 7.93 8.80 14.84 .478 .56 1.08

*p < .05. **p < .01. ***p < .001. The diagonal shows correlations between follow up-test and pre-test.

Table 4. Descriptive statistics for variables at pre-test, post-test and follow-up-test.

Intervention

Pre-test Post-test Follow-up test

Measure M SD M SD M SD

Engagement – – 1.00 .27 – – Intelligence 36.43 7.21 39.15 7.18 37.28 7.77 Language, spelling 128.74 8.96 133.85 9.09 136.98 8.29 Reading comprehension 2.96 1.13 2.93 1.16 – – Mathematics 67.47 15.31 75.83 15.13 82,57 13,28 Singing, melody 2.03 .78 3,27 .69 3.18 .50 Singing, rhythm 2.32 .85 3.65 .63 3.57 .65 Singing, comprehensibility & expression

4.48 1.19 6.20 1.31 6.63 1.10

Control

Pre-test Post-test Follow up-test

Measure M SD M SD M SD

Engagement – – −.94 .19 – – Intelligence 32.29 8.47 34.52 7.92 35.22 8.16 Language, spelling 123.50 7.07 127.67 7.83 133.97 8.60 Reading comprehension 3.14 1.22 2.66 1.32 – – Mathematics 58.05 16.37 66.46 15.53 73.36 15.32 Singing, melody 1.89 .62 3.35 .65 3.18 .46 Singing, rhythm 2.92 .57 3.85 .40 3.72 .45 Singing, comprehensibility & expression

5.20 .94 6.81 1.25 7.03 1.05

Hogenes et al. 43

variances has been met. Moreover, for none of the ANCOVA models the covariates interacted with the condition (p > .05), indicating that the data met the assumption of homogeneous regression slopes.

Moreover, for none of the ANCOVA models the covariates interact with condition (p > .05), indicating that the data met the assumption of homogeneous regression slopes. Table 5 shows the adjusted marginal means and standard errors for the dependent variables. The ANCOVA showed a statistically significant effect of condition on the engagement measure after adjustment by covari- ates. That is, consistent with the hypothesis, students participating in the composition as a class- room activity intervention (M = 1.00, SD = 0.27) indicated more engagement in music activities than did students in the control condition (M = -0.94, SD = 0.19), with F(1, 121) = 1805.29, p < .001, partial η2 = .937. The covariate of age, F(1, 121) = .29, p = .59, partial η2 = .002 did not reach a conventional level of statistical significance. This dependent variable was measured once after the intervention.

The second research question concerned the effect of composing music as a classroom activ- ity on intelligence. The ANCOVA showed no statistically significant effect of condition on intelligence measured at the post-test after adjustment for covariates: age and intelligence measured at the pre-test. The ANCOVA showed no statistically significant effect of condition on intelligence measured at the follow-up test after adjustment for covariate intelligence meas- ured at the pre-test.

Table 5. Adjusted marginal means and standard errors for dependent variables at post-test and follow-up test.

Measure Intervention Control

Post-test M SE M SE

Engagement 1.0 .03 −.93 .03 Intelligence 37.63 .71 35.97 .69 Language, spelling 131.51 .72 129.88 .70 Reading comprehension 3.01 .11 2.58 .11 Mathematics 71.95 1.10 70.22 1.08 Singing, melody 3.24 .08 3.38 .08 Singing, rhythm 3.75 .07 3.75 .07 Singing, comprehensibility & Expression

6.62 .13 6.43 .12

Follow up-test Intervention Control

M SE M SE

Engagement – – – – Intelligence 35.98 .67 36.47 .65 Language, spelling 134.57 .72 136.18 .68 Language, vocabulary 63.27 2.65 61.47 1.63 Reading comprehension – – – – Mathematics 78.73 1.07 77.01 1.04 Singing, melody 3.14 .06 3.23 .06 Singing, rhythm 3.68 .08 3.62 .07 Singing, comprehensibility & expression

6.84 .13 6.84 13

44 International Journal of Music Education 34(1)

Next, to address the third research question, concerning the effects of the intervention on aca- demic achievement, a MANCOVA with language (spelling), reading comprehension, and mathe- matics being dependent variables, controlling for age and pre-test results was performed, followed by three separate ANCOVAs. In the MANCOVA and ANCOVA models all covariate pre-test scores were statistically significant, p < .005. The results of the MANCOVA tests with the use of Wilks’ criterion indicates a marginally significant effect of condition on combined dependent vari- ables, Wilks’ λ = .933, F(3, 104) = 2.48, p = .065, partial η2 = .067.

Follow-up separate ANCOVAs showed a significant effect of the condition on one out of three academic achievement indicators. The ANCOVA showed a statistically significant (F(1, 112) = 7.42, p = .007, partial η2 = .06) effect of condition on reading comprehension measured at the post-test after adjustment for covariates: age and pre-test score of reading comprehension. The covariate of age, F(1, 112) = .22, p = .638, partial η2 = .00 and did not reach a conventional level of statistical significance.

The results for both language (spelling) and mathematics have shown higher average scores in the experimental group than in the control group, controlling covariates age and pre-test scores but these differences were not statistically significant at p < .05.

To address the effects of the intervention on academic achievement at the follow-up, we per- formed a MANCOVA with language (spelling) and mathematics being dependent variables, con- trolling for pre-test results. Reading comprehension is not used as a dependent variable due to the fact that there were no data available at the follow-up. The results of the MANCOVA tests did not show a significant effect of condition on combined dependent variables (p = .101).

The fourth research question, concerning the effects of the intervention on children’s music achievement, was addressed through analyses of variance focused on effects of intervention on the performance of singing: (1) melody; (2) rhythm; (3) comprehension and expression.

The results of the MANCOVA tests with the use of Wilks’ criterion did not show a significant effect of condition on singing achievement, Wilks’ λ = .996, F(3, 117) = 0.17, p = .92, partial η2 = .004. The effects of the intervention on children’s music achievement at the follow-up also did not show a significant effect of condition on singing p <. 05.

Conclusions and discussion

The overall results of the present study comparing two types of music education, an intervention that emphasizes music production by composition and an intervention with an emphasis on music reproduction, showed the following. First of all, the study demonstrates positive effects on students’ engagement in both types of music education, but greater effects in the music pro- duction condition, which confirms the hypothesis that music production would lead to more engagement in students than music reproduction. Secondly, this study does not support the hypothesis that music education as such contributes to nonverbal intelligence. Also, no differ- ences in this respect were found between students in the music production condition versus students in the reproduction condition. This deviates from the findings reported by Bastian (2002) in which significant positive effects of music education on intelligence were reported. Thirdly, the findings of this study partly confirmed the hypothesis that the students of the experi- mental group would perform better with regard to academic skills than the control group. The students of the experimental group performed better with regard to academic skills than their counterparts in the control group; at least, such an effect was found for reading comprehension. It remains uncertain to what this difference should be attributed. One possibility is that students in the experimental group made extensive use of symbolic notation and were more focused on the text-like dimensions of their compositions. A transfer to reading comprehension could have

Hogenes et al. 45

taken place. Fourth, although both groups showed progress with regard to singing, no significant difference between both groups was found in this respect, despite the fact that the control group sang much more than the experimental group. This might indicate, since relevant variables here are melody and rhythm performance as well as expression and comprehensibility, that these can apparently also be improved by non-singing music activities.

In conclusion, this study highlights the surplus value on several dimensions of composition as a classroom activity, such as on engagement and on academic abilities, such as reading com- prehension. This study has shown that music composition is feasible and useful in elementary school. Students are able to compose music in the same way as they are able to sing songs, play instruments, and perform dances. The authors of this study conclude that productive music education is evidently more engaging for students than reproductive forms of music education. However, productive music education requires teachers to have different pedagogical, didacti- cal, organizational, and reflective skills than reproductive music education. For example, using forms of cooperative learning and differentiated instruction. This has consequences for teacher education.

The value of these findings are substantiated by the fact that the current study meets scientific standards (Slavin, 2008) with regard to: (1) randomized assignment: classes were randomly assigned to the experimental and control group; (2) sample size: 131 students participated in this study; and (3) duration: the study lasted 10 months.

Nevertheless, there are a number of limitations to this study. First of all, randomization was only at the class level and not at the level of students or school. Secondly, the same teacher carried out both interventions. A possible proclivity towards one or the other approach to music education can- not be excluded as having affected the outcomes. Third, it can also not be excluded that differences between both conditions had to do with factors other than the presence or absence of composition as an activity. For example, music composition might require different forms of interaction between teacher and students than reproductive music activities. For example, in music composition activi- ties students get more individual feedback and feedback in small groups than in music reproduc- tion activities where the class is addressed as a whole.

Notwithstanding these possible limitations, the findings with regard to engagement and certain academic skills, such as reading comprehension skills, as a result of music education through com- position, are intriguing enough to warrant further research and reconsideration of the content of music education in other elementary school settings and with other teachers.

Acknowledgements

The authors of this article would like to thank the students, teachers and staff of elementary school ‘De Vijver’ to make it possible to conduct this study with their full and enthusiastic cooperation.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Note

1. CITO is a Dutch testing and assessment company. Measuring and monitoring human potential has been their core competence since 1968. The vast majority of Dutch elementary and secondary schools use tests developed by CITO. CITO does not provide information about the reliability of their tests. Considering the status of the institute we can safely assume that the tests are reliable and valid. CITO tests are com- parable with what in the USA is designated School Aptitude Tests (SAT).

46 International Journal of Music Education 34(1)

References

Abikoff, H., Courtney, M. E., Szeibel, P. J., & Koplewicz, H. S. (1996). The effects of auditory stimulation on the arithmetic performance of children with ADHD and nondisabled children. Journal of Learning Disabilities, 29(3), 238–246.

Bastian, H. G. (2002). Musik(erziehung) und ihre Wirkung. Eine Langzeitstudie an Berliner Grundschulen [Music education and its effects. A long-term study in elementary schools in Berlin]. Mainz: Schott.

Bilhartz, T. D., Bruhn, R. A., & Olson, J. E. (2000). The effect of early music training on child cognitive development. Journal of Applied Developmental Psychology, 20(4), 615–636.

Bodovski, K., & Farkas, G. (2007). Mathematics grow in early elementary school: The roles of beginning knowledge, student engagement, and Instruction. The Elementary School Journal, 2(108), 115–130.

Bolduc, J. (2009). Effects of a music programme on kindergartners’ phonological awareness skills. International Journal of Music Education, 27(1), 37–47.

Campbell, P. S., & Scott-Kassner, C. (2006). Music in childhood: From preschool though the elementary grades. Boston, MA: Schirmer.

Cole, M. (1996). Cultural psychology: A once and future discipline. Cambridge, MA: Belknap/Harvard University Press.

Cole, M., & Engeström, Y. (1993). A cultural-historical approach to distributed cognition. In G. Salomon (Ed.), Distributed cognitions: Psychological and educational considerations (pp. 1–46). New York: Cambridge University Press.

Costa-Giomi, E. (1999). The effects of three years of piano instruction on children’s cognitive development. Journal of Research in Music Education, 47(3), 198–212.

de Wijs, A., & Krom, R. (2008). Spelling groep 6. Handleiding [Spelling group 6. Manual]. Arnhem: Cito. de Wijs, A., Krom, R., & Berkel, S. van (2007). Spelling groep 5. Handleiding [Spelling group 5. Manual].

Arnhem: Cito. Eastlund Gromko, J. (2005). The effect of music instruction on phonemic awareness in beginning readers.

Journal of Research in Music Education, 53(3), 199–209. Eastlund Gromko, J., & Smith Poorman, A. (1998). The effect of music training on preschoolers’ spatial-

temporal task performance. Journal of Research in Music Education, 46(2), 173–181. Elliott, D. J. (1995). Music matters: A new philosophy of music education. New York: Oxford University

Press. Engeström, Y., Engeström, R., & Suntio, A. (2002). Can a school community learn to master its own future?

An activity-theoretical study of expansive learning among middle school teachers. In G. Wells & G. Claxton (Eds.), Learning for life in the 21st Century (pp. 211–224). Malden, MA: Blackwell.

Feenstra, H. (2008). Begrijpend lezen groep 6. Handleiding [Reading comprehension group 6. Manual]. Arnhem: Cito.

Feenstra, H., Krom, R., & Berkel, S. van (2007). Begrijpend lezen groep 5. Handleiding [Reading compre- hension group 5. Manual]. Arnhem: Cito.

Furnham, A., & Stephenson, R. (2007). Musical distracters, personality type and cognitiveperformance in school children. Psychology of Music, 35(3), 403–420.

Furnham, A., & Strbac, L. (2002). Music is as distracting as noise: the differential distraction of back- ground music and noise on the cognitive test performance of introverts and extraverts. Ergonomics, 45(3), 203–217.

Gardner, H. (2004). Frames of mind: The theory of multiple intelligences. New York: Basic Books. Geoghegan, N., & Mitchelmore, M. (1996). Possible effects of early childhood music on mathematical

achievement. Australian Research in Early Childhood Education,1, 1–9. Gordon, E. E. (2003). Learning sequences in music: Skill, content, and patterns. A Music Learning Theory.

Chicago: GIA. Hallam, S., Price, J., & Katsarou, G. (2002). The effects of background music on primary school pupils’ task

performance. Educational Studies, 28(2), 111–122. Hallam, S., & Price, J. (1998). Can the use of background music improve the behaviour and academic per-

formance of children with emotional and behavioural difficulties? British Journal of Special Education, 25(2), 88–91.

Hogenes et al. 47

Hickey, M. (2012). Music outside the lines: Ideas for composing in K–12 music classrooms. Oxford: Oxford University Press.

Hogenes, M., van Oers, B., & Diekstra, R. F. W. (2010a). Musical Ability Test: Singing. The Hague: The Hague University. Unpublished.

Hogenes, M., van Oers, B., & Diekstra, R. F. W. (2010b). Musical Ability Test: Listening. The Hague: The Hague University. Unpublished.

Hogenes, M., van Oers, B., & Diekstra, R. F. W. (2015). The impact of music on child functioning. The European Journal of Social & Behavioural Sciences, 135, 1507–1526.

Hogenes, M., van Oers, B., & Diekstra, R. F. W. (2014). Music composition in the music curriculum. US-China Education Review A, 2(3), 149–162.

Jansen, J., Scheltens, F., & Kraemer, J. M. (2006a). Rekenen-Wiskunde groep 5. Handleiding [Arithmetic- Mathematics group 5. Manual]. Arnhem: Cito.

Jansen, J., Scheltens, F., & Kraemer, J. M. (2006b). Rekenen-Wiskunde groep 6. Handleiding [Arithmetic- Mathematics group 6. Manual]. Arnhem: Cito.

Jones, P. E. (2011). Signs of activity: Integrating language and practical action. Language sciences, 33(1), 11–19

Jongen, I., Krom, R., & Roumans, P. (2009). Technisch lezen groep 5. Handleiding [Technical reading group 5. Manual]. Arnhem: Cito.

Jongen, I., Krom, R., & Roumans, P. (2010). Technisch lezen groep 6. Handleiding [Technical reading group 6. Manual]. Arnhem: Cito.

Kagan, S. (1994). Cooperative learning. San Clemente: Kagan. Karpov, Y. V. (2005). The neo-Vygotskian approach to child development. Cambridge: Cambridge University

Press. Kaschub, M., & Smith, J. (2009). Minds on music: Composition for creative and critical thinking. Lanham,

MD: MENC. Kaschub, M., & Smith, J. (2013). Composing our future: Preparing music educators to teach composition.

New York: Oxford University Press. Koopman, C. (2005). Muziek maakt slim? Over de rechtvaardiging van muziekonderwijs [Music makes

smart? About the justification of music education]. In J. Herfs, R. van der Lei, E. Riksen & M. Rutten (Eds.), Muziek Leren. Handboek voor basis- en speciaal onderwijs [Learning music. Handbook for ele- mentary and special education] (pp. 19–35). Assen: Van Gorcum.

Koutsoupidou, T., & Hargreaves, D.J. (2009). An experimental study of the effects of improvisation on the development of children’s creative thinking in music. Psychology of Music, 37(3), 251–278.

Kratus, J. (2012). Nurturing the songcatchers: Philosophical issues in the teaching of music composition. In W. D. Bowman & A. L. Frega (Eds.), The Oxford handbook of philosophy in music education (pp. 367–385). New York: Oxford University Press.

Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge: Cambridge University Press.

Leont’ev, A. N. (1981). The problem of activity in psychology. In J. Wertsch (Ed.), The concept of activity in Soviet psychology (pp. 37–71). Armonk, NY: Sharpe.

Lidz, C, & Elliott, J, (Eds.). (2000). Dynamic assessment: Prevailing models and applications, Vol. 6. New York: JAI Press.

Mahn, H., & John Steiner, V. (2002). The gift of confidence: A Vygotskian view of emotions. In G. Wells & G. Claxton (Eds.), Learning for life in the 21st century. Sociocultural perspectives on the future of education (pp. 46–58). London: Blackwell

Markova, A. K. (1983). Formirovanie motivacii ucenija v škol’nom vozraste [The formation of motivation to learn during school years]. Moscow, Prosvešcenie.

Marks, H. M. (2000). Student engagement in instructional activity: Patterns in elementary, middle, and high school years. American Educational Research Journal, 20(37), 153–184.

McKelvie, P., & Low, J. (2002). Listening to Mozart does not improve children’s spatial ability: Final cur- tains for the Mozart effect. British Journal of Developmental Psychology, 20, 241–258.

48 International Journal of Music Education 34(1)

Palmer, C., & Meyer, R. K. (2000). Conceptual and motor learning in music performance. Psychological Science, 11(1), 63–68.

Parkhurst, H. (1922/2007). Education on the Dalton Plan. New York: E. P. Dutton & Company. Raven, J. C. (2004). Raven: Standard progressive matrices. Manual. The Netherlands: Pearson Assessment

and Information. Rogan, J. C., & Keselman, H. J. (1977). Is the ANOVA F-Test robust to variance heterogeneity when sample

sizes are equal? An investigation via a coefficient of variation. American Education Research Journal, 21(14), 493–498.

Rossini, J. W., Jr. (2000). A study of the relationship of music instruction and academic achievement among elementary school students. Unpublished doctoral dissertation, The Graduate School of Education, Boston College, USA.

Ruthmann, A. (2007). The composers’ workshop: An approach to composing in the classroom. Music Educators Journal, 93, 38–43.

Scinto, L. F. M. (1986). Written language and psychological development. New York: Academic Press. Schellenberg, E. G. (2004). Music lessons enhance IQ. Research report. Psychological Science, 15(8). Schellenberg, E. G., Nakata, T., Hunter, G., & Tamoto, S. (2007). Exposure to music and cognitive perfor-

mance: tests of children and adults. Psychology of Music, 35(5), 5–19. Slavin, R. E. (1983). Cooperative learning. New York: Longman. Slavin, R. E. (2008). Perspectives on evidence-based research education. What works? Issues in synthesizing

educational program evaluations. Educational Researcher, 37(1), 5–14. Stahl, G. (2006). Group cognition: Computer support for building collaborative knowledge. Cambridge, MA:

MIT Press. Sternberg, R., & Grigorenko, E. (2002). Dynamic testing: The nature and measurement of learning potential.

Cambridge: Cambridge University Press. Tolman, C. (1999). Society versus context in individual development: Does theory make a difference? In Y.

Engeström, R. Miettinen & R. L. Punamäki (Eds.), Perspectives on activity theory (pp. 70–86). New York, NY: Cambridge University press.

Tzuriel, D. (2001). Dynamic assessment of young children. New York/ Dordrecht: Plenum Publishers/Kluwer Academic.

Ulfarsdottir, L. O., & Erwin, P. G. (1999). The influence of music on social cognitive skills. The Arts in Psychotherapy, 26(2), 81–84.

van Oers, B. (2005). Dwarsdenken. Essays over Ontwikkelingsgericht Onderwijs [Cross thinking. Essays on Developmental Education]. Assen: Van Gorcum.

Vygotsky, L. S. (1978a). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.

Vygotsky, L. S. (1978b). The prehistory of written language. In L. S. Vygotsky, Mind in society: The devel- opment of higher psychological processes (pp. 105–119). Cambridge, MA: Harvard University Press.

Vygotsky, L. S. (1981). The genesis of higher mental functions. In J. V. Wertsch (Ed.), The concept of activity in Soviet psychology (pp. 144–188). Armonk: Sharpe.

Waller, G. D. (2007). The impact of music education on academic achievement, attendance rate, and student conduct on the 2006 senior class in one Southeast Virginia Public School division. Unpublished doctoral dissertation, Virginia Polytechnic Institute and State University, Blacksburg.

Wallin, N. L., Merker, B., & Brown, S., (Eds.). (2000). The origins of music. Cambridge, MA: MIT Press. Wiggins, J. (1990). Composition in the classroom: A tool for teaching. Reston: MENC. Zachopouloua, E., Tsapakidoub, A., & Derric, V. (2004). The effects of a developmentally appropriate music

and movement program on motor performance. Early Childhood Research, 19, 631–642.