USF Libraries
USF Digital Collections

Upper elementary boys' participation during group singing activities in single-sex and coeducational classes

MISSING IMAGE

Material Information

Title:
Upper elementary boys' participation during group singing activities in single-sex and coeducational classes
Physical Description:
Book
Language:
English
Creator:
Bazzy, Zadda
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Mixed-gender
Same-sex
Single-gender
Music education
Fourth and fifth grade
Dissertations, Academic -- Music -- Masters -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: As boys in the upper elementary grades become increasingly influenced by peer pressure, many are less likely to participate in singing activities because singing is considered a "feminine" activity. The purpose of this research was to explore if there was an effect on upper elementary boys' level of participation during group singing activities when they attended music classes in a single-sex setting. This study employed a true experimental design and a mixed method. Boys (N = 186) were videotaped during their regular coeducational music classes on two occasions to establish baseline data. Then the students were randomly assigned to attend music classes in either a single-sex or coeducational group. Boys were videotaped again after seven music classes (approximately 9 weeks later). The videos were scored using the author-designed Singing Participation Measure, and the scores (N = 123) were analyzed using an analysis of variance (ANOVA). In addition, qualitative data were collected in the form of music teacher interviews and journal entries. The ANOVA showed no statistically significant differences between groups (single-sex or coeducational) or within groups (baseline scores versus post-treatment scores). In contrast, the qualitative data showed substantial differences in most of the boys' participation in single-sex classes. The teachers reported a sudden increase in the boys' singing participation and described numerous advantages of single-sex music education. Further research is needed. Implications for music educators suggest teachers could create single-sex singing opportunities, choose repertoire mindfully, and establish a "singing culture" at the school to increase boys' participation during singing activities. In addition, music educators are encouraged to know their students' strengths, weaknesses, interests, and needs, and to remember that "one size" does not "fit all" when it comes to what is best for developing young musicians.
Thesis:
Dissertation (PHD)--University of South Florida, 2010.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Zadda Bazzy.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains X pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
usfldc doi - E14-SFE0004603
usfldc handle - e14.4603
System ID:
SFS0027918:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0004603
035
(OCoLC)
040
FHM
c FHM
049
FHMM
090
XX9999 (Online)
1 100
Bazzy, Zadda.
0 245
Upper elementary boys' participation during group singing activities in single-sex and coeducational classes
h [electronic resource] /
by Zadda Bazzy.
260
[Tampa, Fla] :
b University of South Florida,
2010.
500
Title from PDF of title page.
Document formatted into pages; contains X pages.
502
Dissertation (PHD)--University of South Florida, 2010.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
3 520
ABSTRACT: As boys in the upper elementary grades become increasingly influenced by peer pressure, many are less likely to participate in singing activities because singing is considered a "feminine" activity. The purpose of this research was to explore if there was an effect on upper elementary boys' level of participation during group singing activities when they attended music classes in a single-sex setting. This study employed a true experimental design and a mixed method. Boys (N = 186) were videotaped during their regular coeducational music classes on two occasions to establish baseline data. Then the students were randomly assigned to attend music classes in either a single-sex or coeducational group. Boys were videotaped again after seven music classes (approximately 9 weeks later). The videos were scored using the author-designed Singing Participation Measure, and the scores (N = 123) were analyzed using an analysis of variance (ANOVA). In addition, qualitative data were collected in the form of music teacher interviews and journal entries. The ANOVA showed no statistically significant differences between groups (single-sex or coeducational) or within groups (baseline scores versus post-treatment scores). In contrast, the qualitative data showed substantial differences in most of the boys' participation in single-sex classes. The teachers reported a sudden increase in the boys' singing participation and described numerous advantages of single-sex music education. Further research is needed. Implications for music educators suggest teachers could create single-sex singing opportunities, choose repertoire mindfully, and establish a "singing culture" at the school to increase boys' participation during singing activities. In addition, music educators are encouraged to know their students' strengths, weaknesses, interests, and needs, and to remember that "one size" does not "fit all" when it comes to what is best for developing young musicians.
590
Advisor: David A. Williams, Ph.D.
653
Mixed-gender
Same-sex
Single-gender
Music education
Fourth and fifth grade
690
Dissertations, Academic
z USF
x Music
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.4603



PAGE 1

Upper Elementary Boys P articipation During Group Singing A ctivities in Single sex and Coeducational C lasses by Zadda M. Bazzy A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Center for Music Education Research School of Music College of The Arts University of South Florida Major Professor: David A. Williams, Ph.D. Robert F. Dedrick, Ph.D. C. Victor Fung, Ph.D. Janet L. S. Moore, Ed.D. Date of Approval July 8 2010 Keywords: Mixed gender, Same sex, S ingle gender Music E ducation Fourth and Fifth Grade Copyright 2010, Zadda M. Bazzy

PAGE 2

! Dedication I dedicate this dissertation to the boys and girls in single sex and coeducational classes at Oneco Elementary. You are truly the inspiration for this research.

PAGE 3

Acknowledgments To begin, I would like to thank the music teachers and administrators who participated in this study, including Helen Abernathy, Ka Cline, Doug Dupouy, Julie Hebert, Cynthia Heidel, Christy Isaacs, Marsha Perry Juday, Judy Kelley, Elizabeth Kimbrell, Jeff Lego, Helene Levin, Allison Rekow, Myra Russell, Barbara Siffermann, and Barbara Sullivan In one way or another, each of these individuals hel ped with the creation and testing of the Singing Participation measure and/or participated in the s tudy Furthermore, I extend my gratitude to the special area teachers, homeroom teachers, staff, parents, and students at the three schools in th e sample. In particular, I would like to thank the three music teachers who participated in this research. I never could have conducted this study without your dedication and support. Next I must thank Dr. Jeffrey Kromrey for assisting in the ge neralizability analysi s of the Singing Participation M easure, Kathy Rolsten for triangulating the qualitative data analysis and Barbara Siffermann for scoring 108 videos! Thank you! Special thanks go to the members of my dissertation committee: Dr. David Williams, Dr. Victor Fung, Dr. Janet Moore, and Dr. Robert Dedrick. You have inspired me to become the best researcher I can be. I am grateful for your guidance. Finally, I would like to thank my principal, Marian Summers, and my friends and family (both furry and otherwise) who have showered me with love and patience during this exciting endeavor.

PAGE 4

! ! ! Ta ble of Contents List of Tables ................................ ................................ ................................ ................... iii List of Figures ................................ ................................ ................................ ..................... v Abstract ................................ ................................ ................................ ................... vi Chapter 1 : Introduction ................................ ................................ ................................ ........ 1 Background ................................ ................................ ................................ .............. 1 Rationale ................................ ................................ ................................ .................. 4 Purpose ................................ ................................ ................................ ..................... 5 Hypotheses ................................ ................................ ................................ ............... 5 Educational Significance ................................ ................................ ......................... 6 Single sex Versus Single gender ................................ ................................ ............. 6 Definitions ................................ ................................ ................................ ............... 7 Delimitations and Limitations ................................ ................................ ................. 8 Chapter 2 : Literature Review ................................ ................................ ............................. 11 G ender Stereotypes and Singing ................................ ................................ ............ 11 Single Sex Education ................................ ................................ ............................. 14 Assessing Participation ................................ ................................ .......................... 18 Conclusion ................................ ................................ ................................ ............. 22 Chapter 3 : Method ................................ ................................ ................................ ............. 23 Design ................................ ................................ ................................ ................... 23 Sample ................................ ................................ ................................ ................... 24 Procedures: Quantitative ................................ ................................ ........................ 27 Procedures: Qualitative ................................ ................................ .......................... 29 Repertoire ................................ ................................ ................................ ............... 30 Measures ................................ ................................ ................................ ................ 31 Data Analysis: Quantitative ................................ ................................ ................... 36 Data Analysis: Qualitative ................................ ................................ ..................... 37 Ethical Concerns ................................ ................................ ................................ .... 38 Validity and Reliability ................................ ................................ .......................... 38 Summary ................................ ................................ ................................ ................ 40 Chapter 4: Results ................................ ................................ ................................ .............. 41 Quantitative Data ................................ ................................ ................................ ... 41 Overview ................................ ................................ ................................ .... 41 Reliability ................................ ................................ ................................ ... 42

PAGE 5

! "" Descriptive statistics ................................ ................................ .................. 44 Analysis of variance ................................ ................................ ................... 55 Assumptions of the ANOVA ................................ ................................ ..... 59 Summary : Quantitative data ................................ ................................ ...... 59 Qualitative Data ................................ ................................ ................................ ..... 61 Overview ................................ ................................ ................................ .... 61 Qualitative data by theme ................................ ................................ .......... 62 Participa tion in coeducational classes ................................ ............ 62 Participation in single sex classes ................................ .................. 64 Adv antages of single sex classes ................................ ................... 64 Time ................................ ................................ ............................... 68 Behavior problems ................................ ................................ ......... 70 Par ticular groupings of students ................................ .................... 72 The impact of th e research design on teachers .............................. 73 Othe r ................................ ................................ .............................. 75 Qualitative data by school ................................ ................................ .......... 80 School 1 ................................ ................................ ......................... 80 School 2 ................................ ................................ ......................... 82 School 3 ................................ ................................ ......................... 84 Overall ................................ ................................ ........................... 86 Summary : Qualitative data ................................ ................................ ........ 8 6 Conclusion ................................ ................................ ................................ ............. 8 8 Chapter 5: Discussion ................................ ................................ ................................ ........ 90 The Re search Hypotheses Revisited ................................ ................................ ...... 91 B eyond the Research Hypotheses ................................ ................................ .......... 9 4 Advantages and disadvan tages of single sex education ............................ 9 4 Song choice ................................ ................................ ................................ 96 The teachers influence ................................ ................................ .............. 9 7 Imp l ications for Music Educators ................................ ................................ .......... 98 Create sin gle sex singing opportunities ................................ ..................... 98 Remember that one size does not fit all ................................ ............ 100 Ch oose repertoire mindfully ................................ ................................ .... 102 Establish a s inging culture at the school ................................ .............. 103 Know your students ................................ ................................ ................. 103 Future Research ................................ ................................ ................................ ... 104 Conclusion ................................ ................................ ................................ ........... 107 References ................................ ................................ ................................ ................. 110 A ppendices ................................ ................................ ................................ ................. 119 Appendix A: Musi c Teacher Re flective Journals ................................ ................ 120 Appendix B: Se m i structured Interview Guide ................................ ................... 121 About th e Author ................................ ................................ ................................ ... End Page

PAGE 6

! """ List of Tables Table 1: School, Distri ct, and State Characteristics ................................ ........................... 25 Table 2: Sample by School and Grade ................................ ................................ ............... 26 Table 3: Inter item Correlations for the Singing Participation Measure ........................... 35 Table 4: Controlling for Potentially Confound ing Variables ................................ ............ 39 Table 5: Mean Participation Totals and I nterrater Reliability by Time ............................. 43 Table 6: Internal Consistency of the Singing Participation Measu re by Time .................. 44 Table 7 : Song Details by Song Num ber ................................ ................................ ............ 46 Tabl e 8: Songs by School and Time ................................ ................................ .................. 46 Table 9: Mean Participation Totals for Each Song by School and Time .......................... 47 Table 10: Descriptive Statistics for Total Scores by School and Group for Time 1 ................................ ................................ ................................ ........ 49 Table 11: Descriptive Statistics for Total Scores by School and Group for Time 2 ................................ ................................ ................................ ........ 50 Table 12: Descriptive Statistics for Total Scores by School and Group for Tim e 3 ................................ ................................ ................................ ........ 51 Table 13: Descriptive Statistics for Total Scores by School and Group for Time 4 ................................ ................................ ................................ ........ 52 Table 14: Correlations Between Total Scores at Each Poin t in Time ............................... 57 Table 15: Mixed ANOVA Resul ts by School, Time, and Group ( N = 123) ..................... 58 Table 16: Qualitative Data Theme 1: Participation in Coeducational Classes .................. 63 Table 17: Qualitative Data Theme 2: Partic ipation in Single sex Classes ........................ 65 Table 18: Qualitative Data Theme 3: Adv a ntages of Single sex Classes .......................... 67

PAGE 7

! "# Table 19: Q ualitative Data Theme 4: Time ................................ ................................ ....... 69 Table 20: Qualitative Da ta Theme 5: Behavior Problems ................................ ................. 71 Table 21: Qualitative Data Theme 6: Par ticular Groupings of Students ........................... 73 Table 22: Qualitative Data Theme 7: The Impact of the Research Design on Teachers ................................ ................................ ................................ ..... 74 Table 23: Qu alitative Data Theme 8: Other ................................ ................................ ....... 76

PAGE 8

! # List of Figures Figure 1: The Gender Hierarchy of School Subjects and Activities ................................ .... 3 Figure 2: Singing Participation Measure ................................ ................................ ........... 34

PAGE 9

! #" Upper Elementary Boys Participation During Group Singing Activities in Single sex and Coeducational Classes Zadda M. Bazzy Abstract As boys in the upper elementary grades become increasingly influenced by peer pressure, many are less likely to participate in singing activities because singing is considered a feminine activity. The purpose of this research was to explore if there was an effect on upper elementary boys level of participation during group singing activities when they attend ed music classes in a single sex setting. This study employed a true experimental design and a mixed method Boys ( N = 186) were videotaped d uring their regular coeducational music cla sses on two occasions to establish baseline data. Then the students were randomly assigned to attend music class es in either a single sex or coeducational group. Boys were videotaped again after seven music classe s (approximately 9 weeks later). The videos were scored using the author designed Singing Participation Measure, and the scores ( N = 123) were analyzed using an analysis of variance (ANOVA). In addition, qualitative data were collected in the form of music teacher interviews and journal entries. The ANOVA showed no statistically significant differences between groups (single sex or coeducational) or within groups (baseline scores versus post treatment scores). In contrast, t he qualitative data showed subst antial differences in most of the boys participation in single sex classes. The teachers reported a sudden increase in the

PAGE 10

! #"" boys singing participation and described numerous advantages of single sex music education. Further research is needed. Implication s for musi c educators suggest teachers could create si ngle sex singing opportunities, c hoose repertoire mindfully, and establish a singing culture at the school to increase boys participation during singing activities In addition, music educators are encouraged to know their students strengths, weaknesses, interests, and needs, and to remember that one size does not fit all when it comes to what is best for developing young musicians.

PAGE 11

! Chapter 1: Introduction As boys in the upper elementary grades become more concerned with social appearances, many are less likely to participate in singing activities in the music classroom. In the United States, singing is typically considered to be a feminine activity (Adler & Harrison, 2004) As such, boys who hope to impr ess girls with their masculinity are often unwilling to sing when there are girls in the room ( Demorest, 2000) However, when boys have the opportuni ty to attend music classes exclusively with other boys, the social dynamics change. This paper describes a research study on boys participation during group singing activities in single sex and coeducational settings. Background Many high school choral directors report more female students than male students enroll in choir. T his has not always been the case. In colonial times, men participated in singing activities more than women did (Gates, 1989). Koza (1993) reviewed articles published between 1914 1924 in the Music Supervisors Journal and found the missing males problem was evident in the early 1900s. During this period, writers of gender related articles were concerned with the decline of boys participating in vocal music and many writers proposed that the effeminate nature of vocal music was a possible cause for this decline. Similar articles were published in the Music Educators Journal in the 1930s and 1940s. Damon (1936) wrote an article in which she described an eighth grade class as the boys who did not sing. Winslow (1946) claimed Probably nothing perplexes the

PAGE 12

! # secondary school teacher more than the vocal education of boys and young men. Many boys enter high school with negative and sometimes hostile attitudes tow ard vocal musicVocal music suggests femininity to boys (p. 58). Gates (1989) reported that in 1932 there was a balanced ratio of female to male singers in the United States, but by 1989 the ratio of female singers to males singers was 5:2 At the seconda ry level, when music is typically an elective class, female enrollment in choir greatly outnumbers male enrollment. Research on student attitudes towards music shows that students in the upper elementary grades become increasingly influenced by gender st ereotypes (Adler, 2002; Pogonowski, 1985; Sherban, 1995; Svengalis, 1978). Adler and Harrison (2004) created The Gender Hierarchy of School Subjects and Activities (see Figure 1). In this hierarchy, music is considered more feminine than reading, writing, or math, and vocal music is considered more feminine than instrumental music. These gender stereotypes may inhibit boys willingness to sing, particularly in front of girls who they might want to impress with their masculinity. In fact, boys in coeducation al elementary music classes participate in singing activities significantly less than girls do (Eccles, Wigfield, Harold, & Blumenfeld, 1993; Haladyna & Thomas, 1979; Mizener, 1993; Moore, 1987; Sherban, 1995). The decline in male participation in singing activities may be due to an emphasis the United States society has placed on categorizing choral singing as a feminine activity. In recent years, males who participate in choirs are often stereotyped as effeminate and possibly homosexual. Once young boys b ecome conscious of gender stereotypes in our culture, they may choose to avoid stereotypically feminine activities such as singing.

PAGE 13

! I N T E R V A R S I T Y S P O R T S I n t r a m u r a l S p o r t s I N F O R M A T I O N T E C H N O L O G Y B u s i n e s s S t u d i e s P H Y S I C S C H E M I S T R Y B i o l o g y M A T H E M A T I C S T R A D E T E C H N O L O G Y ( E l e c t r i c a l W o o d A u t o ) E n g l i s h / L a n g u a g e A r t s G e o g r a p h y H i s t o r y V ISUAL ARTS D R A M A MUS IC: M a r c h i n g B a n d C o m p e t i t i v e B a n d N o n c o m p e t i t i v e B a n d / I n s t r u m e n t a l C l a s s G u i t a r C l a s s S h o w C h o i r O r c h e s t r a J a z z C h o i r C o m p e t i t i v e C h o i r N o n c o m p e t i t i v e C h o i r / V o c a l C l a s s D a n c e Figure 1. The Gender Hierarchy of School Su bjects and Activities created by Adler and Harrison (2004), reproduced with permission by the Canadian Music Educators Association / LAssociation Canadienne des Musiciens ducateurs. M A S C U L I N E F E M I N I N E

PAGE 14

! Rationale One area of concern in the field of music education is the limited participation of males in singing activities once music becomes an elective course (Demorest, 2000; Koza, 1993). Music education at the elementary level may have an important impact on whet her or not boys will elect to take music courses at the secondary level. R esearch su ggests many boys become less interested in singing around the age of 8 years (Mizener, 1993; Moore, 1987; Svengalis, 1978). It is possible a change in approach to music edu cation at the elementary level might maintain boys interest in singing. One strategy for increasing male participation in s inging activities is to allow boys to sing in a single sex environment where the social dynamics are different and bo ys may feel it is safe and acceptable for them to sing. Currently, there is a growing trend of single sex education in the United States. When the National Association for Single Sex Public Education (NASSPE) was founded in March of 2002, only 11 public schools in the United States offered single sex classroo ms. As of April 2010, at least 540 public schools in United States offer ed single sex environments, according to the NASSPE ( #$%$ ). This increase in the number of single sex classrooms aligns with a growing body of research exploring potential advantages for both boys and girls who attend classes with students of the same sex St udies show that there is less gender stereotyping in single sex environments (Colley, Comber, & Hargreaves, 1994), and students are more likely to participate in gender atypical activities in single sex environment s (Swain & Harvey, 2002; Younger & Warring ton, 2006). However, at the present time, there is little research investigating the effects of single sex education as it relates specifically to the field of music education.

PAGE 15

! & Purpose I first became interested in this area of research when I began teach ing both single sex and coeducational classes at the elementary level. During informal observations, I noticed a striking difference in boys participation in singing activities when there were no girls in the room. This study formally investigated to what degree, if any, this phenomenon occurred in other elementary schools with single sex music classes. The purp ose of this study wa s to determin e if there wa s an effect on upper elementary boys level of participation dur ing group singing activities when the y attend ed music classes in a single sex setting Hypothese s This research test ed two hypotheses: '$ Boys in single sex classes will participate during group singing activities m ore than boys in coeducational classes. To test this hypothesis, I compare d participation score s between boys in the treatment groups ( single sex classe s) and the control groups (coeducational classes) within the same school site. In addition, I analyzed qualitative data from the music teacher interviews and journal entries to det ermine what differences, if any, the music teachers reported between the boys participation in the two different groups. ($ Boys in single sex classes will participate during group singi ng activities more at the end of the study than they did when they we re in coeducational classes at the beginning of the study and they will participate more than the boys who remained in coeducational classes.

PAGE 16

! ) To test this hypothesis, I compare d baseline scores for individual boys to the scores those same boys receive d during t he single sex treatment period. In addition, I analyzed qualitative data from the music teacher interviews and journal entries to determine what differences, if any, the music teachers reported in the participation of individual boys at the two dif ferent points in time (during the baseline data collection and at the end of the study). Educational Significance This study explored if the sex composition of a class significantly affected boys willingness to sing. Studies of this kind are important be cause if research shows boys are more apt to participate in singing activities when they are in a single sex environment, then music educators should consider creating opportunities for boys to sing exclusively with other boys. Such single sex opportunitie s may be especially important at the upper elementary grades when many children become strongly influenced by their social environment. If boys develop a love for singing in a socially acceptable environment at the elementary level, they may choose to enro ll in choir in middle school and high school when music courses become elective. This study was one of the first to specifically address the impact of single sex education on music education at the elementary level. The findings might suggest several impli cations for the field. Single sex V ersus Single gender The terms single sex and single gender are synonymous. It is common to find both terms used in the literature to describe classes of all boys or all girls. For this study, I

PAGE 17

! decided to use only on e term consistently A ccording to The Concise Oxford English Dictionary (2008) : Although the words gender and sex both have the sense of the state of being male or female, they are typically used in different ways: sex tends to refer to biological differences, while gender tends to refer to cultural or social ones. (p. 592) I chose to use the term single sex rather than single gender because I am discussing biological differences (whether the students in the class are boys or girls) rather than the social construct of gender (whether the students in the class are masculine or feminine). T he music teachers in this study often refer red to their all boy classes as single gender rather than single sex The staff at the three school sites decided to use the term single gender when referring to the classes because they did not want to use the word sex around the students. Since the teach ers were used to referring to the classes as single gender, they often used this term when answering the interview questions. The only time I use the term single gender subsequently in this paper is when I am quoting the music teachers directly. Defini tions The following definitions apply to this research : Single sex class: A class comprised of either all boys or all girls. Coeducational class: A class comprised of both boys and girls. For this study, coeducational classes were included only if a t least one third of the class wa s boys and a t least one third of the class wa s girls. While a class with 19 boys

PAGE 18

! + and 1 girl would technically be coeducational, such an imbalance in sex composition would not serve the purpose of this study. Participation: Fo r the quantitative portion of the study, participation was defined as, The level of engagement during singing activities based on the movement of the students mouth, the focus of the students eyes, and the position of the students body including postur e, body language, and/or movements. This was measured using the Singing Participation Measure (described in Chapter 3). For the qualitative portion of the study, participation was defined as, The degree of singing during group singing activities. This w as reported via teacher observation. Note that for both definitions of participation the data captured the degree to which students were willing to sing; these data did not refer to the quality of the singing. Group singing activities: Any activi ties in which the entire class wa s supposed to sing si multaneously. The study only included w hole class activities; data were not collected during solo singing or small group singing. Expert elementary music teacher: A teacher who has a masters degree in Music Ed ucation and has taught music at the elementary level for more than 10 years. Delimitations and Limitations Th is study included three delimitations: 1. The sample was limited to schools within one school district. 2. Data w ere only collected on boys, not on girl s. The literature stated that girls show significantly more interest in singing activities than boys at all grade

PAGE 19

! levels ( Adler, 2002; Haladyna & Thomas, 1979 ; Mizener, 1993; Pogonowski, 1985; Sherban, 1995 ) This study focus ed solely on boys participation in singing activities. 3. Data were only collected on students in fourth and fif th grades. I chose to use students in the upper elementar y grades because these students we re too young to elect whether or not to enroll in mus ic class and y et old enough to be significantly influenced by gender stereotypes. Although I collected data from students in two different grades, I neither expected nor intended to explore differences by grade level. I used two grade levels to increase the sample size, but I considered these students to be in the same category: students in the upper elementary grades. For all of the statistical analyses, I kept the scores of the fourth and fifth grade students combined. Limitations of the study include d : 1. The individual s who rate d the bo ys degree of participation knew if the boys we re in the treatment group or the control group because the video show ed whether the class wa s single sex or coeducational. The only way to avoid this would have be en to videotape each boy individually to ensure that the students next to him we re not visibl e in the recording. It would have been impossible to videotape each individual boy simultaneously unless there was one camera for each boy in the classroom. 2. Students may have fake d singing. Data were collected during group singing activities and therefore it was difficult to discern exactly which individual voices were heard. Unfortunately, individually mic ing each student would

PAGE 20

! '! have cause d unwanted student reactions detrimental to the study I acknowledge that a few students may have feign ed their participation in singing activit ies, however this limitation could not be avoided.

PAGE 21

! '' Chapter 2: Literature Review This chapter includes a review of the literature related to the research question: What are the effects of single sex and coeducational classes on upper elementary boys participation during group singing activities ? Although the literature in the field of music education does not offer many s tudies specific to single sex music classes, there are a number of related research topics including gender stereotypes and singing, single sex education in subjects other than music, and the assessment of part ici pation in music classes. I review ed rese a r c h in each of these area s according to the relevant nature of these topics to this study Gender Stereotypes and Singing Several studies indicate that boys may not sing because they consider singing to be a feminine activity. Viggiano (1941) administered a questionnaire to 200 students to assess their attitudes towards music. The questionnaire produc ed both qualitative and quantitative data. Viggiano reported that the qualitative data showed It is not music so much as singing to which the boys seem to object (p. 62). Comments on the quest ionnaires indicated some boys did not like to s ing because they were afraid their peers would mak e fun of them Children are influenced by their peers, and boys who do not feel it is socially acceptab le to sing are less likely to participate in singing activities. Sherban (1995) conducted a qualitative study in which she observed and audio recorded music classes of first grade and fift h grade students on eight occas ions (four

PAGE 22

! '( times a week for two weeks ) In addition, she interviewed eight students (2 boys and 2 girls from each grade), their music teacher and the school administrator The data she collected showed that boys viewed singing as feminine (p. 106) She explained that boys were more willing to sing popular music because it was viewed as more heterosexual than choral music. She attributed the lack of boys participation in singing activities to peer pressure for boys to conform to gender stereotypic al activities Sherban documented nine examples where boys attempted t o deliberately sabotage singing.. .The reason for this sabotage was an attempt to establish their identity as boys (p. 91). In addition, t he differences in the attitudes of the boys and the girls were more pronounced in the fifth grade class. Sherban reported, Grade one boys were involved in singing. They sang with emotion and feeling. The grade five boys were divided, some participated and some did everything they could to subvert sing ing as a class activity (p. 103). These data indicate sex stereotypes of singing are often developed at the elementary level and boys in the upper elementary grades may be less likely to sing than boys in the primary grades Another study on sex stereot ypes and singing provided similar results. Hall (2005) explored the gender stereotypes of 38 five year old boys. He gave the boys a pictorial survey where the boys were asked to identify sex neutral stick figures as male or female. Each stick figure was en gaged in an occupation such as computer operator, teacher, singer, etc. The majority of the boys (25 out of 34) labeled the singing figure as female. In subsequent interviews, one boy explained Singing is something girls do most times (p. 13). Hall conc luded that the missing male trend may begin in early childhood.

PAGE 23

! '. Adler (2002 ) contended b oys who participate in activities which are gender role incongruent may result in being labeled as homosexual and the primary gender related cause to the missing males problem ishomophobia (p. 33). In his qualitative study, Adler o bserved his seventh and eighth grade choral music classes and conducted both group and individual interviews with 16 boys and 2 girls over a two month period The data showed that the students could be grouped in five categories: 1. Jocks of Singing These boys sang in the choir. They were socially popular and extroverted, and they asserted themselves as leaders. They were skilled in sports, labeled as masculine, and had heterosexual rep utations. 2. Sensitive Boys These boys sang in the choir. They were introverted and had an almost androgynous masculinity. They lacked sports prowess. 3. Neutral Boys These boys sang in the choir. They were a year younger than the other boys and had not identified with a particular group yet. 4. Non Singers These boys did not sing in the choir. They appeared shy and nervous. They expressed concern that they might rece ive negative feedback from their peers if they sang. 5. Bad Asse s These boys did not sing in the choir. They were a powerful masculine force at school. Many of these boys harassed the boys in the choir. Adler explained that the groups displayed differing de grees of masculinity. In addition, some of the groups were influenced by gender stereotypes to a greater degree. In the interviews, several of the students labeled singing as a girl thing . Adler concluded that the primary gender related caus e to the miss ing males problem wa s homophobia He

PAGE 24

! '" predicted boys would be more likely to sing in a single sex choir and suggested music educators should consider single sex education. Single s ex Education The biological differences between boys and girls go far beyond the obvious differences in sexual organs. Male and female brain tissues are intrinsically different and the various areas of the brain develop in a different order, time, and rate for boys and girls (Sax, 2005, p. 93). In addition to biological and phys iological differences, boys and girls are subjected to different social roles beginning as early as birth when boys are swaddled in blue fabric and girls are d raped in pink. Banduras Social Cognitive Theory emphasizes that many gender stereotypes are tied to social roles, not biological differences (1989). Whether one considers gender differences in the context of biology or sociology, or both, it is clear that boys and girls are inde ed different. Sax, the founder an d Executive D irector of the National Association for Single Sex Public Education, believe s many boys and girls would benefit from attend ing classes in single sex settings because te achers can better address sex difference s in learning styles attentiveness, and social behaviors. Some researchers have hypothesized that single sex classes would provide opportunities for girls to participate more in science and technology activities and for boys to participate more in the vis ual and perf orming arts and language arts. There are many conflicting results in single sex education research At least three studies show that boys and girls do significantly better both academically and socially when surround ed by students of the same sex First, Swain and Harvey (2002) review e d the literature on the gender gap in te chnology and concluded that females who attended

PAGE 25

! '& single sex computer classes participated more and were more likely to consider a career in technology Second, Younger and Warrington (2006) conducted three case studies with six coeducational secondary schools with single sex classes in the United Kingdom. They used a mixed method design; they collected quantitative data through student questionnaires and test sco res and they collected qualitative data via observations. Their data showed that the students in the single sex classes made substantial gains in achievement. In addition, many of the students felt more comfortable in a single sex setting. One boy stated, You say things you wouldnt say in front of girls because you feel more confident, you know they wont laugh at you and you dont lose face (p. 602). Finally, Gibb, Fergusson, and Horwood (2008) conducted a study in New Zealand with 940 students who eith er attended a single sex or coeducational high school. They measured educational achievement by the amount of education completed at the secondary and post secondary levels. Their data showed statistically significant differences in the educational achie vement of boys and girls at the coeducational schools, but no significant differences between the boys and girls at the single sex schools. The researchers concluded by stating, single sex schooling mitigates gender differences in educational achievement (p. 315). Together, the three studies described here support the practice of single sex schooling to promote academic achievement and socially acceptable environments for pursuing gender atypical interests. Other studies show inconclusive results regardin g the effects of single sex education Hoffman, Badgett, and Parker (2008) conducted a two year study on the effectiveness of single sex instruction on students in an urban, at risk high school. They analyzed course grades, standardized test scores, classr oom observations, surveys,

PAGE 26

! ') teacher interviews, and a focus group discussion. When comparing the single sex classes to the coeducational classes, the researchers found mixed results. In some cases the coeducational classes performed better while in other ca ses the single sex classes performed better. Hubbard and Datnow (2005) conducted an ethnographic study of low income and minority students who attended experimental single sex academies in California. These researcher s found that although these schools wer e successful, the students success was not necessarily related directly to single sex schooling. The researchers found that the schools successes were due more to the schools organizational characteristics (such as smaller class sizes) student teacher relationships, and resources. Salomone (2006) reviewed the literature on the research conundrum surrounding single sex education. She stated that research findings that supported single sex scho oling we re in con clusive and stressed the need for further research in the area of social science and single sex schooling. In the field of elementary music education, there is very little re search on single sex music classes. Most of the research on sex differences in elementary music includes data from coeduca tional classes. For example, Moore (1987) conducted time interval sampling on the attentiveness of 977 students in grade 1 through 5 and reported that boys were significantly less attentive than girls during singing activities and boys accounted for 66% of the off task behavior observed. Several researchers have studied sex differences in the musical creativity of elementary students. Kiehn (2003), Schmidt and Sinor (1986), and Wolfe and Linden (1991) all reported that boys outperformed girls in creative tasks, however Wilson and Wales (1995) showed that girls outperformed boys. As for differences in musical achievement, three studies i ndicate d no significant sex

PAGE 27

! '* differences in achievement (Cooper, 1995; Hallam, 2004; Hedden, 1982) while two studies indicate d that females we re more successful at vocally matching pitch than males (Howle, 1992; Welch, Sergeant, & White, 1997). Many studies showed that girls enjoy music class more than boys (Haladyna & Thomas, 1979; Mizener, 1993; Pogonowski, 1985; Sherban, 1995). In regards to self efficacy, research by Eccles, Wigfield, Harold, and Blumenfeld (1993) indicated that girls had more positive c ompetence beliefs and value beliefs of instrumental music than boys. While all of these studies provide d information on sex differences in music education, each of these studies focused on sex differences within a coeducational classroom. In fact, at the presen t time I have only found one single sex research study in the field of music education and this research was conducted at the secondary level Carp (2004) surveyed 101 choral directors who taught single sex choirs at the secondary level Eighty eigh t percent of these teachers observed differences in behavior in single sex choir students, and the majority of the directors identified student behavior as better in single sex environments. While there is extensive research on sex differences in music education, currently there is very little research specific to single sex music education. Studies on single sex education in fields outside of music help to shed some light on possible advantag es and disadvantages to single sex schooling, however there is a gap in the literature related to single sex music education, particularly at the elementary level. The current study helps to fill this gap in the literature.

PAGE 28

! '+ Assessing Participation When I first became interested in studying single sex music education, I wanted to study boys willingness to sing. I conducted a self report survey with the boys in my single sex classes, and I noticed the results of the survey were incongruent with the behavio rs I saw from the boys during singing activities; the boys reported that they did not like to sing in single sex classes, and yet the boys in my single sex classes outperformed the boys in my coeducational classes. It was then I decided that self report da ta might not be reliable and I would have to find a different way to measure the boys willingness to sing. I began reading the literature to see how researchers defined and measured participation during singing activities. I believed (and still do belie ve) that active participation is an outward, measurable behavior indicative of learning and achieving learning outcomes. My hope was to find an established measure for assessing singing participation for the current study. Interestingly, although many music educators issue participation grades as well as skills grades on quarterly report cards, the existing body of literature offers few research studies that directly address the assessment of student participation. There is a battery of tests for musical achievement (Music, 1988), as well as extensive research on assessing musical skills (Chiodo et al., 1998; Cooper, 1995; Hallam, 2004; Hedden, 1982; Keenan Takagi, 2000; Moore, 1994; Taylor, 1969; Wel ch, 1997; in addition to many others); however, research on assessing participation in the music classroom is limited. Many researchers have addressed the topic of participation by assessing on task and off task behavior. Madsen and Geringer (2000) explained, When a task requires

PAGE 29

! ', an individual to attend to music in some manner, this attention occupies a high degree of participant involvement (p. 106). Researchers have documented on task and off task behaviors by various methods. Many researchers employed time interval sam pling in which they observed students at regularly timed intervals and documented the number of students displaying on task or off task behaviors (Duke, 1987; Forsythe, 1977; Madsen & Madsen, 1974; Madsen & Yarbrough, 1980; Sims, 1986; Wright & Van Der Mar s, 2004; Ya rbrough & Price, 1981). The majority of these studies used a dichotomous system to differentiate between on task and off task behaviors, however Yarbrough and Price (1981) defined the following four classifications: 1. On Task Active when student s are supposed to be performing, they must look at either the music or teacher 2. On Task Passive when students are not supposed to be performing, they must be quiet and look at the music, teacher, or ensemble members who are performing 3. On Task Other stud ents must follow instructions given by the teacher 4. Off Task students are observably not on task (p. 211) Categorizing behavior in this manner offered more information on student behavior than a dichotomous system. In addition to time interval sampling, some researchers have conducted ongoing sampling ( Duke, Buckner, Cavitt, & Colprit 1997 ; Jellison, 2002). During ongoing sampling, the re searcher continuously documents overt behaviors rather than

PAGE 30

! (! documenting behaviors only at timed intervals. Jellison (2 002) employed the following four categories in her study: 1. On Task 2. Off Task Looking 3. Off Task Active 4. Off Task Other (p. 347) Similar to the study by Yarbrough and Price (1981), Jellisons multiple categories provided more detailed data than the dichotomous systems used by many other researchers. In contrast, Jellison categorized three different types of off task behavior, whereas Yarbrough and Price categorize d three different types of on task behavior. The variance in categories suggest s a different focus for these researchers; Jellison examined different types of off task behaviors, and Yarbrough and Price examined different types of on task behaviors. Another method for assessing participation included documenting the number of minutes and/or seconds a s tudent spent on task or off task rather than counting the number of times a certain behavior was observed. Researchers have focused on total time on task (Brand, 2003; Standley, 1992), rather than measuring consecutive seconds spent on tas k. In addition, c omputer software has been used to create a timeline of on task / off task behaviors to illustrate participation trends during music lessons (Duke et al., 1997). In contrast to measuring the frequency of on task or off task behaviors or the amount of time spent on task or off task, some researchers have used rating scales to

PAGE 31

! (' document behaviors. Johnson, Darrow, and Eason (2008) used an overall rating scale of 1 100 to measure student teacher interactions in the music classroom. Gregait, Johnsen, and Nielsen (1997) suggested physical education teachers use a participation checklist with a rating scale of 0 3 in seven different categories. Two of the observable behaviors under the Active Listening category on this participation checklist included sitting up and eyes on speaker (p. 29). Rating scales suggest students may participate in activities to various degrees; rather than a simple no or yes marking for on task participation, these instruments indicate greater or lesser degrees of student particip ation. In summary, the majority of research on assessment in the field of music education has focused on musical skill rather than participation. There have been some studies that measured on task or off task behavior, although in most of these cases the assessment of participation was not the primary focus of the research. Many studies were concerned with teacher effectiveness or rehearsal activity effectiveness, and these studies used measures of on task or off t ask behavior as data to indicate effective ness of the teacher or the rehearsal activity (Forsythe, 1977; Johnson, Darrow, & Eason, 2008; Sims, 1986; Yarbrough & Price, 1981). One study measured student participation as it related to the proximity of typical students to classmates with disa bilities (Jellison, 2002). After careful review of the literature, I decided I should develop and test a measure of singing participation for use in the current study. While the literature showed that researchers have assessed participation in many different ways, the literature did not offer a consensus on how participation should be measured. Therefore, prior to beginning my research on boys participation in single sex and coeducational classes, I conducted

PAGE 32

! (( two studies to develop and test a va lid and reliable measure, titled the Singing Participation Measure. These studies are explained in detail in Chapter 3. Conclusion The literature shows many students view singing to be a fem inine activity and this belief may be formed in early childhood. As such elementary music educators should consider ways to challenge traditional gender stereotypes. One way to create a socially acceptable environment for boys to sing may be through single sex education. The research findings on the success of single sex education programs are mixed, and currently there is very little research available on single sex music education. T he current study help s fill the gap in the literature r egarding single sex music education and whether it could play a role in encouraging boys to part icipate in singing activities.

PAGE 33

! (. Chapter 3: Method This chapter provides details on the study inclu ding the design, sample, quantitative and qualitative procedures repertoire, measures that were used including a review of the research studies conduct ed to develop the measu res, quantitative and qualitative data analyse s, ethical concerns, and considerations related to the vali dity and reliability of the data Design This research employ ed a true experiment al design and a mixed method including the collection of both quantitative and qualitative data The treatment include d the manipulation of one independent variable: the composition of the class based on biological sex I used a random digits table to assign students to either the control group or the treatment group. The control group included students who attend ed music class in a coeducational setting; the treatment gro up included students who attended music class in a single sex setting. The dependent variable wa s the level of participation during group singing a ctivities. Participation was quantified using the Singing Participation Measure In additio n to the quantita tive data, qualitative data were collected including music teacher journal entries and interviews. The qualitative data provide d method triangulation by documenting the music teachers perspectives on the results of the study. The

PAGE 34

! (" conscientious effort to document another perspective is an example of authent icity in this research. Sample The sample include d fourth and fifth grade students who attend ed one of three elementary schools within the same school district in the sout heastern United States The sample wa s limited to fourth and fifth grade stud ents because research indicates boys become less interested in singing in the upper elementary grades (Mizener, 1993; Moore, 1987; Svengalis, 1978) The selecte d schools were a convenience sample chosen on the basis of my professional rapport with the music teachers at these school sites, and the willingness of the administrators, teachers, and staff to restructure the sex compositions of the fourth and fifth g rade classes for t his study Table 1 outlines the defining characteristics of each of the three schools. School 1 and School 2 we re similar in their demo graphics. Both of these schools were Title I schools that had not made adequ ate yearly progress for 6 year s These two sc hools serve d a large number of economically disadvantaged students and ha d a large minority population. School 3 served students of a higher socioeconomic status and had a small minority population. No fourth or fifth grade boy was excluded because of race native language, cognitive or behavioral exceptionalities, or physical, mental, or health status. All 235 fourth and fifth grade boy s at the thre e schools sites were invited to participate in the researc h I was able to obtain assent and parental consent for 198 of these boys. The sample was further reduced due to excessive absences (some students were not present at school for any of the four days when the music teacher videotaped the classes) and

PAGE 35

! (& student s transferring to new scho ols prior to the data collection. The final sample size was N = 186 Table 1 School, District, and State Characteristics ________________________________________________________________ ________ School 1 School 2 School 3 District State _______________________________________________________________________ Title I Yes Yes No School Grade for 2008 2009 a C B A Total N umber of students 380 642 587 > 42, 000 2,671,513 M ales 53 % 51 % 55 % 51 % 51 % F emales 47 % 49 % 45 % 49 % 49 % Economically Disadvantaged 89 % 93% 20 % 47% 50% White 10 % 10% 85 % 56% 45% Black 4 2% 3 0% 1 % 15% 23% Hispanic 46 % 53% 9 % 24% 25% Asian <1% <1% 3 % 2% 3% American Indian <1% <1% <1% < 1% <1% Multiracial 2% 7% 3 % 4% 4% a The Florida Department of Education reports school performance by assigning grades of A, B, C, D, or F to schools based on the annual learning gains of students as measured by the Florida Comprehensive As sessment Test.

PAGE 36

! () I collected baseline data on the boys in their regular coeducational music classes, and then I randomly assigned students to attend music class in either a single sex or coeducational class for the treatment portion of the study. The sample included 21 classes during the baseline data collection and 15 classes (6 single sex and 9 coeducational) during the post treatment data collection. The 6 single sex classes included 1 fourth grade class and 1 fifth grade class from each school site. The 9 coeducational classes that were videotaped at the end of the study served as the control groups. School 1 had 1 coeducational class at each grade level. School 2 had 1 coeducational class in fourth grade and 2 coeducational classes in fifth grade. School 3 had 2 coeducational classes at each grade level. Table 2 outlines the specific number of fourth and fift h grade boys at each school Table 2 Sample by School and Grade _______________________________________________________________ ________ School 1 School 2 School 3 Total Sample ________________________________________________________________________ 4 t h Grade Boys 25 24 41 90 5 t h Grade Boys 21 22 53 96 Total 46 46 9 4 186 _________________________________________________________________ _______ I collected qualitative data from the music teachers at the three school sites. The educational ba ckground and experience of the teachers varied somewhat. The teacher at School 1 he ld a bachelors degree in Music Education and had six years experience teaching music at the elem entary level. She also had three years experience teaching

PAGE 37

! (* band, chorus, guitar, keyboarding, and music theory at the secondary level. The teacher at School 2 also held a bachelors degree in Music E ducation. She had five years experience teaching music at the elementary level. The teacher at School 3 had the most training and experience. She held a bachelors degree in Music Education and a masters degree in E ducational Leadership. She had 10 years experience teaching music at the elementary level. She was certified by the National Board for Professional Teaching Standards and held Level I and Level II certifications in the Orff approach to music education as i ssued by the American Orff Schulwerk Association. Procedures : Quantitative Students began the 2009 2010 school year in coeducational c lasses. Coeducational classes we re sta ndard at these three school s, so the school year began as it normally did The research did not begin until the third quarter of the school year, so the students were used to their daily schedules, the music teachers, and the procedures of the music classes by the time the research began. Students were videotaped during group s inging activities on two occasions prior to the treatment period to establish baseline data. The students were videotaped i n their regular music classroom in an unobtrusive manner. The music te achers at each school place d digital video cameras in their c la ssrooms at least two weeks before they record ed t he students so the children had time to become accustomed to the video cameras. Two of the teachers used two video cameras, each covering half of the classroom. One teacher had to use three video cameras to record every area in her classroom. All of the video cameras were digital. The makes and models of the cameras included Canon ZR50MC, Canon ZR500, Canon ZR800, Canon ZR930, and Panasonic 3CCD. The record ing light

PAGE 38

! (+ on each camera was covered by a piece of ma sking tape so studen ts could not observe when the cameras wer e recording. Also, the students were videotaped only when they we re performing songs they had practiced previously. This wa s important because students who have limited reading abilities may not participate when the class is learning a new song if they cannot read the lyrics efficiently. Limiting the recording to only songs with which the students were familiar helped control for the potentially confounding variable of reading ability. Af ter the baseline video footage wa s collected, all of the students in fourth and fifth gra de, both boys and girls, were randomly assigned to new classes for music, art, physical education, and other special area classes. Studen ts in the treatment groups were assi gned to single sex classes. Students in the control group s were assigned to coeducational classes. Although the students in the control group s were already enrolled in coeducational class es for the first part of the sc hool year, these students wer e reassigned to a different configuration of students for the coeducational classes that served as the control groups for this study. In an attempt to control for extraneou s variables, all students wer e reassigned to new classes of stude nts regardless of w hether they we re in the treatment group or the control group. The treatment took place for eight music classes The length of time for the treatment period varied by school because each school followed a different schedule. For example, at School 3 it to ok 9 weeks for each student to have music class 8 times. At School 1 it took 11 weeks for each student to have music class 8 times. At School 2 it took 11.5 weeks for each student to have music class 8 times. This amount of time wa s appropriate for several reasons. First, this was not a longitudinal study; all of the data had

PAGE 39

! (, to be collected within one school year. Second, due to unusual activities that occur red at the beginning and end of a school year, it was important to collect data during the middle of the year. The baseline data took approximately 5 weeks to collect (2 weeks of having the cameras in place and 2 3 weeks of videotaping the students on 2 occasions) and the treatment data took between 9 11.5 weeks to collect. The total amount of time requ ired for the data collection was approximately 11 14.5 weeks, or almost half of the sc hool year. The treatment could not extend beyond eight music classes because the data collection had to begin during the middle of the school year. The music teachers lef t the video cameras set up in their classrooms for the entire study. Identical to the baseline data collection, at the end of the treatmen t period the music teachers videotape d the classes on two diffe rent occasions when students were singing songs with wh ich they were familiar Since the total sample was 186 students and each student was supposed to be videotaped at four points in time there could have been as many as 744 scores. Due to absences, attrition, and technical difficulties, I was unable to vide otape every boy four times. Some boys were videotaped at one only point in time, while others were videotaped at all four points in time. The total number of scores was 538. Procedures: Qua litative The three mu sic teachers complete d semi structured r eflective journal entries to document their observations of the boys participation during group si nging activities (see Appendix A ). These data we re extremely valuable for two reasons. First, the music teachers ha d an ongoing relationship with the student s in their classrooms. They might have noticed change s in behavior or attitude that we re too subtle to be captured in the

PAGE 40

! .! quantitative data collection a nd analysis. Second, the use of journal entries ensured that data we re recorded when the information wa s fresh in the teachers minds. These journal entries we re opportunities for the music teachers to document specific behaviors they observed that day. Unfortunately, although the teachers were asked to complete journal entries on a weekly (if not daily ) bas is, none of the teachers managed to co mplete a journal entry every week The teachers documented the students participation when they observed events that were striking or poignant. Sometimes teachers completed multiple journal entries within the same wee k; at other times, the teachers did not complete a journal entry for several weeks. At the end of the treatment period I intervie w ed each music teacher to inquire what changes in participation, if any, the teacher observed during the course of the study. I use d a semi structured interview guide (see Appendix B ). The interviews ranged in length from approximately 30 100 minutes, depending on the teacher. These interviews were recorded with a Canon ZR500 digital video camera transcr ibed, and subjected to member checks to verify the accuracy of the transcriptions prior to the data analysis. Repertoire The mus ic teachers at each school used different curriculums. Although the three teachers taught differen t lessons each music teacher taught the same lessons to both the single sex and coeducational classes of fourth and fifth grade students within the school. In other words, the students at each school experience d different lessons from the students at the other schools, but the student s within the same school received the same lessons regardless of grade level or group (either single sex or coed ucational ).

PAGE 41

! .' The teachers were told that they could have their students sing different repertoire during the baseline data collection and the tre atment period. This wa s necessary because musical preference is greatest at a moderate level of arousal (Lehmann, 2007). If students perform ed the same repertoire for an extended period of time, their participation may have decrease d simply because they we re tired of the song I expected the choice of literature would affect student participation. For example, some boys may prefer singing an African American spiritual to singing a lullaby. However, I controlled for changes in participation due to the style of the music by comparing the participation of the control group to that of the treatment group from the sam e school. For example, if the boys we re unwillin g to sing a lullaby, then I expect ed the participation scores to be lower for both the treatment gr oup a nd the control group when they we re videotaped singing a lullaby. Participation s cores for both groups were expected to vary across the four points in t ime based on the repertoire the students sang ; however, in my analysis I looked for significant dif ferences between the treatment group s and control group s By allowing the teachers to use their stand ard curriculums, I maintained high level s of ecological validity in the study. If I had forced the music teachers to teach certain songs at certain points in time, the results could only be generalized to those particular songs. Measures Since no preexisting mea s ure for quantifying the degree of student participation during group singing activities could be found I conducted two phases of research to develop and test a singing participation scoring instrument.

PAGE 42

! .( In April of 2009, I conducted the first phase of research to explore how elementary music teachers assess individual participation during group singing activities. A focus group of six expe r t elementary music teachers viewed video footage of students singing and discussed the best practices for assessing singing participation. I collected qualitative data thro ugh observations and a video recording of the discussion. Overall, the music teacher s felt that participation should be assessed using a rating scale rather than a dichotomous system. The focus group considered what types of items should be on a rubric used to assess participation, and the focus group discussion was videotaped, transcribe d, and subjected to member checks. Then I inductively coded the qualitative data and another expert elementary music teacher coded the data to measure the agreement reliability of the codes ( r = .84). I used the data to create an instrument for quantifying individual student participation during group singing activities: the S inging Participation Measure This instrument measures the degree of student participation during group singing activities based on five categories: Eyes, Mouth, Voice, Body, and Overa ll. The focus group participants reviewed and revised the proposed scoring instrument using the Delphi technique to achieve group consensus of the most valid and reliable way to assess singing participation. In the summer of 2009, I conducted the second p hase of this research to test the reliability and determine the concurrent validity of the Singing Participation Measure. Ten music teachers viewed videos of elementary music classes during group singing activities and scored individual students using the Singing Participation Measure. For this phase, I had to alter the Singing Participation Measure. It was impossible to include the dimension Voice in the measure when scoring the videos since the raters could not

PAGE 43

! .. identify individual voices in a recording where the entire class was singing. The Singing Participation Measure was modified to include only four of the five original dimensions: Eyes, Mouth, Body, and Overall (see Figure 2). Ten music teachers rated the participation of 17 students in 34 videos (two videos of each student performing on two different days). The raters were divided into two groups of five, and the two groups viewed and rated the videos in a different order to account for a potenti al order effect. Two weeks later the same groups watched the same videos in a different order. I used these data to determine the generali zability and inter item correlation coefficients for the scoring instrument. The generalizability analysis estimated a n interrater reliability of .73, an intrarater reliability of .89, and a stability of student performance of .80. The inter item correlation coefficients ranged from .88 to .99 (see Table 3). The content validi ty of this scoring instrument was strong becau se, according to the expert music teachers who developed the instrument, the data th at we re collected by this rating scale were the same as the data that we re collected in a typical music class setting. In addition, the use of the Delphi Technique to achie ve a consensus among the experts strengthened the construct validity of the measure. Furthermore, I compared the participation scores each student received to the participation grades issued to those same students by their music teacher at the time of the video recording and calculated a concurrent validity coefficient of .90. Based on the strength of the Singing Participation Measures validity and reliability, I decided to use this instrument as the quantitative measure of student participation during gro up singing activities in the current study. Once again, I did not include the dimension Voice when using this measure since I rated individual students

PAGE 44

! ." Singing Participation Measure Student #: _____ _____________________________ Total Score: ___________ _______ A. Eyes : The students eyes are focused, looking at the director and the music (if applicable). Never or ~25% of the time ~50% of the time ~75% of the time Always or Almost Never Almost Always 1 2 3 4 5 B. Mouth : The students mouth is open an appropriate amount and moving with the rhythm of the words. Never or ~25% of the time ~50% of the time ~75% of the time Always or Almost Never Almost Always 1 2 3 4 5 C. Body : The students posture, body language, and/or movements indicate active participation. Never or ~25% of the time ~50% of the time ~75% of the time Always or Almost Never Almost A lway s 1 2 3 4 5 D. Overall : The student is actively and appropriately engaged in the singing activity. This may be demonstrated through facial expression or overall affect. Never or ~25% of the time ~50% of the time ~75% of the time Always or Almost Never Almost Always 1 2 3 4 5 Figure 2. Scoring instrument for qu antifying an individuals level of participation during group singing activities.

PAGE 45

! .& Table 3 Inter item Correlations for the Singing Participation Measure _______________________________________________________ ______ ___ Dimension Eyes Mouth Body Overall Total Eyes ___ .88 .93 .95 .96 Mouth ___ .89 .96 .96 Body ___ .98 .97 Overall ___ .99 Total ___ ________________________________________________________________ Note N = 17. during group singing activities by scoring videos and it was impossible to identify individual voices when listening to the recordings. The Singing Participation Measure is designed to quantify an individuals willingness to sing as demonstrated by active participation during singing activi ties. Since the emphasis is on the students singing a child must open his mouth in order to receive scores higher than 1 in each of the four dimensions. For example, it is possible to observe a student whose eyes are focused, looking at the director and the music (if applicable) even though he is not singing. Likewise, a student may choose to participate in body percussion or other movements that go with the song without ever opening his mouth. Since the Singing Participation Measure is specific to sing ing children who do not open their mouths must receive the lowest possible score in all four dimensions regardless of the focus of their eyes or the action of their bodies. Essentially, the movement of the students mouth is considered a prerequisite to r eceiving any score higher than 1 in the dimensions of Ey es, Mouth, and Overall. For the current study, an expert music teacher scored videos in addition to myself. This second rater was trained on the correct way of scoring students using the Singing Parti cipation Measure, including

PAGE 46

! .) how to score students who were focused with their eyes and/or participated with their bodies but demonstrated an unwillingness to open their mouths and sing. Data Analysis : Quantitative An expert music teacher and I viewed the videos and scored each students individual level of participation using the Singing Participation Measure Students who were not videotaped at least once during the baseline data collection and at least once post trea t ment because of absences or att rition, could not be included in the analysis of variance (ANOVA) Before the v ideos were rated, I train ed the e xpert music teacher in using the scoring instrument ; she and I practiced scoring footage from sam ple videos and discussed our scores and the best way to use the measure We did not rate any videos from the study until we we re confident in our abilities to use the scori ng instrument effectively T here were 538 video clips to score I sc ore d all of the videos, and the expert music teache r scored a random sample of 2 0% of the videos or 108 videos We watched the vide os independently and we scored the random sample of videos in the same order. Then I calculate d the interrater reliability for the scores of the 108 videos. After determinin g th e interrater reliability, I analyze d the 538 scores I used d es criptive statistics to report the mean, range, st andard deviations, skewness, kurtosis standard errors of the means, and effect sizes In addition, I conducted a mixed design analysis of variance (ANOVA) to compare the between groups factor of class composition ( single sex or coeducational ) and the within groups factor of time (baseline scores and scores at the end of the treatment). The data were analyzed for each school site ; I did not combine the data across schools since the schools had different music teachers different curriculums, and different student characteristic s

PAGE 47

! .* Data Analysis: Qualitative The qualitative data include d the music teach ers reflective journal entries and the transcriptions of the interviews with the three music teachers. These data were used to tr iangulate the quantitative data Patton warned The credibility of qualitative methodshinges to a great extent on the skill, competence, and rigor of the person doing the fieldwork ( 2002, p. 14). I was highly qualified to con duct this study for many reasons. First, I hold a masters degree in Mu sic Education and am cu rrently pursuing my docto rate in Music Education In addition, I am trained in both qualitative and quantitative research methods and analysis and have experience in conducting both qualitative and quantitative research. Also I have taught elementary ge neral music for 12 years and have experience teaching both single sex and coeducatio nal classes. For the qualitative data analysis, I began by coding the data from the interview transcriptions and the music teacher journal entries I established preliminary codes based upon the interview guide and my research hypotheses. For example, I expected to code the data according to the observations of the single sex classes, coeducation classes, group differences, individual differences, a dvantages and disadvantages of single sex education, and possible reactivity caused by the research. As I reviewed the data, I established more codes inductively. This step was necessary to accurately classify the data. I judged the codes by assessing the internal homogeneity and external heterogeneity of the data within each code. I systematically analyzed the classification system by collapsing some categories and creating new codes until I was confident in the trustworthiness of the analysis. Then I aske d an individual trained and experienced in qualitative data analysis to code the data. I reviewed her codes and made minor revisions

PAGE 48

! .+ to my final analysis. This analyst triangulation improved the credibility of my data analysis. Ethical Concerns The names of the students who participate d in the study remain confidential. I analyzed the quantitative data by class, so there was no need to report scores for individuals. As for the qualitative data, I was careful to omit the names of individual students to prot ect their privacy and keep their information confidential. V alidity and Reliability Several aspects of the research design maintained a high degree of validity and reliability for this study. The mixed method design provide d both qualitative and quantitative data I reviewed these two data sources to see if the different forms of data corroborated This is an example of methods triangulation (Patton, 2002). I allowed the music teachers to teach their regular curriculu ms for t his study. This improved the ecological validity of the research. If I had told the music teachers exactly what to teach, the results would offer limited generalizability The music teachers member check ed the interview transcriptions and a graduate rese ar cher independently coded the qualitative data to provide analyst triangulation and ensure authenticity As for the quantitative data, another expert music teacher rated the vide os in addition to myself, and I calculated the interrater reliability Prior st udies in the development and testing of the Singing Participation Measure examined the content and construct validity of this measure through a focus group discussion and the Delphi technique These prior studies tested the reliability of the measure and t he scores were compared to the students participation grades on their report cards to determine the concurrent validity.

PAGE 49

! ., As I established my research design, I attempted to control for several potentially co nfounding variables (se e Table 4 ). The treatment groups at each school site studied the same repertoire and received the same instruction as the control groups at the same site. Also, since I expected student participation to vary from day to day, I had the teachers videotape the students at mu ltiple points in time. In addition, the music teachers set up video cameras and left them in place for at least two weeks prior to recording so the students had time to get accustomed to the cameras before they were videotaped. Also, a piece of tape was pl aced over the recording light on the cameras to keep students from knowing when the video cameras were actually recording. Finally, the students were only videotaped when they were singing songs with which they were familiar. This was Table 4 Controlling for Potentially Confounding Variables Variable Control Repertoire The music teacher taught the same repertoire to both the single sex and coeducational classes at the same school. Variance in individual Videos were taken at multiple points in time prior to student participation the treatment period and at multiple points in time during the treatment period Reactivity The music teachers left the video cameras in the classrooms for at least two weeks before colle cting video footage. The music teachers covered the recording lights on the video cameras so students did not know when the cameras we re recording. Student reading/language Video recordings were limited to singing lessons in skills which th e students wer e highly familiar with the lyrics of the songs.

PAGE 50

! "! important because students with limited reading abilities often have a difficult time participating when presented with a new song if they cannot read the lyrics efficiently. All of these c hoices related to the research design increased the validity of this research. Summary This study was a true experiment. The sample included 186 boys in fourth and fifth grades at three different school sites. Baseline data were collected prior to beginning the study, then the students were randomly assigned to attend music class in either a single sex class (the treatment group) or a coeducational class (the control group). I employed a mixed method design for this research by collecting both quantitative data in the form of videotapes that were scored and qualitative data in the form of music teacher journal entries and interviews. The videos were scored using the Singing Participation Measure, a measure that I designed and tested previously in preparation for this research. I attempted to control for several potentially confo unding variables and strived to maintain a high level of validity and reliability in the study.

PAGE 51

! "' Chapter 4: Results I used a mixed method design to triangulate the data on boys participation during group singing activities in single sex and coeducational classes. This chapter includes both the quantitative and qualitative results of the study. The quantitative data ar e presented first, followed by the qualitative data and the conclusion. As stated previously the research hypotheses were: 1. Boys in single sex classes will participate during group singing activities more than boys in coeducational classes. 2. Boys in single sex classes will participate during group singing activities more at the end of the study than they did when they were in coeducational classes at the beginning of the study and they will participate more than the boys who remained in coeducational class es. I report data directly related to the two hypotheses, as well as other data related to the topic of singing participation. I used PASW Statistics 17.0 to conduct all of the statistical analyses reported in this chapter. Quantitative Data Overview The total sample size for this study was N = 186. I intended to videotape each boy at four points in time two times prior to being assigned to new groups as baseline data and two times at the end of the study in the new groups. This would have provided 7 44 scores (186 boys scored at four points in time each). Due to

PAGE 52

! "( abs ences, attrition, and technical difficulties, I was unable to videotape every boy four times. Some boys were videotaped at only one point in time, while others were videotaped at all four points in time. The total number of scores was 538. It is important to note that School 3 encountered technical difficulties when videotaping the students at the second point in time. In th e data t hat follow, there is a decrease in the sample size for School 3 at Time 2. For all of the statistical analyses, I kept the scores of the fourth and fifth grade students combined. I neither expected nor intended to explore differ ences by grade level. I use d two grades in my sample to increase the sample size. I kept the data for the two grades aggregate d during the statistical analyses. Reliability. I reviewed and scored all of the videos ( N = 538), and a second rater independently reviewed and scored 20% of the videos ( n = 108). This second rater was highly qualified to score the videos. She had 13 yea rs of teaching experience, held an M.A. degree in Music Education, a nd wa s Na tional Board Certified in Early and Middle Childhood Music. She was 1 of 10 teachers who used the Singing Participation Measure to score videos in an earlier phase of this research. In addition, she was trained on using the measure prior to scoring the videos. I used a random digits table to randomly select the 108 v ideos that the second rater scored. I viewed and score d these videos in the same order that she did however I scored them on a different day, independently from her Then I compared our scores to determine the interrater reliability. Table 5 shows the mean participation totals and standard deviations for the scores issued b y myself and the second rater at each of the four points in time. In addition, this

PAGE 53

! "# Table 5 Mean Participation Totals and Interrater Reliability by Time Time 1 Time 2 Time 3 Time 4 Mean SD Mean SD Mean SD Mean SD Rater 1 13.52 4.03 11.48 5.83 11.70 5.43 14.46 4.11 Rater 2 14.61 3.98 11.14 5.47 9.20 5.09 12.75 3.85 Pearson Corr. .83 .88 .88 .83 n 33 21 30 24 Note. The participation totals had a possible range of 4 20 points with higher numbers indicating a greater degree of singing participation.

PAGE 54

! "" table shows the correlations between my scores and the second raters scores. The Pearson correlation coefficients ra nged from .83 to .88, indicating a strong reliability between the scores given by each rater. Next I compared the scores for each dimension of the Singing Participation Measure (Eyes, Mouth, Body, and Overall) for the 538 videos that I scored Table 6 sh ows the Cronbachs a lpha for the videos taken at each point in time. The ranged from .87 to .94 This shows that the internal consistency of the measure was high. Table 6 Internal Consistenc y of the Singing Participation Measure by Time Time 1 Time 2 Time 3 Time 4 Cronbachs a lpha .87 .94 .92 .9 2 n 143 90 152 153 T he strength of the correlations between the two raters and the high level of internal consistency of the Singing Participation Measure suggest that the scores from this measure are highly reliable Descriptive s tatistics. During the course of this study, the music teachers at the three school sites followed their own curriculums. This means that the three teachers were not teaching the same songs or lessons. One might expect to see dramatic differences in the boys singing participation based upon the repertoire. For example, boys may participate more when singing an African American spiritual than they would when singing a lullaby. For this reason, it is important to look at the differences between the mean scores based on song.

PAGE 55

! "# That being said, one cannot look at the differences between the mean scores of different songs without considering the impact of school, time, and group. Different schools have different students and teachers, and both of these variables may affect singing participation. Time may also have an effect on the scores. If students sing the same song for a long period of time, their pa rticipation may change as t hey become more comfortable with a song or grow bored with a song. Finally, singing participation may be affected by wh ether the boys a re in the single sex group or the coeducational group. At both Time 1 and Time 2, all of the boys were in their regular c oeducational classes. However, the videos taken at Time 3 and Time 4 captured student participation in their n ew groups, either single sex or coed ucational As one reviews the descriptive statistics, it is important to keep in mind the total scores may be affected by the song, school, time, and group, as well as interaction effects among these four variables. Students at the three school sites were videotaped on four occasions. This means it was possible to have as many as 12 songs in this study. In actuali ty, the three music teachers collectively used only nine songs Table 7 identifies the songs by title and composer/lyricist. In addition, this table shows the identification number assigned to each song. Table 8 shows which songs the students sang by both school and time. School 1 used Song 1 (The Star Spangled Banner) at both Time 1 and Time 3. School 3 also used Song 1, but only at Time 2. School 3 used Song 8 (Im Gonna Sing) at both Time 1 and Time 3. All of the other songs were used by one school, at one point in time only.

PAGE 56

! "$ Table 7 Song Details by Song Number Song # Title Composer/Lyricist 1 The Star Spangled Banner Smith and Key 2 La Buena Vida Emerson 3 Free at Last African American Spiritual 4 The Grand Old Duke of York Traditonal 5 Will You Be My V A L E N T I N E? Unknown 6 Parallel or Perpendicular? Jacobs and Bedley 7 The Lion Sleeps Tonight Weiss, Peretti, and Creatore 8 Im Gonna Sing African American Spiritual 9 Sing, Sing, Sing Prima Table 8 Songs by School and Time School Time 1 Time 2 Time 3 Time 4 1 Song 1 Song 2 Song 1 Song 3 2 Song 4 Song 5 Song 6 Song 7 3 Song 8 Song 1 Song 8 Song 9 Table 9 shows the mean totals scores for participation by song, school, and time. Differences between the single sex and coeducational groups do not appear in the table,

PAGE 57

! "% however group differences are shown in Tables 10 13. As shown in Table 9, the song wit h the lowest mean participation was La Buena Vida (Song 2), and the song with the highest mean participation was Sing, Sing, Sing (Song 9). Table 9 Mean Participation Totals for Each Song by School and Time Song School Time n Mean SD 1 1 1 39 10.49 3.80 1 1 3 39 10.56 4.36 1 3 2 15 10.40 2.97 2 1 2 38 9.29 4.49 3 1 4 38 10.87 4.17 4 2 1 36 12.97 4.6 3 5 2 2 37 14.43 4.51 6 2 3 35 11.74 5.34 7 2 4 43 12.51 4.58 8 3 1 68 12.32 3.08 8 3 3 78 14.12 2.95 9 3 4 72 14.61 3.11 Note. The participation totals had a possible r ange of 4 20 points with higher numbers indicating a greater degree of singing participation. Song 1 and Song 8 were used at multiple times. Although The Star Spangled Banner (Song 1) was performed at three different points in time and at two different school sites, the means for this song were very similar at 10.40, 10.49, and 10.56 The

PAGE 58

! "& mean scores for Im Gonna Sing (Song 8) increased fr om 12.32 to 14.12 from Time 1 to Time 3. This may be due to a treatment effect since Time 3 features data from the end of the treatment period when students we re in single sex and coeducational classes. In addition, this increase in scores may be due to an increased familiarity of the song over time. The standard deviations were the lowest for School 3 at 2.95, 2.97, 3.08, and 3.11 indicating that there was less variance in the students singing participation at this school. The standard deviations were t h e highest for School 2 at 4.51, 4.58, 4.62 and 5.3 4 indicating that there was more variance in the students singing participation at this school. Overall, Table 9 indicates that the mean totals for singing participation varied by song, school, and point in time The only song that had similar means regardless of school and time was The Star Spangled Banner (Song 1). Tables 10 13 present the descriptive statistics for the total scores by school and group at each point in time. The first two tables show the baseline data that were collected at Time 1 and 2, before the students were randomly assigned to new groups. Ideally, the group means would be similar and the effect sizes would be close to 0 in Tables 10 and 11 because this would show similarity between the baseli ne singing participation of the boys who woul d later b e assigned to the single sex and coeducational classes. According to Tables 10 and 11, the means were the most similar for School 3 at Time 1 where the boys who would later be in the single sex group had a mean of 12.39 and the boys who would later be in the coeducational group had a mean of 12.29. The means we re the most different for this same school, School 3, at Time 2 where the boys

PAGE 59

! "# Table 10 Descriptive Statistics for Total Scores by School and Group for Time 1 Range Group n Mean Min Max SD Skew ness Kurt osis SE of Mean Cohens d School 1 Single sex 25 10.32 4 16 3.25 0.53 0.16 0.65 0.13 a Coed ucational 14 10.79 4 19 4.74 0.16 0.70 1.27 School 2 Single sex 18 13.22 4 20 4.35 0.25 0.42 1.02 0.11 a Coed ucational 18 12.72 4 19 5.00 0.70 0.54 1.18 School 3 Single sex 26 12.39 4 20 3.69 0.33 0.42 0.72 0.03 a Coed ucational 42 12.29 8 19 2.69 0.64 0.19 0.42 Note These scores were determined under coeducational conditions. The students were not scored in their new single sex and coeducational groups until Time 3. a Effect size wa s computed using this formula: d = Mean S i n g l e s e x Mean C o e d u c a t i o n a l Pooled SD

PAGE 60

! $% Table 11 Descriptive Statistics for Total Scores by School and Group for Time 2 Range Group n Mean Min Max SD Skew ness Kurt osis SE of Mean Cohens d School 1 Single sex 2 6 8.50 4 1 8 4.38 0. 80 0.36 0.86 0.58 a Coed ucational 1 2 11.00 5 19 4.43 0.7 6 0. 31 1.28 School 2 Single sex 1 6 15.25 8 20 3.77 0.48 0. 54 0.94 0.33 a Coed ucational 21 13.81 4 20 5. 01 0.58 0.6 4 1.09 School 3 Single sex 6 b 8.66 8 9 0.52 0. 97 1.88 0.21 1.1 6 a Coed ucational 9 b 11.56 9 1 8 3.40 1.18 0.06 1.13 Note These scores were determined under coeducational conditions. The students were not scored in their new single sex and coeducational groups until Time 3. a Effect size was computed using this formula: d = b The sample size for School 3 at Time 2 was reduced due to technical difficulties with the video equipment. Mean S i n g l e s e x Mean C o e d u c a t i o n a l Pooled SD

PAGE 61

! $& Table 12 Descriptive Statistics for Total Scores by School and Group for Time 3 Range Group n Mean Min Max SD Skew ness Kurt osis SE of Mean Cohens d School 1 Single sex 2 4 10.13 4 1 6 4.40 0. 34 1.67 0. 90 0.27 a Coed ucational 1 5 11. 27 4 1 8 4.37 0.49 0. 62 1.13 School 2 Single sex 1 8 13.44 4 20 4.91 0.4 9 0. 80 1.16 0. 7 1 a Coed ucational 17 9.94 4 19 5. 31 0. 27 1.50 1. 29 School 3 Single sex 32 14.44 5 1 9 2.82 1.31 3.45 0. 50 0.19 a Coed ucational 46 13.89 7 20 3. 05 0.20 0.31 0.45 a Effect size was computed using this formula: d = Mean S i n g l e s e x Mean C o e d u c a t i o n a l Pooled SD

PAGE 62

! $' Table 13 Descriptive Statistics for Total Scores by School and Group for Time 4 Range Group n Mean Min Max SD Skew ness Kurt osis SE of Mean Cohens d School 1 Single sex 2 2 10. 96 4 1 9 4.38 0.15 0.86 0.9 3 0.05 a Coed ucational 1 6 10.75 4 18 3.99 0. 18 0. 44 1. 00 School 2 Single sex 23 13.17 4 20 4. 92 0.18 1.21 1.03 0.32 a Coed ucational 20 11.75 5 19 4.14 0. 11 0.69 0.93 School 3 Single sex 25 15.16 10 20 2. 75 0.09 0.97 0.5 5 0.27 a Coed ucational 4 7 14.32 4 20 3.27 0. 39 0. 71 0.4 8 a Effect size was computed using this formula: d = Mean S i n g l e s e x Mean C o e d u c a t i o n a l Pooled SD

PAGE 63

! "# who would later be in the sing le sex group had a mean of 8.67 and the boys who would later be in the coed ucational group had a mean of 11.56 This variance between the two groups may be due to the small sample size ( n = 15) for School 3 at Time 2. Many of the boys were not videotaped at this school during Time 2 because of technical difficulties during filming. Cohen (1988) established benchmarks for effect sizes from 0.20 indicating a small effect, 0.50 indicating a medium effect, and 0.80 indicating a large effect. The effect sizes at Time 1 were quite small for all three schools at d = 0.13, d = 0.11, and d = 0.03, respectively (see Table 10). At Time 2, the effect sizes were larger for each school at d = 0.58 d = 0.33 and d = 1.16, respectively (see Table 11 ). The very large effect size for School 3 at Time 2 is most likely due to an i ncrease in sam pling error because of the reduced sample size for this school at Time 2. It is unclear why the effect sizes for School 1 and School 2 are larger at Time 2. This may be due to differences in the boys level of participation for different songs. Ideally, th e effect size s for Time 1 and Time 2 would be very small, as this would indicate similarities between the boys who would later be divided into the treatment groups (single sex classes) and control groups (coeducational classes). The effect sizes for Schoo l 1 were negative for both Time 1 and Time 2. This means that the boys who would later be assigned to single sex classes participated less during the baseline data collection than the boys who would later be assigned to coeducational classes. The o pposite was true for School 2; the data for this school showed positive effect sizes for both Time 1 and Time 2. In this case, the boys who would later be assigned to sing le sex classes participated more during the baseline data

PAGE 64

! "$ collection than the boys who would later be assigned to coeducational classes. Since the effect sizes show small to moderate differences between the groups prior to the treatment phase of the study, it is important to keep in mind that there were some group differences in participation prio r to the reorganization of the students into their new single sex or coeducational classes. The highest possible score on the Singing Participation Measure is 20 and the lowest possible score is 4. The range, standard deviation, skewness kurtosis, and sta ndard error of the mean for the data reported i n Tables 1 0 and 11 we re all within acceptable limits. Tables 12 and 13 show the data collected at Time 3 and Time 4 the seventh and eighth time s students attended music class after the y were assigned to their new groups. If my hypotheses were supported one would expect to see a difference between the group means at each school and moderate to large effect sizes for Time 3 and Time 4 The means for the boys in the single sex groups were typically higher than the means for the boys in the c oed ucational groups, although not substantially At School 1, Time 3, the mean for the single sex group was actually lower than the mean for the coed ucational group. The effect sizes in Tables 12 and 13 encompass a wide range of effects from d = 0.27 to d = 0.71. The effect sizes were the highest for School 2, inc luding a moderately large effect size at Time 3 ( d = 0.71). The effect sizes were smaller for School 3 at d = 0.19 and d = 0.27 School 1 showed a small negative effect at Time 3 ( d = 0.27) and a very small positive effect at Time 4 ( d = 0.05) This means the boys in the coeducational classes participated more than the boys in the single sex classes for School

PAGE 65

! "" 1 at Time 3 but the boys in the single sex classes participated more than the boys in the coeducational classes for School 1 at Time 4 All of the other effect sizes in Tables 12 and 13 show a positive effect. This means that most of the time the boys in the single sex classes parti cipated more than the boys in the coeducational classes. The size of the effect varied across schools. The range, standard deviation, skewness, and standard error of the mean reported in Tables 12 and 13 were all within acceptable lim its. The kurtosis for School 3 at Time 3 was quite high at 3.45 The kurtosis may be leptokurtic due to a slightly high skewness for this group. According to Glass and Hopkins (1996), Highly skewed distributions tend to be leptokurtic because they have more scores that are far from the mean than does the normal distribution (p. 92). Collectively, Tab les 10 13 show that School 1 had the lowest means for singing participation in every group at every point in time except at Time 3 when the means of the coed ucational classes at Sc hool 1 were higher than the means of the coed ucational classes at School 2 (see Table 12) Th is means that with only on e exception, the singing participation of the boys at School 1 was noticeably lower than the singing participation of the boys at Schools 2 and 3. Analysis of variance. Before conducting the analysis of variance ( ANOVA ) I had to determine which of the 538 scores to use. My original intent was to take an average of the scores at Time 1 and Time 2 and an average of the scores at Time 3 and Time 4, and then conduct the ANOVA by comparing the two averages. For students who were missing data at one point in time I intended to simply use the score at that one point in time.

PAGE 66

! "% After reviewing the descriptive statistics, I had to reassess my plan. There was a lot of variance in the data by song and I was afraid that averaging some score s (where available) but not oth ers (due to missing data) would skew the results For example, each school used two different songs at Time 1 and Time 2 and there was a fair amount of variance between the means of the different songs. If I did the ANOVA using the average between Time 1 and Time 2 for the students who were filmed both days but I did not use the average for some students because they were only there for one of those two points in time, I might have skewed the data. To check whether this theory was correct, I calculated a c orrelation matrix so I could compare the scores at the different point s in time. If Time 1 and Time 2 were highly correlated, and if Time 3 and Time 4 were highly correlated, then I could have used the average s of the two baseline scores and post treatment scores for the students who were filmed on those occasions and used one baseline score and one post treatment score for the students who were not present for each filming. Table 14 shows the correlation matrix. The Pearson correlation coefficient for Time 1 and Time 2 was r = .48. The Pearson correlation coefficient for Time 3 and Time 4 was r = .62. I did not feel these correlations were strong enough to assume that student s would have performed similarly on the days they were absent. Therefore, I decided to conduct the ANOVA using only one point in time before the students were reassigned to new groups and only one point in time after students had been in their new groups f or 7 music classes.

PAGE 67

! "& Table 14 Correlations Between Total Scores at Each Point in Time Time 1 Time 2 Time 3 Time 4 Time 1 Pearson Corr. ___ .48 .40 .42 n 70 120 123 Time 2 Pearson Corr. ___ .19 .37 n 71 76 Time 3 Pearson Corr. ___ .62 n 127 Time 4 Pearson Corr. ___ n Next I had to decide which points in time to use. For the baseline data, it was clear that I should use Time 1 rather than Time 2 because many of the boys were not filmed at School 3 at Time 2 due to technical difficulties. The sample size for Time 1 was n = 143 while the sample size for Time 2 was only n = 90 As for Time 3 and Time 4, the sample s izes differed by only one child; t he sample size for Time 3 was n = 152 and the sample size for Time 4 was n = 153. Ultimately I decided to use Time 4 because this was the last point in time so it captured the entire treatment period Students who were filmed at Time 4 were i n the new groups for 8 music classes The final statistical analysis I conducted was the ANOVA. For these calculations, the variable Time compared the scores at Time 1 and Time 4. There were 123 students who were present at both Time 1 and Time 4, so the sample size for the ANOVA was

PAGE 68

! "' N = 123. I us ed a mixed ANOVA so I could look at differences between subjects (differences between boys in the single sex and coed ucational classes at each school ) and within subjects (differences betw een the baseline scores and the post treatment scores of t he boys in the single sex and coeducational classes at each school ). Table 15 shows the results by Time, School, and Group. The between subjects ANOVA showed statistical significance for the variable School ( p < .001). The analysis did not show statistical significance for the variable Table 15 Mixed ANOVA Results by School, Time, and Group (N = 123) Source df MS F p Between Subjects School 2 187.9 6 8.98 <.001 Group 1 7.19 0 .34 .56 School Group 2 7.65 0 .3 7 .69 Error 117 20.92 Within Subjects Time 1 39.13 4.49 .0 4 Time School 2 74.00 8.49 <.001 Time Group 1 0 .86 0.10 .75 Time School Group 2 1.3 4 0 .15 .86 Error 117 8.71

PAGE 69

! "( Group ( p = .56) or the interaction of the variables School and Group ( p = .69). These data suggest there were significant differences in the singing participation of the boys at different schools, but there were no statistically significant differences in the singing participation of the boys in the single sex and coeducati onal classes. The within subjects ANOVA showed statistical significance for the variable Time ( p = .04 ) and for the interaction of the variables Time and School ( p < .001). The analysis did not show statistical significance for the interaction of Time and Group ( p = .75) or the interaction of Time, School, and Group ( p = .86) These data suggest that there were significant differences in the singing participation of the boys at different schools and at different points in time, but there were no statistical ly significant differences in the baseline and post treatment scores of the boys in the single sex and coeducational classes. Assumption s of the ANOV A The data met the assumptions of the ANOVA regarding independence of observations and normality of the population distributions, however the data did not meet the assumption of the homogeneity of variance. Boxs M test of the equality of covariance matrices showed statistically significant differences at p = .01. In addition, Leven e s test of the equality of error variances showed statistically significant differences between the error variances at Time 4 at p = .04 Fortunately, ANOVA is fairly robust to violations of the homogeneity of variance (Stevens, 2007) In light of this robustness, I chose to con clude my statistical analyses with the ANOVA. Summary : Quantitative data. The descriptive statistics indicated differences in the mean totals for singing participation according to the song, school, and point in time. The only song that had similar means r egardless of school and time was The Star

PAGE 70

! %) Spangled Banner (Song 1). In addition, the descriptive statistics show ed that the boys at School 1 consistently participated less than the boys at Schools 2 and 3. There was only one point in time when one of the groups at School 1 did not have a lower score than the groups at the other two schools. The between subjects ANOVA showed statistically significant differences in the boys singing participation based upon the school they attended, and t he within subjects ANOVA show ed statistically significant differences in the boys singing participation based upon the school they at tend ed and the point in time. Both the between subjects ANOVA and the within subjects ANOVA showed no statistically significant differences between the single sex and coeducational classes. These d ata fail to support either of the hypotheses. It is importa nt to note that there was a great deal of variance in the data set. The data did not meet the assumption of homogeneity of variance. While the ANOVA is fairly robust against a violation of this nature, it is possible that a Type II error occurred. This may be due to the relatively small sample size. Harrison, Thompson, and Vannest (2009) warn that statistical significance testing is not efficient for estimating the probability that the claim is correct in the population (p. 744, emphasis in original). Sam pling error increases with smaller sample sizes, and this increases the possibility that the sample does not accurately reflect the population from which the sample was drawn. A larger sample would increase power and reduce the possibility of a Type II err or. It is also possible that there were truly no group differences in the boys singing participation as measured by the Singing Participation Measure.

PAGE 71

! %* According to Wilkinson and the Task Force on Statistical Inference (1991), researchers should always pr ovide some effect size estimate when reporting a p value (p. 599). Even though the ANOVA showed no statistically significant differences between the single sex and coeducational groups, I calculated the effect sizes using Cohens d The effect sizes at the end of the treatment period ranged from 0.27 to 0.71 with five of the six effect sizes showing a positive effect on boys singing participation in single sex classes. Qualitative Data Overview To ensure data triangulation, I collected two types of qualitative data journal entries written by the music teachers and interviews I conducted with each teacher. I transcribed the interviews and conducted member checks of the transcriptions to check for authenticity. For the qualit ative data analysis, I began by coding the data from the interview transcriptions and the music teacher journal entries. I established preliminary codes based upon the interview guide and my research hypotheses. For example, I expected to code the data according to observations of the single sex classes, coeducation classes, group differences, individual differences, advantages and disadvantages of single sex education, and possible reactivity caused by the research. As I reviewed the data, I establ ished more codes inductively. For example, under the broader category of Coeducation Classes, I developed two different codes: Limited Singing Participation and Participation in Non Singing Activities. This step was necessary to accur ately classify the d ata. Once all of the data were coded, I looked for themes in the data. I reorganized the previously coded data under broader thematic headings. I judged the theme s by assessing the internal

PAGE 72

! %+ homogeneity and external heterogeneity of the data within each the m e. I systematically analyzed the classification system by collapsing some categories and creating new theme s until I was confident in the trustworthiness of the analysis. Then I asked an individual trained and experienced in qualitative data analysis to i ndependently code the data and look for themes I reviewed her codes and themes, and we had a conference regarding our individual analyses. Then I mad e revisions to my final analysis. This analyst triangulation improved the credibility of my data analysis. Table s 16 23 present the results of the qu alitative data analysis. I organized the data in to the following eight theme s: Participation in Coeducational Classes, Participation in Single s ex Classes, Advantages of Single s ex Classes, Time, Behavior Problems Particular Groupings of Students, The Impact of the Research Design on Teachers, and Other. All but one of these categories includes multiple subcategories. I identified the teachers as Teacher 1, Teacher 2, and Teacher 3. These identification numbers c orrespond with the numbers for School 1, School 2, and School 3. A full comparison of the schools is provided in Chapter 3. As a brief r eminder, School 1 and School 2 wer e Titl e 1 schools, whereas School 3 was not. Qualitative data by theme Participation in coeducational classes. The teachers reported limited singing participation of the boys in the coeducational classes, both in the coeducational classes that were in place prior to this study and in the new coeducational classes that were formed when the students were randomly assigned to new groups (see Table 16) The teachers said that some boys just sat there, some boys moved their mouths but did not produce a sound, some boys sang very little, and some boys sang badly on purpose. One

PAGE 73

! %# Table 16 Qualitat ive Data Theme 1: Participation in Coeducational Classes Description Teacher Comments Limited singing participation was standard for the boys in coed ucational classes. Teacher 1: They just sit. If its a song they like, they might sing. Butthey just sit. I dont think [singing] would be their first choice. Teacher 1: I had to drag more out of them in that mixed group. It was almost like, Alright, Im going to take the blame because Im going to say that you have to sing, so whether you wanted to o r not, youre going to do it. So you get to save face. Teacher 1: What I get a lot is singing badly on purpose. Teacher 2: Normally the fifth grade boys act like they do not want to participate. If they do, theyll just barely open their lips a little bit and if I stand really close to them I cant hear any sound coming out. Teacher 2: The boys did not really sing that much and I really had to almost beg them to sing the song and we would have to do it again and again in little parts. Teacher 3: Th ere were some boys who just didnt sing as much as I thought they should. In the coed ucational classes, some boys would participate in non singing activities only Teacher 2: Normally, even with the coed, the boys would participate just as much when we would be playing instruments or doing speech pieces and stuff like that, but the singing is what was the most noticeable. Teacher 3: Some [boys] werent even moving their mouths. Although, if there was any kind of movement or body percussion involved, t hey would do that. But they were silent.. .Maybe 30% were participating but not fully, not doing the singing part. teacher said she had to drag more out of them and another said she had to almost beg them to sing. Two of the teachers mentioned that s ome boys would participate in non singing activities only. For example, the boys were willing to play instruments or perform the movements or body percussion for a song without actually singing. Overall, the singing participation for the boys in the coeduc ational classes was limited.

PAGE 74

! %$ Participation in single sex classes. The teachers observed more singing participation from the boys in the single sex classes (see Table 17) The teachers commented on singing participation for the single sex classes in general, as well as for individual boys. One teacher reported that it was easier to get her boys to sing in the single sex classes. Another teacher said more of her boys would offer to sing solos in the single sex classes. This same teacher explained that some of her boys who would normally sabotage singing activities by intentionally singing badly started to lip sync in the single sex classes. While lip syncing is not considere d full singing participation, this teacher felt this was a step in the right direction. All three of the teachers told stories about certain boys who participated in singing activities more in the single sex classes. Teacher 2 stated that a small change in participation for one student was actually a really big deal for him. The teachers provided many examples, both substantial and subtle, of increased singing participation of individuals and the groups overall in the single sex classes. Advantages of s ingle sex classes. The data indicated that there were many advantages to the single sex groupings (see Table 18). Teacher 1 liked the fact that she could focus on stuff just for them, such as how to read the non top line of music and how to transpose t he melody down an octave when it gets too high. She also felt that the antsy boys blended in more in the single sex classes because these classes tended to be more animated in general than the coeducational classes. Teacher 2 was surprised by the boys attitudes in the single sex classes. She noticed that they were more excited about coming to music and they admitted they liked

PAGE 75

! %" Table 17 Qualitative Data Theme 2: Participation in Single sex Classes Description Teacher Comments The teachers observed m ore singing participation in the single sex classes. Teacher 1: The boys in the all boys class volunteered to sing solos, whereas in the mixed group, [the boys] werent volunteering. Not as much, I dont think. Teacher 1: The middle of the road people tended to participate more on average in the single sex classes. Teacher 1: There are certain boys who sing bad just to be silly, to get attention. Some of that was reduced in the course of the study. I got more lip syncing. I see that as a step in the right direction. Its a compromise. Theres some effort in trying to match your mouth to the words. Teacher 2: [ referring to the boys in the single sex classes ] They would sing out without me having to convince them of it like I would have to with the co ed classes. Teacher 2: In the mixed classes I felt as if I had to persuade the students to participate and work really hard to drag it out of them. In the all boy classes I did not have to do that. I worked less to get them to participate. Teacher 3: I was like, You guys can do this! Youve been hiding behind the girls all this time and shying away, and here you are you are singing and you sound good! Teacher 3: I thought, Wow! [ B eing in a single sex class] really does make a difference. Teacher 3: Their participation was impressive. Teacher 3: My boys are singing really well. Teacher 3: I had my fifth grade boys today. They continue to impress me with their enthusiasm for singing. Teachers noticed differences in individual boys si nging participation in the single sex classes Teacher 1: I think [ names a boy ] really benefitted from that all boys class. He got into it. He got into singing. Teacher 1: [ referring to two boys in the single sex class ] They perked up a little bit, and I dont know if they wouldve done that if there were girls. Teacher 1: [ referring to one boy ] Hes not into [music ]. Hes shy. Its not his thing.. .He would still sing. Teacher 2: There was this boy. I dont really remember him ever even trying to si ng in class. And now with the all boyshe was actually moving his lips and participating some, which was a really big deal for him because he went from not participating at all to doing that.

PAGE 76

! %% Table 17 (continued) Qualitative Data Theme 2: Participation in Single sex Classes Description Teacher Comments Teachers noticed differences in individual boys singing participation in the single sex classes. (continued) Teacher 3: [ describing one boy ] Heskind of quiet, just doesnt say a whole lot, not a behavior problem, but just really quiet and shy. And he got into the all boys group and all of the sudden, hello! Hes singing and hes in his head voice and hes participating and doing really well. the musical activities more readily. She explained that the single sex classes completed their lessons faster than the coeducational classes because the boys were willing to do what she asked them to do the first time she asked them to do it. As for Teacher 3, she described the students in the single sex c lasses as proud, eager to please, excited, comfortable, and confident. Also, she explained that she noticed the boys singing participation more in the single sex classes because the boys could not hide behind the girls. In fact, she said that she picked two boys to participate in the All County Chorus because of this research study. She explained that she never really heard them singing in their coeducational classes. However, once these two boys switched to a single sex class, she noticed they were outstanding singers and she chose them to represent her school in the All County Chorus. She said she probably would not have noticed these boys singing participation if they remained in a coeducational class. She also claimed that a single sex envi ronment is risk free for boys who wanted to sing. These data show that all three teachers noticed numerous advantages to conducting singing activities with boys in single sex classes.

PAGE 77

! %& Table 18 Qualitative Data Theme 3: Advantages of Single sex Classes Description Teacher Comments There were advantages to the single sex groupings. Teacher 1: I got to focus on stuff that was just for them. Like, I got to say, Cant hit that note? Sing it an octave lower. Teacher 1: I had an opportunity to teach them about reading the non top line of music. Teacher 1: The [boys] who are normally spirited and cant sit still and all over the place, blended in more in the boy group than they would have boy girl. Teacher 1: In the single gender, the antsy boys di d not stick out as much because everybody was a little more antsy. So because I wasnt worried about the outliers the ones who were, like, laying under a chair I got to just pay attention to some of the ones I wouldnt normally get to hear. Teacher 1: The boy girl starting to be boyfriend girlfriend, you dont have it so much in the all boy classes. Teacher 1: [ referring to one student in the single sex class ] I think that group is good for him. Teacher 2: [T he boys] would act like they liked the song they were singing, by their facial expressions, smiling while doing it. Teacher 2: Advantages, definitely is that its easier to get them to participate and also to admit that they like what were doing. Teacher 2: I think just getting them to adm it that they like what were doing in music class is a big step for some of them because they wouldnt have done that before. Teacher 2: I ended up being able to move at a faster pace [with the single sex classes] becausethey were more willing to do wha t I asked them to do the first time so I didnt have to keep spending as much time working on the same thing. Teacher 2: I really liked how receptive that they were to everything. They were so willing to do what I asked them to do. Teacher 2: The boys seemed a little bit more excited about coming to music. Teacher 3: [T he boys] were proud of themselves. Teacher 3: [T he boys] seemed eager to please. Teacher 3: [The boys] we re excited. Teacher 3: The advantage is that I noticed the participation.

PAGE 78

! %' Table 18 (continued) Qualitative Data Theme 3: Advantages of Single sex Classes Description Teacher Comments There were advantages to the single sex groupings. (continued) Teacher 3: The students [in the all boys class] were less inhibited. Students were comfortable with one another. Teacher 3: [The boys] actually said, Were not afraid to sing when its us. Teacher 3: It was good for them to be in an environment where they were risk free and they felt comfortable and confident and not hindered by what somebody else was going to think. Teacher 3: In my opinion, the single gender setting will always be a risk free environment for a boy. Time. I designed this study so the students would be in their new groups for eight music classes b efore the final data collection because I thought it might take some time for the treatment to take effect. Contrary to what I expected, Teacher 2 and Teacher 3 noticed an increase in singing participation the very first day they had the single sex classes (see Table 19) Teacher 2 also explained that she saw the opposite effect in the coeducational classes. She claimed the boys in the coeducational classes participated less when they first switched to their new coeducational groupings. She hypothesized tha t, even though the boys were switching from one coeducational class to another, they needed time to adjust to the new group of students. Until these boys felt comfortable with their new classes, several of the boys participated less in the coeducational cl asses immediately following the shift to the new groupings. Time also had an effect on the boys behavior. Teacher 1 and Teacher 3 reported that the students were better behaved immediately after they were switched to the single sex classes. These teachers explained that the boys behavior deteriorated over time. They

PAGE 79

! %( Table 19 Qualitative Data Theme 4: Time Description Teacher Comments Two of the teachers noticed a sudden change when some of the boys switched to single sex classes. Teacher 2: I noticed right away that it changed.I could actually hear sounds coming out of their mouths from the first day that we switched. Teacher 2: Right from when we switched them to all boys, they would sing out. Teacher 2: It surprised me at how quick there was a noticeable change. With the all boys class I instantly noticed that they participated more. The boys would sing to where I could hear them. Teacher 3: The first day of the switch I had the all boy fourth grade class, and they floored me! Teacher 3: Right away I noticed that the boys, as a whole, were less inhibited and more willing to participate. Time had an effect on the students behavior and participation. Teacher 1: When we first started, the [single sex] boys were less antsy and they were getting through a lot of stuff, more than the mixed classes. The first couple, maybe three times, we did limbo at the end because they finished their lesson. But then they got used to each other. The more they got used to each other, the more they wou ld get off task. Teacher 2: In the coed classes I think that when we switched them around they started participating less than they had. They were used to coming to music with their class. They had already gotten to know it. And then as soon as they were mixed up with different kids, still being coed but they had new kids in there to face, they kind of shut down a lot more and were much harder to get to do anything. Teacher 3: [ describing the boys behavior ] The boys started off in the beginning, when t hey didnt know each other that well, not too much of a problem. Teacher 3: Maybe it was just the first daywere not comfortable acting up in front of each other yet syndrome. noticed the boys acted up and were more off task once they got used to the other boys in the single sex class. In conclusion, the data indicate that reconfiguring the s tudent groups had an immediate e ffect on the boys singing participation. Many of the boys in the single sex

PAGE 80

! &) classes immediately started to sing more, and some of the boys in the coeducational classes immediately started to sing less. Time had an effect on the boys behavior as well. The boys in the single sex classes were relatively well behaved at first, but their behavior deteriorated the more time they s pent in the single sex groupings. Behavior problems. When I began coding the data I used a code called Disadvantages of Single sex Classes, however as I coded the data I was only able to identify one disadvantage: behavior problems. All of the teachers reported issues with the boys behavior in the single sex classes (see Table 20). They described the boys as spirited, more rambunctious, energetic, and rough. They explained that the boys fed off each other in such a way that undesirable behaviors were more prominent. In addition, Teacher 1 pointed o ut that in her coeducational classes she separates misbehaving boys by placing girls in between them, and she could not use this strategy in the single sex classes. In total, the teachers described behavior problems in 5 of the 6 single sex classes. These findings differ from those reported in my review of the literature. Carp (2004) surveyed 101 choral directors who taught single sex choirs at the secondary level, and the majority of the directors identified student behavior as better in single sex enviro nments. It is possible the differences in the findings are because Carps study described the behavior of secondary students while this study describes the behavior of students in the upper elementary grades. Two of the teachers in this study reported hav ing difficulties with students at times of transition. While these teachers may feel transitions are a time of chaos and combustion throughout the school year, the experimental design of this study added to the

PAGE 81

! &* Table 20 Qualitative Data Theme 5: Behavior Problems Description Teacher Comments There were behavior problems in most of the single sex classes. Teacher 1: The boys were spirited in general in the all boys classes. In the mixed group the boys and girls seemed to balance each other out. Teacher 1: All the boys were a little more rambunctious. Teacher 1: Those boys are rough. The fifth grade boys are rough. Teacher 1: The spirited ones start to drag the other ones with them a little bit. Teacher 1: We took three classes worth of boy s and concentrated it into two so your behavior problems are more concentrated. Teacher 1: Theres so many boys. Its not like you can separate them with girls. The problem boys, you can put empty chairs in between them, but its easier to walk past an e mpty chair to get to somebody than it is to walk past two girls with their legs [ she sticks her legs out in front of her chair to demonstrate ]. In the coed, I just had more options before I had to find a way to put them in a corner. There would be somebody to separate them with. Teacher 2: They kind of fed off each other. If one would start talking, the other ones they were almost in competition with each other to see who could push the limits of the expectations that we have in here. Teacher 3: It wa s hard to contain them and keep them focused and on task. Theyre just all over the place. Teacher 3: It was tough. It was tough to man those all boy classes. Teacher 3: The boys were just hard to contain because of the dynamic in the room. Teacher 3: The on ly downside is that they are so energetic. I really have to keep my thumb on them because if there is any downtime (changing CDs, getting new materials, etc.) I lose them. Transitioning into the new groups right before class may have affected the boys behavior. Teacher 1: The fifth grade boys will break out into a fight before you know it. Weve had a bunch suspended because of that. And its during transitions. Teacher 3: Transitions were a disaster. I mean, just a nightmare.

PAGE 82

! &+ potential ch allenge with transitions. Students were reassigned to their groups immediately before going to music class each day. This regrouping may have exacerbated the difficulties normally faced when students transition from their homerooms to the music classroom. Particular groupings of students. The particular groupings of students in each class had an effect on both the students behavior and participation (see Table 21) Teacher 2 and Teacher 3 felt an unusually high number of students with behavior problems ended up in the single sex classes at their schools To avoid this in future studies, researchers could match boys according to the amount of behavior problems they cause and then randomly assign the boys to the new groups so a similar number of boys with behavior problems would end up in both the single sex and coeducational classes. In response to the effect caused by particular students being grouped together, Teacher 1 and Teacher 3 expressed a desire to be able to choose which students were in which g roups so they could have some control over the personality of the class and choose what was best for individual students. Both Teacher 1 and Teacher 3 noticed a few boys who participated more in their new coeducational groups. One might not expect this cha nge since these boys moved from one coeducational group to another, but apparently the personality of the new coeducational class had a positive effect on some students behavior and participation. Teacher 3 clarified by explaining that she noticed many bo ys who participated more in their new single sex group, but that for one boy he did better in his new coed ucational group. She believed that single sex classes would always be a risk free environment

PAGE 83

! &# Table 21 Qualitative Data Theme 6 : Particular Grouping s of Students Description Teacher Comments Some of the t eachers felt like a disproportionate amount of boys with behavior problems ended up in the single sex classes. Teacher 2: In my fourth grade group some of my boys have problems with beha vior, this one particular group of boys. So I think in that cla ss with that mixture of boys [behavior problems were] heightened a little bit. Teacher 3: It may have just been the random grouping of kidsWe ended up with a lot of our behavior problems in the al l boys classes. Teacher 3: There really were, in the all boys fifth grade class, like, eight behavior problems who are major behaviors problems in school. The combination of particular students had an effect on individual boys participa tion in the coe d ucational classes Teacher 1: In the mixed group I had a couple of boys who were normally pretty quiet and they seemed to perk upI got a little more spirit out of them. Teacher 3: [ describing one boy who was in a coed ucational class in his new group ] When he got into a different group all of the sudden he started singing. Away from his friends and his buddies, and all of the sudden hes singing and its beautiful. Teacher 3: The personality of the class impacted the coed groups. It could be that th e cultures in the coed classes were more comfortable than that of the original coed class for some students, but that would be hit or miss. Some of the teachers wished they could have placed students in the groups. Teacher 1: Wouldnt it be nice if we could just hand pick them? Look! Youre doing so great being in a single gender all boys. Lets put you here and lets take out [ names some students ] and lets just put them somewhere else. Teacher 3: If we were placing them we could spread those beha vior problems out over the whole grade level. conducive to increasing boys singing participation, but she admitted that occasionally a change in coed ucational classes may create a safer environment and increase some boys participation in a new coeducational setting. The impact of the research design on teachers. In order to participate in this study, the music teachers at the three school sites had to have th e students sing on a

PAGE 84

! &$ regular basis. In my final interviews with the teachers, two of them mentioned that they had their students sing more during this study than they usually would (see Table 22) I did not intend for the teachers to change their regularly scheduled lessons for this study, as I had assumed that the students sang on a regular basis anyway. It wa s unclear from the data if this increased amount of singing had an effect on the students attitudes, behavior, or participation. Teacher 3 admitted that throughout the study she was highly aware of the research. She paid close attention to the boys singing partic ipation, more than she normally would. I questioned her about possible biases or reactivity that may have occurred because of the study. Sh e assured me that although she was more aware of the boys singing participation, she strived to present herself the way she normally would at all times. For example, she did not go out of her way to either encourage or discourage the boys from participati ng during group singing activities in the single sex classes. Table 22 Qualitative Data Theme 7: The Impact of the Research Design on Teachers Description Teacher Comments One teacher was more aware of singing participation because of the study. Teacher 3: [The research] drew my attention to the boys singing participation. Two of the three teachers had their students sing more than normal. Teacher 1: Without a performance it would tend to be odd for us to sing this much. Just sitting and singing is not what I would normally do unless it was prepping for a performance. Teacher 3: In the older grades, I kind of dont focus on the singing as much.

PAGE 85

! &" Although researchers work to minimize reactivity, it is common for participants to be affected by research even in a naturalistic setting. The qualitative data indicate that two of the teachers did more singing with their students than normal, and one tea chers awareness of the boys singing participation was heightened. There is no evidence to indicate that either of these factors had a significant impact on the data. Other. There were several codes that appeared less frequently and/or less substantially in the qualitative data. Some of these factors were mentioned by only one teacher, only one time. However, after reviewing the data, I decided it was important to report each of these findings. They appear in Table 23 under the theme Other. First, the d ata indicated that the teachers words and actions may have influenced student participation during singing activities. Two teachers mentioned that if they encouraged or coerced their students, they could get them to participate more. One of the teache rs worked at a school that used a token economy as an incentive program. In the videos, I observed this teacher offering Paw Bucks to students who would participate fully during singing activities. At the other two schools, I noticed the teachers used pr oximity control to increase participation. In the videos, these teachers can be seen walking over to students who were not participating and standing near them to encourage their participation. Many of the 538 videos featured examples where the students s inging participation appeared to be a result of the teachers words and actions Second, when I asked the teachers to compare the participation of the boys in the single sex classes to the boys in the coeducational classes, some of the teachers reported that they did not notice a substantial difference. This surprised me because at other points in the qualitative data these same teachers indicated that they observed noticeable

PAGE 86

! &% Table 23 Qualitative Data Theme 8: Other Description Teacher Comments Students sang more when the teacher encouraged them. Teacher 2: [ referring to the students in coed ucational classes ] Coercing them could maybe get them to sing a little bit. But thats about it. Teacher 3: I found that no matter which group I had whether it was the boys or the coed if I encouraged them to sing, they did, but if I didnt say anything, there were some in both groups that would not participateIf I didnt encourage, they didnt participate fully. When asked to compare the boys in the single sex and coed ucational classes, some teachers noticed a difference and some did no t. Teacher 1: I still got a lot of just sitting [by the boys in the single sex classes]. Fifth grade. In the fourth grade all boys class there were some boys who sang more. Teacher 2: There was definitely a noticeable difference between having the coed class [and the single sex class]. Teacher 2: I really believe that I saw a difference in participation in the all boys class, and I wasnt even thinking ahead of time that I would. Teacher 2:Motivation and energy levelThose are two things that definitely were different in my eyes for the all boys classes. Teacher 3: I didnt notice a huge difference [between the single sex classes and the coed ucational classe s]. The teachers believed the boys were influenced by the presence of the girls. Teacher 1: I think [the boys] are normally trying to be macho. Some of those boys wouldnt sing normally in a mixed class. Teacher 1: I definitely think that if I were g oing to be able to get boys to sing because they loved it, it would have to be without the girls. Im very, very positive of that now. Teacher 2: That stood out to me, that [the boys] actually do like coming to music, they just dont always want to admit it around other peoplegirls. Teacher 3: I asked [the boys], Why dont you do this when there are girls in the room? And they actually said, Well, because they might make fun of us, or What if our voice cracks? or They might say we sound like gir ls. The boys in the single sex classes needed to move around more. Teacher 1: Movement, movement, movement, movement. All right. Youre driving me crazy. Stand up. Too crazy. Sit down. Stand up. Sit down. A lot of that. Teacher 1: Its the roughho using aspect. It goes like that [ snaps ] from playing to fighting. And to be fair, until it crosses the line, theyre just being boys. I try to ignore it as

PAGE 87

! && Tabl e 23 (continued) Qualitative Data Theme 8: Other Description Teacher Comments The boys in the single sex class es needed to move around more. (continued) much as possible, but I think thats how they play. Thats why the movements important. Some of the boys fake singing. Teacher 1: There may be some boys in the videos who appear to be si nging but actua lly arent.. .Theyll fake it but they wont actually make the sound come out. One teacher thought some boys would never sing and some would always sing. Teacher 1: There are some boys who will never sing regardless [of being in a single sex or coed ucational class]. Theyre not. Theyre just not! And they dont need to. I hate to say it that way. If its not their passion, if its not something they enjoy, its probably not something they are going to do. Teacher 1: Of course, you have the kids who are fearless and will participate no matter what. Student participation may be influenced by the teachers attitude towards singing. Teacher 1: Singing is not necessarily my thing. I was definitely an instrumentalist way before I was a singer and that probably comes through. I prefer instruments. Ive found that unless we are preparing for a performance, my kids are like, Whats the point? Why are we singing? And its hard f or me to give them that answer Unless theres a performance I dont feel theres that motivation for them to get it really, really, really, really good. differences. For example, when I asked Teacher 3 to Talk a little bit about how the boys participated in singing activities when there werent girls in the r oom , the teacher responded by explaining the substantial change she noticed in the single sex classes. She ended her response by saying, I thought, Wow! It [ referring to the single sex grouping ] really does make a difference. My very next question was Compare the boys participation in the sing le sex class to the coed group, and this time Teacher 3 stated, I didnt notice a huge difference.

PAGE 88

! &' It is possible the apparent inconsistencies in the teachers comments were caused by subtle differences in t he teachers impressions of the results. For example, perhaps a teacher noticed substantial differences in the students in the single sex classes compared to their participation in their regular coed ucational classes (a within groups factor that is, the same students in coed ucational versus single sex classes ), but the teacher did not notice a difference between the single sex classes and the coed ucational classes during the treatment phase (a between groups factor that is, comparing different students i n the two different classes ). Each teacher reported differences in the boys singing participation during the course of the study, but not all of the teachers noticed a substantial difference between the single sex classes and the coeducational classes. Th e third code in this theme addresses the teachers impressions of why boys do not sing very much in coeducational classes. All three teachers attributed the boys limited singing participation to the presence of the girls. Teacher 1 believed the boys tried to act macho in front of the girls by refusing to sing. Teacher 3 asked the boys why they did not sing as much in front of the girls, and the boys expressed a fear that the girls would make fun of them. Although this research does not directly address w hy boys may or may not participate more in single sex classes, the teachers in the study were curious about the influence of the girls and speculated on what effect the girls might have on the boys singing participation The final four cod es i n Table 23 all contain data from Teacher 1 only. This teacher spent a great deal of time reporting information that was not specifically addressed i n the interview guide. The next four paragraphs report the f our cod es that emerged from the data from Teacher 1 alone.

PAGE 89

! &( Teacher 1 noticed the boys needed to move around much more than the girls. The boys would rock in their chairs, leave their seats, lie down on the floor, and roughhouse with one another. She felt it was important to give the boys additional opportunitie s to move, and the single sex classes made it possible for her to increase movement activities specifically for the boys. At the end of the data collection, Teacher 1 warned me that some of the students in the videos might have appeared to be singing even though they actually were not. Apparently, a number of her boys lip sync in class and she does not reprimand them for this. She considers lip syncing a compromise to participating fully in singing activities. She was concerned about the validity of the me asure since I could not score the boys on what I heard; I could only score the boys on what I saw. She warned me that looks could be deceiving. During the course of scoring the videos, I noticed that some of the boys always scored very high in singing part icipation while others never even opened their mouths. For some students, their participation was consistent regardless of the song, the group, or the point in time. During my interview with Teacher 1, she reported similar findings. At one point she said, There are some boys who will never sing regardless. Shortly thereafter, she added, Of course, you have the kids who are fearless and will participate no matter what. When she reported her overall impression of the study, she suggested it was the middl e of the road people who demonstrated changes i n participation due to the single sex setting Finally, Teacher 1 admitted that she did not enjoy doing singing activities with her students and she thought the students may have picked up on her aversion to singing

PAGE 90

! ') activities. She did not maintain a high level of expectation for the students participation in the videos. It was common to observe students talking to their neighbor for most of the song, or lying down on the floor under their chair, or balancing their book on their head while they were singing. Teacher 1 rarely reprimanded these students. In one of our discussions, she told me that she picked her battles with her students, and it appeared that singing participation was a battle in which she did not engage. The quantitative data corroborate this observation; the mean scores for the students at School 1 were consistently lower than the scores for the students at the other two schools. Qualitative data by school. While one can look collectively at the qualitative data as it is organized by theme, it is also possible to look at the similarities and differences in the qualitative findings when comparing the results by school. Earlier in this chapter I provided descriptive statistics by school. In this section I provide the qualita tive findings by school as this makes it easier to identify areas of convergence and divergence between the qualitative data and the descriptive statistics. School 1. School 1 was a Title 1 school. The music teacher reported many behavior problems with the students at her school, not just in music class but also throughout activities on campus. Teacher 1 did not use assigned seats in her classroom, and her class was les s structured than the classes at the other two schools. She reported that most of her male students did not enjoy singing activities, and she admitted that she did not do a lot of singing activities with her students unless they were preparing for a perfor mance. Singing participation was limited at this school both during the baseline data collection and at the end of the treatment period. In the videos, boys could be seen lying

PAGE 91

! '* under their chairs, balancing their music books on their heads, and talking to their neighbors during singin g activities. The music teacher used proximity control and occasional verbal and nonverbal praise to encourage her boys to sing, but overall she did not maintain a high expectation for singing participation. Teacher 1 believed that the boys in her fourth grade single sex class participated more during singing activities than the boys in her fourth grade coeducational class. She did not notice a difference in singing participation with the boys in her fifth grade single sex and coeducational classes. When asked about the singing participation of the boys in her fifth grade single sex class she reported, I still got a lot of just sitting. She also reported that the boys in the single sex classes were more spirited than the boy s in the coeducational classes. Teacher 1 warned that several of her boys lip sync during singing activities. She explained that she encourages them to fake it by moving their mouths even if they are unwilling to sing. She reported that some of her boys who would normally lip sync actually sang in the single sex classes. She also stated that some of the boys who typical ly would not open their mouths started to lip sync in the single sex classes. Although these boys were not actually singing, Teacher 1 bel ieved that lip syncing was a step in the right direction for them. In general, Teacher 1 reported an increase in participation in the single sex classes. She offered many examples of individual students who participated more in their new single sex groupin gs. She also noticed a couple of boys who perked up when they were assigned to a new coeducational group, probably due to the chemistry of the new coed ucational class. In addition, Teacher 1 reported that more boys would volunteer

PAGE 92

! '+ during solo singing a ctivities in the single sex classes than the coed ucational classes. Overall, she expressed support for single sex music classes. She liked that she could tailor her instruction to meet the specific needs of the boys. She also believed the boys were more li kely to sing in a single sex class. She emphasized, I definitely think that if I were going to be able to get the boys to sing because they loved it, it would have to be without the girls. Im very, very positive of that now. The descriptive statistics c orroborated the qualitative data from Teacher 1. She reported that her boys did not like to sing, and this was evident in the low mean scores for both the boys in the single sex and coed ucational groups at all four points in time. There was only one time w hen one of the two groups from School 1 had a higher mean score than one of the groups at the other two schools. Other than that one exception, the boys at School 1 consistently participated less than the boys at the other two schools. As for the effect si zes, School 1 showed negative effect sizes during the baseline data collection. This suggests that prior to the treatment period, the boys who would later be assigned to single sex classes participated less than the boys who would later be assigned to coe ducational classes. By the end of the treatment period, School 1 s howed a positive effect size ( d = 0.05). While this effect size is very small, it signifies a change from the small to moderate negative effect sizes at Time 1 and Time 2. School 2. Like School 1, School 2 was also a Title I school. Teacher 2 mentioned that there were several students with behavior problems at her school, although I did not notice any major issues with behavior when I observed the videos of her classes. Her students s at in assigned seats, followed directions, and almost always met the teachers expectations. School 2 used a positive behavior support system, and the videos showed

PAGE 93

! '# evidence of Teacher 2 implementing this system by offering Paw Bucks from the schools to ken economy system to students who were well behaved and participating. The class activities were structured, and the students spent a great deal of time on task. Of the three teachers, Teacher 2 was the most adamant about r eporting a noticeable increase i n the boys singing participation in the single sex classes. She admitted that she was skeptical about the research before it began, and she was surprised not only by the amount of the increase in the participation of the boys in the single sex classes but also by how quickly the boys participation changed. She noticed an immediate substantial increase in the boys participation in single sex classes. She also reported that the boys in the single sex classes were more willing to sing the first time she ask e d them, so they were able to mov e through their lessons at a faster pace than the boys in the coeducational classes. She mentioned that sometimes she could coerce the boys in the coeducational classes to sing, but it took a longer time and a great deal of e ffort and begging on her part She also said that the boys in her single sex classes were more willing to admit that they liked the singing activities. The teachers at School 1 and School 3 reported behavior problems with both their fourth grade and fi fth grade single sex classes. In contrast, Teacher 2 only reported behavior problems with her fourth grade class. She attributed the behavior problems to one particular group of boys in the fourth grade class. She did not have behavior problems with her fifth grade single sex c lass, and unlike the teacher at School 3, Teacher 2 did not think that behavior prob lems would always be an issue with a class of all boys. The effect sizes for School 2 at Time 1 and Time 2 were small and positive ( d = 0.11 and d = 0.33). This means the boys who would later be assigned to single sex

PAGE 94

! '$ classes participated more than the boys who would later be assigned to coeducational classes. While there were small differences in the groups prior to being assigned to new classes, t here was a large difference between the groups at Time 3 ( d = 0.71) The boys in the single sex classes participated in singing activities much more than the boys in the coeducational classes at Time 3. The effect size was smaller at Time 4 ( d = 0.32). Thi s may be due to differences in participation based on song. Overall, School 2 showed the largest positive effect sizes and Teacher 2 was the most adamant about concluding that she saw a substantial increase in singing participation for the boys in her sing le sex classes. School 3. Unlike Schools 1 and 2, School 3 was not a Title 1 school. Many of the students came from affluent families, and School 3 reported a lower suspension rate than Schools 1 and 2. The music teacher maintained a very structured class room, and her expectations for the students were high. She used proximity control and both verbal and nonverbal feedback to manage the students in her class. Teacher 3 reported a substantial increase in the boys singing participation in the single sex cla sses. Like Teacher 2, Teacher 3 reported that the increase in singing participation was immediate. S he asked the boys why they did not sing as much in the coeducational classes in which they were previously, and the boys told her that they were afraid the girls would make fun of them. Teacher 3 attributed the boys increase in singing participation to th e risk free environment that wa s created in the single sex setting. The vast majority of the comments from Teacher 3 indicated that she noticed an increase in singing participation for the boys in the single sex classes, however at one

PAGE 95

! '" point Teacher 3 reported that she did not notice a substantial difference between the si ngle sex and coeducational groups. She explained that while many boys in the single sex classes participated more than they had when they were in their coeducational classes earlier in the year, there were also some boys who participated more in their new coeducational groupings. She believed this occurred because some of the boys found the personality of their new coeducational class to be less threatening than their previous coeducational class. Teacher 3 stated: It could be that the cultures in the coed classes were more comfortable than that of the original coed class for some students but that would be hit or miss.. .In my opinion, the single gender setting will always be a risk free environment for a boy. Overall, she believed the single sex setting would increase boys singing participation, although she acknowledged that some of her students participated more in their new coeducational groupings. Teacher 3 reported behavior problems with both her fourth and fifth grade single sex classes. She belie ved that part of the behavior problems were due to an unusually high numbe r of challenging students that were assigned to the single sex classes. She also thought the behavior issues were confounded by the single sex setting. The effect sizes for School 3 were all positive with the exception of a very large negative effect size at Time 2 ( d = 1.16) The sample size was greatly reduced at Time 2 for this school due to technical difficulties during filming, and it is likely the reduction in sample size incr eased the sampling error in the data. T herefore, t he data for S chool 3 at

PAGE 96

! '% Time 2 has limited reliability. As for the effect sizes at Time 3 and Time 4, they were small and positive at d = 0.19 and d = 0.27. Overall. Each of the teachers reported increases in singing participation for the boys in the single sex classes, however the qualitative data varied by school. Teacher 2 spoke consistently about the substantial changes she saw in the singing participation of the boys in the single sex classes. Teacher 3 also reported changes, but she admitted that she didnt notice a huge difference between the boys in the single sex and coed ucational classes. Teacher 1 noticed an increase in singing participation for individual students, but overall she reported that many of her students were still unwilling to sing even in a single sex class. The qualitative data for the three schools corroborate d the descriptive statistics reported earlier in this chapter. The largest positive effect sizes were for School 2 ( d = 0.71 and d = 0.32). There were smaller positive effect sizes for School 3 ( d = 0.19 and d = 0.27). School 1 showed a negative effect size at Time 3 and a very small positive effect size at Time 4 ( d = 0.27 and d = 0.05, respectively). In addition, Teacher 1 reported the least amount of singing participation for the students at her school, and with only one exception, the group me ans for the students at School 1 w ere always lower than the group means for the students at the other two schools. Summar y : Qualitative data. Overall t he qualitative data show that there were noticeable differences in the boys singing participation in the single sex classes. The teachers noticed that many boys sang more, appeared more comfortable singing, and participated willingly without having to be coerced. Teacher 1 did not notice a major change in the boys participation in her fifth grade single sex group, but she noticed

PAGE 97

! '& increased participation for individual students. In the coeducational classes, the teachers felt like they had to beg the boys to sing, and some of the boys would sabotage singing activities by singing badly or simply not sing at all. The teachers described many varied advantages for teaching boys in single sex classes. The teachers reported only o ne disadvantage, and that was the boys behavior. In the single sex classes the boys were more spirited, rambunctious , and hard to contain. Time played a factor in this study. The music teachers noticed an immediate increase in the boys singing par ticipation in the single sex classes. Teacher 2 also reported an immediate decrease in singing participation for the boys who switched to a new coeducational class. As for issues with behavior in the single sex classes, the boys were relatively well behave d at the beginning of the treatment period. The boys in the single sex classes got more off task as the weeks progressed. In addition to differences in the single sex and coeducational classes, the data highlight other factors that affect ed singing partici pation. For example, t here were a few cases of boys who participated more in their new coeducational class es This suggests that the chemistry of the students in the class may affect singing participation regardless of whether the class is single sex or co educational. The teachers expressed a desire to hand pick which students were in each class. They thought that if the students were intentionally assigned to their new groups, there would be fewer behavior problems and a greater increase in singing parti cipation. In addition to changes based on the personality of the class there were significant diffe rences in participation based on the teacher who taught the lesson. The teachers who held higher expectation s for their students observed more singing parti cipation from their students.

PAGE 98

! '' Conclusion According to Patton (2002), Triangulation of qualitative and quantitative data constitutes a form of comparative analysis (p. 558). This mixed method design provided method triangulation and a wealth of rich data. When comparing the results of the quantitative and qualitative data in this study one area of divergence is clear. The ANOVA show ed no statistically significant differences between the single sex classes and the coeducational classes however in most cases the qualitative data showed a substantial increase in the boys singing participation in the single sex classes. There are a couple of probable explanations for the differences in the results. First, although the Singing Participati on Measure scores appear to be both valid and reliable, there is one major limitation for using this measure in a study of this design. The original Singing Participation Measure included a dimension called Voice. For this dimension, each boy was scored based on the sound that he produced during singing activities. Since this study focused on boys singing participation when they were singing with the entire class, rather than singin g solo, it was impossible to score each boy based on the dimension Voice While I could hear the overall sound of the classes in the videos, there was no way to identify individual voices within the group. Therefore, for this research I removed the dimension Voice from the Singing Participation Measure If I had chosen to observe the classes in person, rather than by means of a video, I may have been able to keep the dimension Voice as part of the measure. Second, the videos that I scored captured only one moment in time. I scored the final performance of the songs at each point in time. The quantitative data did not capture the effort or length of time it took to prepare the song to be sung. One of the teachers in

PAGE 99

! '( the study stated, I worked less to get the all boy classes to participate. She said that while the end results of the performances of the single sex classes and coeducational classes may have appear ed similar in the videos these brief moments in time did not show how the boys in the single sex classes were ready and willing to sing while the boys in the co educational classes had to be coerced into singing. This might have been a reflection of the difference between data based on brief moments in time and data based on extended periods of observation by the teachers. In summary, w hile the ANOVA wa s an import ant part of this researc h, it is vital not to dismiss the hypotheses based on the results of the ANOVA alone. The qualitative data show ed different results It is possible that scoring videos of groups of students might not be the best way to explore the v ariable of individual boys singing participation. One of the advantages of collecting qualitative data is that these data bring context to the issue at hand. Qualitative data can capture feelings, perceptions, and other information that may be difficult to quantify. One of the strengths of this study is that the mixed method provided both quantitative and qualitative data. Further more, although the ANOVA showed no statistically significant differences by group, the descriptive statistics corroborated the qualitative data. F ive of the six effect sizes at the end of the treatment period showed a positive effect on boys singing participation in single sex classes, including small, medium, and large positive effects for the groups at School 2 and School 3 for Time 3 and Time 4 (ranging from d = 0.19 to d = 0.71) Continued research on the effect of single sex education on boys singing participation is warranted.

PAGE 100

! () Chapter 5: Discussion The purpose of this research was to det ermine if upper elementary boys level of participation during group singing activities would be affected by attending music classes in a single sex setting. The hypotheses were: 1. Boys in single sex classes will participate during group singing activities more than boys in coeducati onal classes. 2. Boys in single sex classes will participate during group singing activities more at the end of the study than they did when they were in coeducational classes at the beginning of the study and they will participate more than the boys who remained in coeducational classes The sample ( N = 186) included students at three different school sites where students normally attend ed music classes in coeducational settings. Students were videotaped in their regular coeducational classes on two occasions. Then students were randomly assigned to new groups. Some students were assigned to single sex classes for music, and others were assigned to new coeducational classes. The students remained in their new groups for 8 music class es. Then students were videotaped two more time s at the end of the study period I collected quantitative data in the form of scoring individual boys on their participation using the Singing Participation Measure. This instrume nt measures singing participation only. It does not measure quality of singing, on task/off task behavior, or

PAGE 101

! (* any other factors. I collected qualitative data in the form of music teacher interviews and music teacher journal entries. The Research Hypotheses Revisited The task of measuring the degree of singing participation is somewhat complex. It is difficult to understand the issues at hand by looking solely at one form of data. One of the strengths of this study was the mixed method design. By collecting both quantitative and qualitative data, I was able to triangulate the findings. This triangulation was especially important in my data set as the quantitative and qualitative data did not corroborate. If I had only looked at one type of data, I would hav e limited the breadth and depth of the results. The ANOVA reveal ed statistically significant differences between the boys at the three school sites, as well as statistically significant differences between the boys participation at different poi nts in time. The differences in time may be the result of the teachers using a variety of songs on different days as the descriptive statistics show differences in the mean participation scores for the various songs The ANOVA showed no statistically sign ificant differences between the participation of the boys in the single sex and coeducational classes. In addition, the data indicat ed no statistically signific ant interaction effect for group (either single sex or coeducational) and time Although the ANO VA showed no statistically significant differences by group five of the six effect sizes at the end of the treatment period illustrat ed a positive effect on boys singing participation in single sex classes, including small, medium, and large positive effects for the groups at School 2 and School 3 for Time 3 and Time 4 (ranging from d = 0.19 to d = 0.71)

PAGE 102

! (+ The qualitative data showed substantial differences between the participation of most of the boys in the single sex and coeducational classes. The teachers noticed an increase i n many of the boys willingness to sing in the single sex settings. The teachers reported differences between the single sex and coeducational classes overall, as well as differences in the participation of individual boys These differences were evide nt immediately after the boys were reassigned to their new groups for music class. As for the students who switched to new coeducational classes, some of these boys participated less after the new groups began. Although the boys moved from one coeducational class to another, they were affected by the specific combination of students in the room. It took time for some of the boys to get comfortable with their new classmates before they would participate in singing activities. In a couple of cases, boys participated in their new coeducational classes more than they did previously. This was most likely due to the boys being influenced by the presence of certain individuals in their class. It is possible that some boys felt their new coeducational class was a safer environment than the group of students with whom they normally attend ed class. However, the majority of the boys participated less in the coeducational classes than in the single sex classes. In reference to the research hypotheses, the ANOVA did not support the hypoth eses but overall the qualitative data did This may be a testament of the complexity of the issue at hand. Although the Singing Participation Measure appears to be both val id and reliable, this measure may not capture all of the factor s related to singing participation. For example, w hile the original Singing Participation Measure i ncluded a dimension for Voice, I had to remove this dimension from the measure for the current

PAGE 103

! (# stud y. Since I was assessing individual boys within a group setting using videos of the entire class singing I could not score boys on the sound s they produced individually. As I watched the videos, I simply could not accurately assess which sounds were comin g from which students If I had scored the boys in person, I may have been able to hear individual voices within the group and maintained the dimension Voice on the measure. In addition, if I had scored the boys in person, I may have been able to identif y which boys were feigning participation by lip syncing. Another factor that was not captured by the quantitative data was the amount of effort the teachers had to exert to get the students to participate. The teachers reported that the boys were more willing to sing in the single sex classes and sometimes these boys even finished their lessons early. The teachers often had to beg the boys in the coeducational classes to sing. While it was possible for the teachers to get many boys to sing in the coed ucational classes, the boys begrudged participating in singing activities. While the degree of participation in the videos may appear similar between the single sex and coeducational cla sses, the quantitative data did not reflect the i ncreased effort exert ed by the music teachers to coerce the boys in the coeducational classes to sing. Furthermore, this study i ncluded a relatively small sample size. Smaller samples include a greater degree of sampling error, less power, and a higher chance of a Type II err or. This may explain why the descriptive statistics included some moderate to high effect sizes even though the ANOVA showed no statistically significant differences between the single sex and coeducational classes. Since the quantitative data do not corr oborate the qualitative findings, it is difficult to come to a definitive conclusion regarding the research hypotheses. The

PAGE 104

! ($ ANOVA did not support the hypotheses but overall the qualitative findings d id I believe the qualitative findings offer a more comp lete picture of the phenomenon of boys participation during group singing activities in single sex and coeducational classes. The qualitative data encompas s ed changes the teachers s aw throughout the research period not just at the two brief moments in time included in the ANOVA The teachers were able to identify individual voices when the classes sang, and this was a dimension I could not assess using th e Singing Participation Measure In addition, the teachers reported a decrease in the amount of effort they had to exert to get the boys in the single sex classes to sing. This wa s another factor that could not be reflected in the quantitative data. In summary, the qualitative data support ed the research hypotheses and explore d the issue of boys sin ging participation i n ways the quantitative data could not. Further study on the effect of single sex music education is warranted. Beyond the Research Hypotheses The music teacher interviews and journal entries included a wealth of information. While man y of the themes were not directly related to the research hypotheses, I believe it is important to discuss these additional considerations since they relate to single sex music education. In the following section, I discuss the advantages and disadvantages of single sex education, the impact of song choice on boys singing participation, and the influence the teacher has on singing participation. Advantages and disadvantages of single sex e ducation The teachers reported several advantages of single sex music education. One teacher gave her boys an opportunity to read the non top line of music. She also allowed her boys to sing one octave lower when the melody was too high. Another teacher said the boys in the single

PAGE 105

! (" sex class es were more like ly to admit t hey enjoyed activities in music class She found the boys in the single sex classes progressed through the lesson s faster because they were willing to participate the first time they were asked to s ing, and the teacher spent less time having to convince the boys to participate. The third teacher reported that the boys in the single sex classes were more confident and comfortable. She believed that the single sex classes were risk free environments for boys who wanted to sing. All of the teachers touted the many advantages of single sex music education. The single sex model enables music educators to tailor their instruction to meet the differing needs of boys and girls. This specialization can im pro ve music teacher effectiveness. The qualitative data revealed only one disadvantage for single sex education: behavior. All of the teachers reported some difficulties with the boys behavior in the single sex classes. The boys were more spirited, rambunctious, energetic, off task, and hard to contain. The students had to transition into their new groups right before coming to music, and this may have added to the boys high energy level. T wo of the teachers reported that an unusually high number of students with behavior problems ended up in several of the single sex classes. The teachers thought the behavior problems would have been reduced if they had been able to assign students to their new groups. Although the teachers witnessed some problem s with the boys behavior in the single sex classes, it appears the many advantages outweigh the one disadvantage. Teachers with strong classroom management strategies may not have a problem with a class comprised of energetic and spirited boys. For example one of the teachers noticed that her boys had fewer problems if she gave them multiple opportunities to move. Sax (2005) stresses, There are no differences in what girls and boys can learn. But there are

PAGE 106

! (% big differences in the best ways to teach them ( p. 106 ). Teachers of all boy classes can choose to incorporate more kinesthetic activities to accommodate boys need to move and alleviate some of the behavior challenges that may occur in a classroom full of boys. Harrison (2008) posits a number of strat egies for developing a sensory approach when teaching music to boys (p. 138). He suggests music educators provide opportunities for physical involvement in learning to help boys stay actively engaged in the lesson at ha nd. In addition to helping boys r emain attentive to the learning tasks, movement activities may help boys express some of their need for kinesthetic action and reduce unwanted behaviors that may occur when boys are asked to sit still. Music educators who are concerned with potential behav ior problems in an all boys class may benefit from adopting a sensory approach to music education and allowing boys opportunities to move. The advantages of single sex music education appear substantial, and skilled music educators may find any possible ch allenges in the behavior of the students in an all boys class may be a small price to pay for the potential benefits. Furthermore, teachers with well developed classroom management strategies and a variety of approaches to teaching music to boys may experi ence very few issues with behavior in a class comprised of all boys. Song choice. The quantitative data showed significant differences in boys singing participation at different points in time. Since the music teachers used different songs on different days, it is possible that the boys participation may have differed, in part, based upon the song that was used. Although this is beyond the scope of this research, it is possible that girls would also participate to vary ing degrees based upo n the repertoire. In addition, research shows differences between the song preferences of boys

PAGE 107

! (& and girls ( Lehmann, Sloboda, & Woody, 2007). Whether teaching boys in a single sex or coeducational setting, it is important for tea chers to be aware that boys m ay be more willing to sing certain songs Of course, teachers do not have to choose repertoire based solely on the preferences of the students. For example, a teacher may feel it is important for all students to sing The Star Spangled Banner regardless o f whether or not they enjoy singing it. That being said, teachers can strive to include a variety of son gs that appeal to boys when choosing repertoire for their classes. The teachers influence The singing participation of the boys at School 1 was signi ficantly less than the singing participation of the boys at the other two schools. In the videos, one can see the boys talking to their neighbors, lying down under chairs, and balancing their books on thei r heads. Although the teacher at this school said that she expected all students to sing, her words and actions did not appear to reinforce this expectation. Her primary form of classroom management was proximity control. When she would stand next to students, most of them wo uld sing, ho wever it was impossible for her to stand next to every child at every moment. As she would move away from boys, many would stop singing. She did not ask the boys lying on the floor to sit in the chair s properly, and she did not reprimand the students who b alanced the books on their heads. The students were allowed to choose their own seats, and the classroom was less structured than many other classroom environments. My general impression was that the boys knew the teacher wanted them to sing, but they also believed there would be no incentives or consequences as a result of their participation or lack thereof. The teacher admitted that she did not enjoy doing singing activities with the students, and she suspected that the students were aware of her reserva tions. She also said that she chose

PAGE 108

! (' her battles with her students, and it was clear that she chose not to battle with the boys regarding their limited singing participation. These observations reinforce the important role of the teacher in education. The teachers expectations and attitudes have a direct impact on the students he or she teaches. Essentially, the teacher can create an environment in which students think of themselves as singers. The teacher can establish a routine that involves singing on a daily basis, and the teacher can find creative ways to motivate students both extrinsically and intrinsically. Teachers can award incentives to both individuals and entire classes who perform well Since the teacher is a paramount factor in cultivatin g singing participation in students, one cannot have a comprehensive discussion of the topic without acknowledging the great impact teachers have on student performance. Implications for Music Educators Based upon the findings of this research, I posit fi ve implications for music educators : 1. Create single sex singing opportunities 2. Remember that one size does not fit all 3. Choo se repertoire mindful ly 4. Establ ish a singing culture at the school 5. Know your students I explain each of these implications in the sections that follow. Create single sex singing opportunities. According to the qualitative data, many of the boys participated more during singing activities in the single sex classes. One of the teachers stated, In my opinion, the single gender set ting will always be a risk free

PAGE 109

! (( environment for a boy. Music teachers at schools with coeducational classes can create single sex singing opportunities to establish a risk free environment where boys might be more willing to sing. In addition, teachers can modify their instruction to meet the specific needs of boys in a single sex setting. The teacher can choose specific repertoire, include more gross motor movement, and employ instructional strategies that are appropriate for teaching boys. There are se veral different ways to create a single sex environment. For this study, the teachers rearrange d students into new groups immediately prior to music class to form the single sex classes At these schools, the entire grade level of students attended music, art, physical education, and other special area classes simu ltaneously. This meant that rearranging the student s affected all of the special area teachers. This plan would work at other schools where the entire special area team is willing to rearrange the students to have the opportunity to teach single sex classes. At many schools, the special area teams may not be willing to rearrange their classes. If this is the case, the music teacher might be able to esta blish a single sex choir before, after, or du ring school instead of rearranging the students in the regularly scheduled music classes. Boys could either self select to participate in the choir, or the teacher could invite certain individuals to join the choir based upon the teachers belief that thes e students would benefit from a single sex setting. If teachers cannot establish a single sex environment, either in the context of the regular music class or in the form of a single sex choir, then the teacher could provide single sex singing opportunities within the coeducational class. For example, if a teacher taught partner songs, he or she could have the boys sing one song while the girls sing the

PAGE 110

! *)) other. This arrangement is not ideal since the boys still may be inhibited by the presence of girls in the room. However, dividing the coed ucational class wo uld make it so the boys could not hide behi nd the girls by letting the girls do the majority of the singing. In the example using two partner songs, it would be blatantly obvious if the boys were not singing their song. Sometimes boys can fake singing participation when boys and girls sing together because often the girls will take on the majority of the singing responsibility. Creating single sex singing opportunities within the coeducational class will spotlight the boy s singing participation Of cou rse, the ideal situation would be to establish a single sex class or choir since the boys may feel more confident and willing to sing in an environment solely comprise d of boys. Teachers who are able to work with boys in a single sex setting should be pre pared to accommodate the needs of boys. All of the teachers in this study reported some difficulties with the boys behavior in the single sex setting. The boys may seem more spirited or energetic when surrounded by other boys. It may be helpful to all ow ample opportunities for the boys to participate in gross motor movements. In addition, the pacing of the instruction can have a substantial impact on th e boys attentiveness. One teacher warned that she could not have any down time or she would lose t he boys focus. Teachers should be prepared for potential challenges with the behavior of students in a single sex boys class and may need to alter their teaching methods and classroom management strategies accordingly. Remember that one size does not fit all. While some boys may flourish in a single sex environment, other boys may excel in a coeducational setting. There were several boys who received the highest possible scores on the Singing Participation

PAGE 111

! *)* Measure during the baseline dat a collection when t hey attended music class with their regular coeducational class. These boys were not only willing to sing, but sang enthusiastically in a mixed setting. There was no need to use single sex education as an intervention for these students since they we re already successful singers in a coeducational class. Every teacher is responsible for knowing his or her students and differentiating instruction to best meet the needs and interests of individual students whenever possible. There is no one magic int ervention that will always prove effective with every child. While single sex music education may increase some boys willingness to sing, it would be an overstatement to suggest that every boy must attend music class in a single sex setting to be success ful. Two of the teachers in this study expressed a desire to hand pick which students were in each group. Teacher 3 mentioned this in reference to splitting up boys with behavior problems, however Teacher 1 made this suggestion in reference to meeting th e needs of individual students. She recognized that some of her students did better both academically and socially in a single sex environment while others excelled in a coeducational environment. I believe the best case scenario is to offer both single se x and coeducational classes so teach ers, students, and parents can select the type of class that would best serve the needs of individual students. Unfortunately, many elementary music teachers teach the same lessons to the students in each grade level in the same manner. While teachers are responsible for teaching certain music standards to all students, music educators can teach the same benchmarks in a variety of ways. For example, singing activities can include singing as a large ensemble, singing in s mall groups, singing solo into a microphone, and singing

PAGE 112

! *)+ different types of music such as folk songs, partner songs, and popular music, just to name a few. Just as the one size fits all approach may not be the best choice when it comes to single sex educ ati on, teachers may find that one size does not fit all when it comes to other approaches to music education as well. The students in one third grade class may benefit from spending a lot of time engaged in singing activities, while the students in ano the r third grade class may benefit from fewer singing activities and more composition activities. Effective music educators differentiate their instr uction to meet the needs of specific classes as well as students within c lass es Choose repertoire mindfull y. According to the data, the boys level of participation varied by song. It is important for teachers to be aware of the impact repertoire has on students singing participation. Research shows that boys and girls have different song preferences (Lehmann, Sloboda, & Woody, 2007). Therefore, if a teacher has the opportunity to teach single sex groups, he or she may choose entirely different repertoire for the all boys groups and the all girls group s Choosing appropriate repertoire can increase singing participation. For example, fifth grade boys might refuse to sing a song about kittens but might be willing to sing a song about a journey to the moon. Careful selection of the repert oire may positively impact the boys singing participation. That be ing said, teachers should beware of choosing gender typical repertoire as this reinforces gender stereotypes. While I believe it is appropriate to choose gender typical repertoire to cultiv ate a love for singing in students who refuse to sing I believe the ultimate goal is to create an environment where it is safe for boys and girls to explore gender atypical songs. Essentially, it may be acceptable to use stereotypically masculine songs to g et boys interested in singing,

PAGE 113

! *)# however once the boys are intereste d in singing, the music teacher should work to b roaden the repertoire of the boys to gender atypical songs Establish a singing culture at the school. The teacher sets the tone for the culture of his or her classroom. A music teacher can create an environment where it is safe to take risks, where students know they will not be laughed at if they sing at the wrong time or if their voice cracks as they sing. A music teacher can encourage students to sing, offer incentives for good singing, and inspire the students to think of themselves as singer s. Even in an all boys class, boys might not be willing to sing if the teacher has an unfavorable at titude towards singing Children are very impr essionable. Music teachers can foster singing participation by establishing what I refer to as a singing culture . Students who identify with the role of singer, are praised for their singing efforts, and know their teacher has high expectations for the ir participation are more likely to engage fully in singing activities. Know your students. One of the teachers in this study made a brave confession in her interview; she admitted that prior to the study she never noticed that some of her boys were not participating during singing activities. She explained that d uring the course of the study she paid closer attention to the boys singing participation and she began to recognize which students faked singing. As she became more aware of the lack of part icipation from some of the boys, she wanted to encourage the boys to sing more. It is possible that many educators in the field lack an accurate assessment of each boys singing ability and willingness to sing. As educators, we can strive to know our stu dents and know them well. Each student has strengths and weaknesses as a developing musician. Some students may even have their own musical goals and

PAGE 114

! *)$ interests, and these goals and interest s may lie within the music classroom, outside the music classroom, or both. Music teachers who know their students well may be able to identify musical aptitude and foster the development of students who might be successful in a career in music. In some cases, boys may enjoy singing but may not feel like they have the s upport of their family and peers to pursue singing activities. Adler (2002) conducted a qualitative study on boys experie nces of singing in school. H e grouped students into five categories: Jocks of Singing, Sensitive Boys, Neutral Boys, Non Singers, or Bad Asses. He concluded that boys in the different groups were influenced by gender stereotypes to different degrees. Music educators who are aware of students social influences may have a better understanding of why students behave as they do and what te achers can do to motivate students to participate in singing activities. Music educators can serve as m entors who support the singing interests of young boys and girls alike. When m usic teachers know their students well, they can modify their instruction within the classroom setting and offer encouragement and assistance outside of the classroom to positively impact the development of young musicians. Future Research Currently, there is very little research available on single sex music education at the e lementary level According to the National Association for Single Sex Public Education (NASSPE) there are 540 schools in the United States that offer single sex environments, but the majority of these are coeducational schools that offer single sex classe s ( ,-.). In some of these cases, boys and girls may attend music class in a coeducational setting even though they attend their regular classes in a single sex group.

PAGE 115

! *)" So, while the NASSPE reported an increase in single sex education over the past 8 years, there has been less of an increase in single sex music classrooms. Many high schools offer single sex choirs, but single sex music classes are rare at the elementary level. Since there is very little research on the phenomenon of single sex elementary mus ic education, further research is needed. Replication studies would be highly valuable. The cu rrent study has a relatively small sample size ( N = 186 total with N = 123 for the ANOVA), so power was limited and the possibility of a Type II error was increased. Ideally this study would be reproduced with a much larg er sample size. In addition, a future researcher could score individual boys on their degree of participation by visiting them in person rather than by watchi ng videos. If the researcher were present during the singing activities, he or she could include the Voice dimension of the Singing Participation Measure to improve the validity of the scores. Furthermore, it would be useful to conduct replication studies with students at schools with male music teachers. All of the music teachers in this study were female; it is possible the results would be different if some or all of the music teachers were male. If rep lication studies are conducted, future researchers may consider collecting data on boys in single sex classes at one school and boys in coeducational classes at another school. In the current study, there were single sex and coeducational classes at each s chool. I chose this design because I wanted to collect data from boys from the same school population with the same music teacher studying the same curriculum in single sex and coeducational settings. However, some researchers may be concerned that having single sex and coeducational classes at the same school could result in treatment

PAGE 116

! *)% diffusion Boys in the various classes had opportunities to interact throughout the rest of the school day at lunch, at recess, and in their homeroom classes A s the boys interacted, they may have discussed the new configurations of classes, and their discussions may have impacted their behaviors. For example, one boy might tell another, You are so lucky! You get to be in the all boys class! Then, the boy in th e single sex class could feel a sense of pride and might perform better in music class. A researcher can reduce this contamination by creating single sex classes at one school and coeducational classes at another school since students at different school s ites will have limited contact, if any Unfortunately, there is usually a sizeable tradeoff when changing the design of a study ; if students in the treatment and control groups attend different schools, they will probably come from different school populat ions, have different music teachers, and experience different curricula. These extraneous variables limit the internal validity of the study. Future researchers may choose to alter the design of the current study rather than replicate it, however researche rs must weigh the strengths and weakness of various designs when planning research on this topic. Another important study would be a qualitative study on why many boys in the upper elementary grades are hesitant to sing in a coeducational setting. This st udy might include interviews with the students, as well as the teachers and parents. In addition, students could complete a self report survey or an open ended questionnaire to highlight some of the issues they might not be willing to report in a face to f ace interview. This rese arch would go beyond the question of Do boys in single sex classes sing more than boys in coeducational classes? and delve into the realm of w hy boys might participate less when there are girls in the room.

PAGE 117

! *)& Further study of music preferences of boys and girls may also be helpful. If it were possible to identify which songs boys and girls prefer, one could compare the songs to the repertoire in various elementary music curricula. It is possible there is a gender bias in favor of fe males when it comes to the choice of songs taught in the elementary music classroom. This may reinforce gender stereoty pe that singing is for girls. Ideally, songs used in a coeducational class would appeal to both boys and girls, and the songs used in s ingle sex classes could accommodate differences in the song preferences of boys and girls. Continued research is necessary to explore the many facets of boys willingness to sing in educational settings. Conclusion Research shows that g irls participate in singing activities much more than boys ( Eccles, Wigfield, Harold, & Blumenfeld, 1993; Haladyna & Thomas, 1979; Mizener, 1993; Moore, 1987; Sherban, 1995 ). In fact, boys participation in singing activities has declined over the past century (Gates, 1989). This decline may be due to gender stereotypes that singing is for girls and boys who sing must be gay. Studies show that there is less gender stereotyping in single sex environments (Colley, Comber, & Hargreaves, 1994). Single sex classes may create s afe environments where boys who like to sing feel it is acceptable to participate in singing activities. This research employed a true experi mental design. Baseline data were collected on boys singing participation in coeducational classes, and then boy s were r andomly assigned to attend music in either a sin gle sex or coeducational class I collected more data on the boys singing participation at the end of the treatment period. The research hypotheses were:

PAGE 118

! *)' 1. Boys in single sex classes will participate d uring group singing activities more than boys in coeducational classes. 2. Boys in single sex classes will participate during group singing activities more at the end of the study than they did when they were in coeducational classes at the beginning of the study and they will participate more than the boys who remained in coeducational classes. I used a mixed method design to triangulate the data. I collected quantitative data in the form of videos that were scored using the Singing Participation Measure; I collected qualitative data in the form of music teacher journals entries and interviews. The teachers reported that the boys in the single sex classes typically participated in singing activities more than the boys in the coeducational classes. Each teacher also reported changes in individual boys participation in the all boys classes. The quantitative data did not corroborate these findings. I t is possible that the Singing Participation Measure was unable to capture the full spectrum of changes that the teachers observed. Further research is necessary to explore whether or not boys will participate during singing activities more in single sex c lasses than in coeducational classes. As one of the first studies on single sex education in the elementary general music classroom, this research just begins to scratch the surface of the issues surrounding boys willingness to sing in a school setting. T he findings show that many boys in the upper elementary grades are hesitant to participate in singing activities in coeducational classes Some of the boys appeared more comfortable and willing to sing in a single sex environment. Boys who feel it is safe and socially acceptable to sing are more likely to participate in singing activities and may develop of love of singing. This love of singing

PAGE 119

! *)( may influence whether or not boys continue with music once it becomes an elective course. Furthermore, if elementa ry music teachers consciously strive to cultivate a love of singing in boys while they are young these boys may grow up to be lifelong singers and challenge the stereotypes surrounding boys and singing.

PAGE 120

! **) References Adler, A. (2002). A case study of boys experiences of singing in school. (Doctoral dissertation, University of Toronto, Canada, 2002). Dissertations & Theses: Full Text Database. (Publication No. ATT NQ69210). Adler, A., & Harrison, S. (2004). Swinging back the gender pendulum: Addressing boys needs in music education research and practice. In L. Bartel (Ed.), Questioning the music education paradigm (pp. 270 289). Toronto: Canadian Music Educators Association. Bandura, A (1989). Social cognitive theory. In R. Vasta (Ed.), Annals of child development. Vol. 6. Six theories of child development (pp. 1 60). Greenwich, CT: JAI Press. Brand, E. (2003). Childrens beliefs about learning: Structures and strategies. Bulletin of the Council for Research in Music Education, 157, 9 17. Byo, J. (1991). An assessment of musical instrument references of third grade children. Bulletin of the Council for Research in Music Education, 110, 21 32. Carp, R. (2004). Single sex choral ensemble s, attitudes and practices: A survey of Southern California high school directors. (Doctoral dissertation, University of Southern California, 2004). (UMI No. 3145167)

PAGE 121

! *** Chido, P., Frakes, L., MacLeod, S., Pagel, R., Shuler, S., Thompson, J., & Watts, D. ( 1998). Making the grade authentic assessment in music, K 8. Teaching Music, 6(2), 34 36. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2 n d ed.) Hillsdale, N J : L. Erlbaum Associates Colley, A., Comber, C., & Hargreaves, D.J. (1994). School subject preferences of pupils in single sex and co educational secondary schools. Educational Studies 20 (3), 379 385. Cooper, N. A. (1995). Childrens singing accuracy as a function of grade level, gender, and individual versus union singing. Journal of Research in Music Education, 43 (3), 222 231. Damon, I. F. (1936). The boys who did not sing. Music Educators Journal, 23 (1), 41+43. Delzell, J. K., & Leppla D. A. (1992). Gender associations of musical instruments and preferences of fourth grade students for selected instruments. Journal of Research In Music Education, 40 (2), 93 103. Demorest, S. M. (2000). Encouraging male participation in chorus. Music Ed ucators Journal, 86 (4), 38 41. Duke, R. A. (1987). Observations of applied music instruction: The perceptions of trained and untrained observers. Applications of Research in Music Behavior, 115 124.

PAGE 122

! **+ Duke, R. A., Buckner, J. J., Cavitt, M. E., & Colprit E. (1997). Applications of SCRIBE: Systematic observation and analysis of teacher student interactions in music. Proceedings of the Fourth International Technological Directions in Music Education Conference. San Antonio: University of Texas. eJournal retrieved from http://music.utsa.edu/tdml/conf IV/IV Duke_et.al.html Eccles, J. S., Wigfield, Al, Harold, R. D., & Blumenfeld, P. (1993). Age and gender differences in childrens self and task perceptions during elementary school. Child Development, 6 4 (3), 830 847. Forsythe, J. L. (1977). Elementary student attending behavior as a function of classroom activities. Journal of Research in Music Education, 25(3), 228 239. Gates, J. T. (1989). A historical comparison of public singing by American men and women. Journal of Research in Music Education, 37 (1), 32 47. Gibb, S. J., Fergusson, D. M., & Horwood, L. J. (2008). Effects of single sex and coeducational schooling on the gender gap in educational achievement Australian Journal of Education, 52 (3), 301 317. Glass, G. V., & Hopkins, K. D. (1996). Statistical methods in education and psychology (3 r d ed.). Needham Heights MA : Allyn and Bacon, Inc Gregait, L. H., Johnsen, D. R., & Nielsen, P. S. (1997). Improving evaluation of student participation i n physical education through self assessment. Masters thesis, Saint Xavier University, Chicago. Haladyna, T. & Thomas, G. (1977, April). The attitudes of elementary school children toward school and subject matters. Paper presented at the annual meeting of the American Educational Research Association, New York, NY.

PAGE 123

! **# Hall, C. (2005). Gender and boys singing in early childhood. British Journal of Music Education 22 (1), 5 20. Hallam, S. (2004). Sex differences in the factors which predict musical attainment in school aged students. Bulletin of the Council for Research in Music Education, 161/162, 107 115. Harrison, J., Thompson, B., & Vannest, K. J. (2009). Interpreting the evidence for effective interventions to incr ease the academic performance of students with ADHD: Relevance of the statistical significance controversy. Review of Educational Research 79 (2), 740 775. Harrison, S. (2008). Masculinities and music: Engaging men and boys in making music. Newcastle upon Tyne: Cambridge Scholars Publishing. Hedden, S. K. (1982). Prediction of music achievement in the elementary school. Journal of Research in Music Education, 30 (1), 61 68. Hoffman, B. H., Badgett, B. A., & Parker, R. P. (2008). The effects of single sex instruction in a large, urban, at risk high school. The Journal of Educational Research. 102 (1), 15 35. Howle, M (1992). An examination of selected aspects of pitch matching problems among children. Research Perspectives in Music Education, 3, 22 27. Hub bard, L., & Datnow, A. (2005). Do single sex schools improve the education of low income ad minority students? An investigation of Californias public single sex academies. Anthropology and Education Quarterly, 36 (2), 115 131.

PAGE 124

! **$ Jellison, J. A. (2002) On task participation of typical students close to and away from classmates with disabilities in an elementary music classroom. Journal of Research in Music Education, 50(4), 343 355. Johnson, C., Darrow, A., Eason, B. J. A. (2008). Novice and skilled musi c teachers nonverbal behaviors and their relationship to perceived effectiveness and rapport. Bulletin of the Council for Research in Music Education, 178, 73 83. Keenan Takagi, K. (2000). Assessment in choral teaching. Music Educators Journal, 86(4), 42 46+63. Kiehn, M. (2003). Development of music creativity among elementary school students. Journal of Research in Music Education 51 (4), 278 288. Koza, J. E. (1993). The missing male and other gender issues in music education: Evidence from the Music Supervisors Journal, 1914 1924. Journal of Research in Music Education, 41 (3), 212 232. Lehmann, A. C., Sloboda, J. A., & Woody, R. H. (2007). Psychology for musicians. New York: Oxford University Press. Madsen, C. H., & Madse n, C. K. (1981). Teaching discipline: A positive approach for educational development. Boston: Allyn and Bacon, Inc. Madsen, C. K., & Geringer, J. M. (2000). A focus of attention model for meaningful listening. Bulletin of the Council for Research in Mus ic Education, 147, 103 108. Madsen, C. K., & Yarbrough, C. (1980). Competency based music education. Englewood Cliffs: Prentice Hall, Inc.

PAGE 125

! **" Mizener, C. P. (1993). Attitudes of children toward singing and choir participation and assessed singing skill. Journal of Research in Music Education, 41 (3), 233 245. Moore, R. S. (1987). Effects of age, sex, and activity on childrens attentiveness in elementary school music classes. In C K. Madsen, C. A. Prickett, & J. T. Gates (Eds.), Applications of Research in Music Behavior (pp. 26 31). Tuscaloosa: University of Alabama Press. Moore, R. S. (1994). Effects of age, sex, and melodic/harmonic patterns on vocal pitch matching skills of talented 8 11 year olds. Journal of Research in Music Education, 42 (2), 5 12. Music. Annotated bibliography of tests. (1988). Educational Testing Service, Princeton, N. J. Retrieved from ERIC (ED 366 608). National Association for Single Sex Public Education. (n.d.). Retrieved from http://www.singlesexschools.org/schools schools.htm Patton, M. Q. (2002). Qualitative research & evaluation methods (3 r d ed.). Thousand Oaks: Sage Publications. Pogonowski, L. M. (1985). Attitude assessment of upper elementary students in a process orie nted music curriculum. Journal of Research in Music Education, 33 (4), 247 257. Salomone, R. C. (2006). Single sex programs: Resolving the research conundrum. Teachers College Record, 108 (4), 778 802. Sax, L. (2005). Why gender m atters New York: Broadway Books.

PAGE 126

! **% Schmidt, C., & Sinor, J. (1986). An investigation of the relationships among music audiation, musical creativity, and cognitive style. Journal of Research in Music Education, 34 (3), 160 172. Sherban, R. A. (1995). Peering through the transp arencies of singing, gender and the music classroom. Published thesis, University of Calgary, Alberta, Canada. Sims, W. L. (1986). The effect of high versus low teacher affect and passive versus active student activity during music listening on preschool childrens attention, piece preference, time spent listening, and piece recognition. Journal of Research in Music Education, 34(3), 173 191. Soanes, C., & Stevenson, A. (Eds.). (2008). The Conci se Oxford English Dictionary (11 t h ed., revised). New York: Oxford University Press. Standley, J. M. (1992). Research note: Preschoolers responses to auditory and vibroacoustic stimuli. Psychology of Music, 20(1), 80 85 Stevens, J. P. (2007). Intermed iate statistics: A modern approach (3 r d ed.). New York: Lawrence Erlbaum Associates. Svengalis, J. N. (1978). Music attitude and the preadolescent male. (Doctoral dissertation, University of Iowa, 1978). Dissertation Abstracts International, 39 4800A 4801A. Swain, S. L., & Harvey, D. M. (2002). Single sex computer classes: An effective alternative. Tech Trends, 46 (6), 17 20. Taylor, S. (1969). Development of children aged seven to eleven. Journal of Research in Music Education, 17 (1), 100 107.

PAGE 127

! **& V iggiano, F. A. (1941). Reaching the adolescent who thinks its sissy to sing. Music Educators Journal, 27 (5), 62 63. Welch, G. F., Sergeant, D. C., & White, P.J. (1997). Age, sex, and vocal task as factors in singing in tune during the first years of sch ooling. Bulletin of the Council for Research in Music Education, 133, 153 160. Wilkinson, L., & Task Force on Statistical Inference. (1999). Statistical methods in psychology journals: Guidelines and explanations. American Psychologist 54 (8), 594 604. Wil son, S. J., & Wales, R. J. (1995). An exploration of childrens musical compositions. Journal of Research in Music Education, 43 (2), 94 111. Winslow, R. W. (1946). Male vocal problems in the secondary school. Music Educators Journal, 32 (4), 58+61. Wolfe, E. W., & Linden, K. W. (1991). Investigation of the relationship between intrinsic motivation and musical creativity (Tech. Rep. No. 143). Indiana: Purdue University. Wright, M. T., & Van Der Mars, H. (2004). Blending assessment into instruction: Practical applications and meaningful results. Journal of Physical Education, Recreation, and Dance, 75(9), 29 34. Yarbrough, C., & Price, H. E. (1981). Prediction of performer attentiveness based on rehearsal activity and teacher behavior. Journal of Re search in Music Education, 29(3), 209 217.

PAGE 128

! **' Younger, M. R., & Wa r rington, M. (2006). Would Harry and Hermione have done better in single sex classes? A review of sing sex teaching in coeducational secondary schools in the United Kingdom. American Educat ional Research Journal, 43 (4), 579 620.

PAGE 129

! **( Appendices

PAGE 130

! *+) Appendix A Music Teacher Reflective J ournals Date of journal entry____ _______ _____ Class__________ _____________________ In general, describe the participation of the boys during group singing activities during this class period. If any individual boy participated in singing activities differently today than he typically does, describe his behavior in detail. You may provide examples from multiple boys, if appropriate.

PAGE 131

! *+* Appendix B : Semi structured Interview Guide 1. The boys at your school typically attend music in coeducational classes. Tell me about the boys participation during group singing activities when they attend music in coeducational classes. 2. Du ring the course of this study, some of the boys at your school attended music in single sex classes. Des cribe the boys participation during group singing activities when there were no girls in the class. 3. How did the boys participation in the single sex class es compare to the boys participation in the coeducational class es ? 4. How did the individual boys participation compare when you consider how each boy participated in singing activities prior to the reorganization into new classes and after the boys were in their new classes ? 5. Can you give a specific example of a boy whose participation in singing activities in the single sex class was different than it was when he was in the coeducational class? 6. Can you share other examples as well? 7. What aspects of this study, if any, surprised you? 8. Describe any advantages and/or disadvantages you have seen when boys attend music classes as a single sex group? 9. In what ways, if any, did you alter your delivery of instruction when teaching the all boys groups? 10. How d id you feel about teaching single sex classes? 11. Is there anything else you would like to share with me? The researcher may also include questions about specific students or incidents reported by the music teacher in her journal entries.

PAGE 132

! ! ! ! ! ! ! About the Author Zadda M. Bazzy was born in Warren, Michigan She earned a Bachelor of Fine Arts degree in Musical Theatre from the Univer sity of Michigan and a Master of Arts degree in Music Education from the University of South Florida. She was certified in Early/ Middle Childhood Music by the National Board for Professional Teaching Standards and holds a Level III Orff certification. She receiv ed the Provosts Award for Outstanding Teaching by a Graduate Teaching Assistant in 2010 for teaching at the University of South Florida during her doctoral studies. She has taught elementary general music for 12 years including 6 years of teaching both single sex and coeducationa l classes. She frequently presents workshops for inservice and preservice music teachers.