Florida State University Libraries

Florida State University Libraries Electronic Theses, Treatises and Dissertations The Graduate School 2014 Factors Influencing Listener Preference for Music Therapists' Performances of Familiar Popular Songs Chelsea Chason Follow this and additional works at the FSU Digital Library. For more information, please contact lib-ir@fsu.edu

FLORIDA STATE UNIVERSITY COLLEGE OF MUSIC FACTORS INFLUENCING LISTENER PREFERENCE FOR MUSIC THERAPISTS PERFORMANCES OF FAMILIAR POPULAR SONGS By CHELSEA CHASON A Thesis submitted to the College of Music in partial fulfillment of the requirements for the degree of Master of Music Degree Awarded: Spring Semester, 2014

Chelsea Chason defended this thesis on April 17, 2014. The members of the supervisory committee were: John Geringer Professor Directing Thesis Dianne Gregory Committee Member Jayne Standley Committee Member The Graduate School has verified and approved the above-named committee members, and certifies that the thesis has been approved in accordance with university requirements. ii

ACKNOWLEDGMENTS I would like to thank my professors, Dr. Jayne Standley, Dr. Clifford Madsen, Dr. Alice- Ann Darrow, and Professor Dianne Gregory for their knowledge, leadership, inspiration, and encouragement. A special thanks goes to Dr. John Geringer who has overseen this research project, offering valuable guidance and support. Thank you also to Eitaro Kawaguchi, for his expertise in recording and editing the musical excerpts used in the study. A very special thank you to the four music therapists who allowed me record their renditions of popular songs for use in my survey thank you for lending your voices to this investigation. Thank you also to the students who volunteered to complete my survey, and their professors for allowing me to recruit research participants during their classes. Lastly, I would like to thank my family, my friends, music and music therapy colleagues, and coworkers for allowing me to vent and brainstorm while they listened, and for offering me all the encouragement and support I could ever need. iii

TABLE OF CONTENTS List of Tables... vi List of Figures... viii Abstract....ix 1. INTRODUCTION...1 Purpose of Study...1 2. REVIEW OF LITERATURE...3 Music Preferences...3 Influence of listener s background and experience...3 Preferences for popular music...4 Listener Discrimination and Music Preference...5 Patient Preferred Music in Music Therapy...6 Live versus Recorded Music...7 Music Competency in Music Therapy...8 Rationale for Study...10 Research Questions...10 3. METHOD...12 Recruitment of Participants...12 Survey...12 Song selection...13 Recording of music therapist performances...14 Procedure...14 4. RESULTS...16 Participant Data...16 Data Analysis for Research Question One...17 Data Analysis for Research Question Two...23 Data Analysis for Research Question Three...24 Data Analysis for Research Question Four...27 iv

Data Analysis for Research Question Five...28 5. DISCUSSION...30 Limitations of Study and Suggestions for Future Research...31 Differences from music therapy settings...31 Study design...32 Conclusions...33 APPENDICES A Text of Web-Based Survey...35 B Song Selection Survey and Results...41 D Performer Background Information...46 E Text Responses to Survey Questions...47 F Text Response Analysis Categories...51 G IRB Approval Memo and Participant Consent...52 REFERENCES...55 BIOGRAPHICAL SKETCH...60 v

LIST OF TABLES Table 1: Participants ages...16 Table 2: Frequency of participants responses to questions about music experiences...17 Table 3: Correlation between ratings of overall preference and music/performance elements for all performances of all songs, video and audio-only conditions combined....18 Table 4: Spearman Correlation between ratings of overall preference and music/ performance elements for the first performance of The Climb, video and audio-only conditions combined...19 Table 5: Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of The Climb, video and audio-only conditions combined...19 Table 6: Spearman Correlation between ratings of overall preference and music/performance elements for the first performance of Don t Stop Believing, video and audio-only conditions combined...20 Table 7: Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of Don t Stop Believing, video and audio-only conditions combined...20 Table 8: Spearman Correlation between ratings of overall preference and music/performance elements for the first performance of Firework, video and audio-only conditions combined...21 Table 9: Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of Firework, video and audio-only conditions combined...21 Table 10: Spearman Correlation between ratings of overall preference and music/performance elements for the first performance of Hey Jude, video and audio-only conditions combined...22 Table 11: Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of Hey Jude, video and audio-only conditions combined...22 Table 12: Means and standard deviations of overall preference for participants with differing levels of experience in private lessons...24 Table13: Means and standard deviations of overall preference for participants with differing levels of experience in music ensembles or groups...24 vi

Table 14: Means and standard deviations for overall preference for participants with differing levels of self-reported familiarity with alternate versions of a familiar popular song...25 Table 15: Means and standard deviations for overall preference for participants with differing levels of self-reported familiarity with the original recorded version of a familiar popular song... 26 Table 16: Means and standard deviations for overall preference for participants with differing levels of self-reported preference for the original recorded version of a familiar popular song...26 Table 17: Means and standard deviations for overall preference for participants who rated audioonly or audio-video performances of a familiar popular songs...27 Table 18: Correlation between ratings of overall preference and music/performance elements for all performances of all songs, audio only condition...27 Table 19: Correlation between ratings of overall preference and music/performance elements for all performances of all songs, video condition....28 Table 20: Songs chosen in song selection survey...42 vii

LIST OF FIGURES Figure 1: Frequency of participants answers to questions about music listening...17 Figure 2: Frequency with which different musical elements were found to have the highest three or lowest three correlation scores (r), across eight different performances...23 Figure 3: Participants responses to the question How familiar are you with the original recorded version of this song?...25 Figure 4: Participants responses to the question What do you think of the recorded version of this song?...26 Figure 5: Frequency of positive, negative, and neutral/unknown text responses, by category...29 viii

ABSTRACT The purpose of this study was to investigate factors that may influence listener preference for familiar popular songs as they might be performed live by music therapists. Participants, who were college undergrads (N = 86) enrolled in music appreciation courses, completed a web-based survey in which they rated their overall preference for video or audio excerpts of music therapists performing familiar popular songs, as well as their perceptions of performers musical skills and ability to match the original artists recordings of the same songs. Listeners responses were analyzed to find correlations between ratings of overall preference and other performance elements. Results indicate that the strongest correlations with preference were found for ratings of performers overall musical skill, singing skill, and musical expressiveness, while the lowest correlations were found for match of lyrics, match of rhythmic style, and match of accompaniment style. These results suggested that in preparing live music for use in music therapy settings, therapists may be more effective by focusing on musical skills, especially singing and expressiveness, rather than seeking to emulate rhythmic or stylistic elements from the original recording of a song. ix

CHAPTER ONE INTRODUCTION Music therapists use patient preferred live music with patients or clients for a variety of reasons: to build rapport with clients, to reduce pain, stress, and anxiety, to increase relaxation, to introduce and facilitate discussions of counseling topics, to structure interventions, and to motivate clients to progress towards a wide variety of clinical outcomes. Research and clinical training support that live performance of a patient or client s preferred music is often most effective in music therapy interventions. However, with the staggering amount of music available to listeners today, the likelihood is quite small that a music therapist will know every client s favorite songs well enough to perform live. Music therapists focus on learning a wide variety of music from various genres and eras with the goal of having at least one song that could meet the preferences of a given client in a moment s notice. Even if the therapist knows a favorite song, the client most likely knows and prefers a familiar recording of the song. Popular music, the preferred music of many music therapy clients, is almost always heard in a recorded format. Listeners are accustomed to the recorded versions of the songs, which contain many musical elements that a solo performer cannot easily reproduce (for example, multi-instrumental arrangements, vocal harmonies, or electronic backing tracks). Even in a song with a simpler arrangement, listeners are accustomed to hearing a favorite artist s distinct instrumental and vocal inflections and timbres. Are listeners likely to overlook these differences when listening to a music therapist s unfamiliar rendition of a familiar favorite song? Purpose of Study The purpose of this study is to explore the way several factors influence listeners stated preference for popular songs as they might be performed by music therapists in music therapy 1

settings. The study will investigate how listener preference for a music therapist s performance is affected by the listener s perceptions of the performer s skill, sincerity, and how well the musical elements of the performance match the familiar popular recording of the same song. The study will also take into account participants musical backgrounds and familiarity with popular recorded music. The main objective of the study is to determine what aspects of a music therapist s performance are most important in determining whether or not the listener likes the performance, since listener-preferred music is more effective in music therapy interventions. 2

CHAPTER TWO REVIEW OF LITERATURE Music Preferences For the purposes of this study, preference is the act of choosing, esteeming, or giving advantage to one thing over another-- that is, having propensity toward something (Price, 1986). At the most basic level, music preference is a question of whether a listener likes or dislikes a performance or piece of music at any given point in time. Musical preference is determined by a complex interaction of multiple factors. LeBlanc s (1982) model of music preference outlined these factors in an 8-level hierarchy that considered properties of the musical stimuli, social and cultural influences, and the listener s personal experiences, perceptions, and cognitive process in determining listener preference for a single musical experience. Berlyne (1971) and Walker (1980) examined the relationship between music complexity and listener preference, suggesting that the relationship can be represented as an inverted U-shaped curve, where each listener has an optimal level of complexity. Further research has explored the influence of other factors on music preference, including individual personality traits (Rentfrow & Gosling, 2003; Kopacz, 2005), the influence of emotion and meaningfulness (Schubert, 2007; Craig, 2009), and the relationship to how well the music fulfills the needs of the listener in response to various possible functions of music (Schafer & Sedlmeier, 2009, 2010). Influence of listener s background and experience Individual differences in listeners backgrounds, including levels of musical training and experience were found to have an effect on musical preferences. Factors including age, home country, and gender were found to have effects on listening preferences, though the interaction of multiple factors were too complex to make any useful generalizations (LeBlanc, Jin, Stamou, & 3

McCrary, 1999). No substantial differences were found in aesthetic responsiveness between musicians and non-musicians (Madsen, Byrnes, Capperella-Sheldon, & Brittin, 1993). However, individuals with musical training were found to have greater preferences for non-popular styles of music, including jazz and western art music (Gregory, 1994; Ginocchio, 2009). Preferences for popular music Although much of the research on music preference focuses on western art music, one need only to peruse current Billboard charts to realize that popular music is preferred by the majority of listeners as a sort of tautology, popular music is the most popular music. In a survey of over two thousand subjects across all age groups, listeners had a higher preference for rock music over jazz and western art music (LeBlanc, Sims, Siivola, & Obert, 1996). Holbrook and Schindler reported that most listeners over age 24 had developed a taste for pop music (1989). North and Hargreaves found that popular songs were rated higher for preference and for artistic merit when compared to critically acclaimed art pop songs (1998). As to questions of why listeners prefer pop music, Boyle, Hosterman, and Ramsey (1981) had young listeners rate their reasons for liking their favorite pop songs. In order of most relevant to least relevant factors in determining preference, respondents cited: 1) melody 2) mood 3) rhythm 4) lyrics 5) instruments 6) sentiment 7) harmony 8) singer or group 9) radio exposure 10) dance-ability and 11) peer influence. Other research (Alpert, 1982; Webster & Hamilton, 1981/1982) suggested that peer influence may have a stronger effect than listeners may realize or admit, especially for young people. Familiarity could also be a large part of the reason that pop music is most preferred, as popular music is ubiquitous in western culture. Research has shown that repetition and familiarity are very important factors in determining a listener s preference (Bradley, 1971; Getz, 4

1996; Peretz, Gaudreau, & Bonnel, 1998). In studies of preferences for world music, familiarity was found to increase listener preferences in both children (Demorest & Schultz, 2004) and college-aged students (Fung, 1996). In fact, Demorest and Schultz found that familiarity, both of specific pieces and of western style arrangements of world music songs, may have influenced children s preferences for arranged versions of songs over more stylistically authentic versions. Listener Discrimination and Music Preference As discussed above, music preferences are determined by extra-musical attributes, including listeners social, cultural, or musical background and life experiences. These factors are in play before the listener even hears the musical stimuli in question. However, properties of the musical stimuli are also important in determining preference. The research literature reflects that listeners are able to perceive and discriminate differences in various musical elements, and that this discrimination in musical perception can influence preferences. In light of the above findings that familiar music is usually most preferred, it is important to note that listeners with and without musical training are surprisingly adept at recognizing deviations from the expected in familiar music stimuli. Both musicians and non-musicians were able to detect deviations in tempo and pitch in recorded versions of familiar songs in orchestral and popular styles (Geringer & Madsen, 1984, 1987). It was found that subjects with and without musical backgrounds seemed to have long-term memory of absolute pitch and tempo for familiar recorded pop songs (Levitin, 1994; Levitin & Cook, 1996). These studies demonstrate that listeners often come to associate preferred popular songs with a definitive familiar recording, and they have shown marked preferences for the original un-altered versions of familiar songs (Geringer & Madsen, 1987). 5

In addition to recognizing deviations from what is expected in familiar music, listeners have been found to discriminate based on other aspects of musical stimuli. Researchers have investigated listeners abilities to discriminate based on tone quality and intonation (Geringer & Madsen, 1981; Webster & Hamilton, 1981/1982), vocal vibrato, (Leblanc & Sherrill, 1986), tempo and accompaniment style (Brittin, 2000). Listener preference may also be influenced by non-musical elements of a performance that are not acoustical properties of the musical stimulus. Research has found that a performer s gender (LeBlanc & Sherrill, 1986; Millar, 2008), race (McCrary, 1993), physical attractiveness, dress, and stage behavior (Wapnick, Mazza, & Darrow, 1998), or even perceived humor in song lyrics (LeBlanc, Sims, Malin, & Sherrill, 1992) can have effects on listeners evaluations of performances and ratings of overall preference. Patient Preferred Music in Music Therapy In Standley s 2000 meta-analysis of music research in medical treatment settings, studies using patient preferred live music demonstrated the greatest effect size (ES=1.40, d=30), while those using live music presented by a trained music therapist also showed a reasonably large effect size (ES= 1.13, n=16). The largest effects were shown for improvement in grasp strength following stroke, perceived effectiveness of music, EMG, self-reported reduction in pain, relaxation, and reduction of anxiety. Pelletier s 2004 meta-analysis supported the use of music with various relaxation techniques to reduce arousal due to stress and anxiety. In this analysis, preferred relaxing music selected by patients showed a smaller effect size (ES=.51) than use of music selected by therapist or researcher (ES=.83), and subjects with music experience showed the greatest reductions in stress overall. In a more recent analysis of 50 studies dealing with the neural effects of music on emotional regulation, Moore (2013) found that desired neural 6

activation (regulation of emotion) occurred with the use of preferred and familiar music, with singing, and when musicians engaged in improvisation, while undesired activation (stress and arousal) occurred when complexity, dissonance, and unexpected musical events were introduced. A small sample of recent studies using music in medical and psychiatric settings confirmed that patient preferred music was effective in reducing emotional distress during radiation therapy (Clark et al., 2006), reducing reported levels of pain following open heart surgery (Jafari, Zeydi, Khani, Esmaeili, & Soleimani, 2012), and reducing time needed to complete MRI scans while improving patient perception of the MRI experience (Walworth, 2010). Listening to patient preferred heavy metal music was found to increase the positive affect of adolescents in an inpatient psychiatric setting, showing that even music that is not commonly thought to be soothing can have positive effects when it is preferred by listeners (Wooten, 1992). Preferred music was found to reduce experimentally induced pain (Mitchell & MacDonald, 2006; Perlini & Viita, 1996) and anxiety (Walworth, 2003; Jiang, Zhou, Rickson, & Jiang, 2013). Interestingly, Walworth (2003) found that songs selected by the researcher within the subject s preferred genre and specific songs selected by subjects were both effective in reducing anxiety, with no significant difference in effects for music selected by the researcher or the research subject. Live verses Recorded Music Comparison of the effects of live and recorded music have shown live music to be somewhat more effective in achieving various desired clinical outcomes. In patients undergoing MRI scans, the music therapist s live performance of preferred music was more effective than recorded music performed by the original artists (Walworth, 2010). Harmon (2011) reported inconclusive results in comparisons between the effects of live music, recordings of therapist, or 7

recordings of original artists on participation and behavior states of adults with dementia, though levels of participation (hand and foot tapping) were slightly higher during both live performance and recordings by the original artist. When compared to recorded music, live music was found to be more effective in eliciting alert behavior states in non-responsive hospice patients (Segall, 2007). Although it is acknowledged that it is important to use patient preferred music in therapy, outside of song selection there is little discussion of how music elements may influence patient preference of live performances. Groene (2001) compared the effects of simple vs. complex guitar accompaniment and live vs. recorded performances of familiar songs on the attentional and responsive behaviors of older adults with dementia, finding that patients appeared to pay more attention to live performances with complex accompaniment. In his discussion of the results of a survey of music therapists using lyric analysis interventions in mental health settings, Silverman (2009) suggested that although it is suggested that music therapists use live music whenever possible, it has been the author s experience that some songs do not translate well to a live performance on solo acoustic guitar ( p. 53). According to Standley & Jones, (2007) In order for the client s objective to be successfully reached, the music therapist must have selected a relevant song and recreated the song in a musically sophisticated manner (p. 223). Music Competency in Music Therapy The American Music Therapy Association (AMTA) is an organization dedicated to the support of the music therapy profession through the advancement of music therapy research, education, training, professional standards, and credentials. The AMTA has listed the following basic music competencies for music therapists: the abilities to lead and accompany proficiently on instruments including, but not limited to, voice, piano, guitar, and percussion, to play basic 8

chord progressions in several major and minor keys with varied accompaniment patterns, to play and sing a basic repertoire of traditional, folk, and popular songs with and without printed music, and to sing in tune with a pleasing quality and adequate volume both with accompaniment and a cappella (American Music Therapy Association, 2014). The Certification Board for Music Therapists has made very little mention of music skills in the Scope of Practice, stating only that music therapists must employ functional skills with voice, keyboard, guitar, and percussion, and suggesting that professional development should include the expansion of music skills (Certification Board for Music Therapists, 2014). Competency assessments for music therapy students specify that music therapy performances should be free of wrong notes and unpleasant sounds, vocals should be in tune with a pleasing quality, and guitar accompaniment should be smooth and maintain a steady beat (Standley & Jones, 2007). Gregory and Belgrave (2006) identified basic requisite guitar accompaniment skills for music therapists and a method for evaluating the development of these skills. No mention was made of attempts to emulate stylistic elements of the original/popular recording of a song, though another proposed evaluation instrument, the Guitar Songleading Performance Scale, did rate expressive quality and overall aesthetic affect in a performance (Silverman, 2011). In summary, live patient-preferred music is shown by the research to be effective in music therapy interventions, especially those to reduce pain and anxiety, or increase positive affect. Music therapists are expected to have the music competencies to sing and play guitar accompaniment for live versions of a wide variety of patient preferred songs in an aesthetically pleasing manner. 9

Rationale for Study This review of literature demonstrates that even though research has shown that listener preference for music is based on listeners abilities to perceive and discriminate based on differences in musical stimuli, there is a lack of research into how music therapy patients and potential patients evaluate and respond to musical elements of music therapists live renditions of familiar popular songs. The purpose of this study is to explore the way several factors influence listeners stated preference for popular songs as they might be performed by music therapists in music therapy settings. The study will investigate how listener preference for a music therapist s performance is affected by the listener s perceptions of the performer s musical skills, sincerity, and confidence, as well as how closely the musical elements of the performance match the familiar popular recording of the same song. The study will also take into account participants musical backgrounds and familiarity with popular recorded music. The main objective of the study is to determine what aspects of a music therapist s performance are most important in determining whether or not the listener likes the performance, since listener-preferred music is more effective in music therapy interventions. Research Questions This study will attempt to answer the following research questions about how listeners and potential music therapy patients respond to a music therapist s performance of a popular song as it might be performed in a music therapy session: 1) Which rated performance elements have the strongest relationship to listeners ratings of overall preference for music therapists performances of familiar popular songs? 2) Will listeners musical backgrounds and music experience levels have a significant effect on their rated preferences for familiar popular songs performed by music therapists? 10

3) Will listeners familiarity with a recorded version or alternate renditions of a familiar song have a significant influence on their rated preferences for music therapist performances of the same song? Will listeners familiarity with alternate renditions of a familiar popular song have a significant effect on their rated preference of music therapists performances of the same song? 4) Will listeners who see videos of music therapists performing familiar popular songs have significantly different ratings of overall preference from those who hear an audio-only version of the same performance? 5) What do participants like most and/or least about music therapists renditions of familiar popular songs? 11

CHAPTER THREE METHOD Recruitment of Participants Participants were undergraduate students at Florida State University who were enrolled in music appreciation courses for non-music majors (two sections of Music in Popular Culture, and one section of American Roots Music ). Prospective participants were approached by the researcher during regular class meeting time, with the permission of the class instructors. One hundred sixty-two students submitted an email address in order to receive a hyperlink to the web-based survey. Of these, 91 began the survey, and 86 submitted usable responses. Criteria for inclusion stipulated that participants were at least 18 years of age and enrolled in undergraduate courses at Florida State University. Survey responses were included in the study only when participants agreed to consent to participate, completed survey questions about personal and musical background, and completed questions for at least one performance. Partial responses were included for respondents who left some questions blank, unless respondents failed to answer questions about overall preference for the performance. Survey A web-based survey (see Appendix A) was created by the researcher using the website Qualtrics.com. Participants were asked to complete the web-based survey on a personal computer equipped with personal earphones. The survey prompted each participant to choose their favorite/most familiar song from a list of four popular songs. The survey contained an embedded video or audio excerpt of two performances of two different music therapists performing the song selected by the participant. The participants completed a questionnaire with 13 seven-point Likert-type scales to rate the following: 1) their own preference for each of the 12

performances they watched, 2) their perception of the performer s musical skills (overall musical skill, singing skill, sound quality of voice, guitar skill, and musical expressiveness), confidence, and sincerity, and 3) how closely musical elements ( rhythmic style, singing style, musical accompaniment style, lyrics, and overall mood) of the performance matched their memory of a familiar recording of the same song. Optional text entry fields allowed participants to make comments on what they liked most and/or least about each performance, and comments on how they thought the performance did or did not match the familiar/original recording of the same song. Participants also answered questions about their level of familiarity with recorded versions of the song selection, and a brief survey on their own musical background and music listening habits. Song selection The songs used in the survey were selected based on an informal survey of music therapists about which songs they use most often in music therapy sessions with young adult clients (see Appendix B for survey questions and anonymous responses). This informal webbased survey was created using Qualtrics.com, and distributed through the researcher s online social network, and through a national music therapy listserv, a LISTPROC mailing list hosted on a Unix server at the University of Kansas. The most commonly listed songs were selected for possible inclusion in this study. This list was modified to accommodate the song repertoire of the music therapists who agreed to participate in the creation of recordings for use in the study. The songs selected for use in the study were: The Climb by Miley Cyrus, Don t Stop Believing by Journey, Firework by Katy Perry, and Hey Jude by The Beatles. 13

Recording of music therapist performances Music therapists who were current or former post-graduate students at Florida State University were recruited to perform in recordings to be embedded in the web-based survey. A total of four music therapist performers were used (see Appendix C for music and music therapy backgrounds of performers). Performers were assigned to play songs with which they were already familiar. The researcher requested that the performers not focus on creating a perfect and polished rendition of the song, rather that they perform the song as they would play for a client. Two different recordings were made of each of the four songs included in the study, using different performers on each of the two renditions. The songs were recorded in a medium sized recital hall in the college of music, with only the researcher, a recording technician, and a faculty adviser present in addition to each individual performer. A Zoom H2 digital audio recorder was used to record audio, and a Sony DCR-SR200 video camera was used to record the video. The audio and video were edited together, and a black screen replaced the video of the performance for the audio-only condition. The exact same audio was used in both conditions. Excerpts were uploaded to a private channel on youtube.com, and then embedded in the web-based survey. Procedure The researcher obtained permission from the University s IRB to recruit students as research subjects (see Appendix F). Participants were approached by the researcher during regular class meeting times, with the permission of the class professors. The researcher spoke to 3 separate classes in order to recruit participants for the web-based survey. The researcher provided a brief verbal description of music therapy and a description of the study, and then collected the email addresses of students who volunteered to participate in the study. An email with a link to the web-based survey was sent all participants (N=162). The email included a brief 14

description of the study, instructions and conditions for completing the survey, contact information for the principal investigator, faculty adviser, and Florida State University s Human Subjects Committee, and a hyperlink to the survey. The email advised that participation was voluntary, there would be no compensation for participation, and there would be no penalty for withdrawing participation. It was stated that consent was implied upon completion of the survey. The survey was open for four weeks, and during this time participants received two reminder emails containing the hyperlink to the survey. Participants were randomly assigned to either the audio-only, or the audio-video condition, and were sent the hyperlink to the corresponding survey. The surveys were identical with the exception that the audio-video condition displayed a video of the performance, while the audio-only condition used the same audio, but displayed a black screen. Participants were prompted to select one of the four available songs. They were asked to choose the song that they liked best, or in the event that they did not particularly like any of the songs, to choose the song that was most familiar. The web-based surveys were configured to display the song choices in random order, as well as to randomize the order in which the two performances were displayed, and randomize the order of questions following each performance. 15

CHAPTER FOUR RESULTS Participant Data Participants were students enrolled in music appreciation courses (2 sections of Music in Popular Culture, and 1 section of American Roots Music) for undergraduate non-music majors at Florida State University. Of the participants who returned a usable web-based survey (N=86), 50 were female and 36 were male (58% female, 42% male). Participants ages ranged from 18-32 years (M=19.1 years), with 85% between the ages of 18 and 20 (see Table 1) Table 1 Participants ages. Age Frequency Percentage 18 41 48% 19 21 24% 20 11 13% 21 9 10% 22 1 1% 26 1 1% 32 1 1% No Response 1 1% Most participants were not music majors (n=81, 94% non-music majors vs. n=5, 6% music majors), though participants had varying levels of music experience in private lessons or participation in music ensembles as illustrated in Table 2. A majority of participants (64%) responded to a question asking them to list their primary musical instrument or voice, while 36% did not indicate whether they played an instrument or sang. A large majority of participants reported that they often choose to listen to music in their daily lives, and that they often have a song or a part of a song stuck in their heads (Figure 1). 16

Most participants (79%) reported that they consider music to be a very important part of their lives, with 20% reporting that music is somewhat important and only 1% reporting that music is not an important part of daily life. Table 2 Frequency of participants responses to questions about music experiences. Level of Experience Music Lessons (% of Total) Experience in Music Ensembles (% of Total) Never 21 (24%) 24 (28%) Less Than 1 Year 13 (15%) 8 (9%) 1-2 Years 15 (18%) 10 (12%) 3-5 Years 15 (18%) 18 (21%) 6-10 Years 13 (15%) 18 (21%) 10+ Years 9 (10%) 8 (9%) Very Often 44 64 Quite Often Sometimes Rarely 2 11 20 31 How often do you have a song "stuck in your head"? How often do you choose to listen to music? Never 0 10 20 30 40 50 60 70 Figure 1: Frequency of participants answers to questions about music listening. Data Analyses for Research Question One Which performance elements have the strongest relationship to listeners ratings of overall preference for music therapists performances of familiar popular songs? A composite of the correlations between overall preference ratings and ratings of performance qualities for all performances of all songs combined can be found in Table 3. 17

Table 3 Correlation between ratings of overall preference and music/performance elements for all performances of all songs, video and audio-only conditions combined Ratings r t p (2-tail) M SD n Overall Preference 3.22 1.46 171 Musical Skill*.64 10.87 <.001 3.88 1.32 170 Singing Skill*.60 9.84 <.001 3.16 1.52 170 Sound Quality of Voice*.47 6.93 <.001 3.49 1.49 170 Guitar Skill*.51 7.70 <.001 4.16 1.34 171 Expressiveness*.54 8.22 <.001 3.72 1.50 170 Sincerity*.49 7.21 <.001 3.81 1.43 171 Confidence*.37 5.19 <.001 4.11 1.51 170 Match of Rhythmic Style.21 2.73 =.001 3.81 1.58 171 Match of Singing Style*.50 7.57 <.001 2.99 1.47 170 Match of Accomp. Style*.33 4.60 <.001 3.33 1.41 171 Match of Lyrics.06.72 >.45 5.86 1.32 170 Match of Mood*.51 7.79 <.001 3.18 1.54 170 Spearman Correlations between participants ratings of overall preference and ratings of each of 12 performance qualities for all eight performances of the four songs can be found in Tables 4-11. There was some agreement about which performance elements had the strongest and weakest correlations with overall preference across the different performances. A composite of the 3 highest and 3 lowest correlations for each individual performance can be found in Figure 2. Across the eight separate performances that were rated by participants, the strongest correlations were found most frequently between overall preference and performer s skill as singer, overall musical skill, and expressiveness. The weakest correlations were found most frequently between ratings of overall preference and match of lyrics, match of rhythmic style, and match of accompaniment style. In general it appears that evaluations of the performer s musical skills, and especially vocal skill, were more strongly related to ratings of overall 18

preference than were evaluations of how well the performance matched the familiar original recording of the same song. Table 4 Spearman Correlation between ratings of overall preference and music/performance elements for the first performance of The Climb, video and audio-only conditions combined. Ratings r t p (2-tail) M SD n Overall Preference 2.91 1.30 11 Musical Skill.39 1.25.24 3.36 1.03 11 Singing Skill.50 1.74.12 3.00.89 11 Sound Quality of Voice*.90 6.15 <.001 3.27 1.10 11 Guitar Skill -.04 -.13.90 3.18 1.08 11 Expressiveness.34 1.07.31 2.91 1.38 11 Sincerity.73 3.18.01 3.09 1.30 11 Confidence.23.71.50 2.73 1.28 11 Match of Rhythmic Style.08.24.82 3.82 2.14 11 Match of Singing Style.05.16.88 3.64 1.86 11 Match of Accomp. Style.25.76.47 3.18 1.33 11 Match of Lyrics.43 1.43.19 5.91 1.81 11 Match of Mood.28.84.43 3.00 1.63 10 Table 5 Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of The Climb, video and audio-only conditions combined. Ratings r t p (2-tail) M SD n Overall Preference 2.55.93 11 Musical Skill.57 2.07.07 3.55 1.13 11 Singing Skill.40 1.30.23 3.00 1.27 11 Sound Quality of Voice.40 1.29.23 3.27 1.79 11 Guitar Skill.03.09.93 4.46 1.29 11 Expressiveness.28.88.40 3.18 1.25 11 Sincerity.17.53.61 3.46 1.21 11 Confidence -.06 -.20.85 4.09 1.45 11 Match of Rhythmic Style -.05 -.15.88 3.36 1.75 11 Match of Singing Style.40 1.31.22 2.64 1.75 11 Match of Accomp. Style -.24 -.75.47 3.36.92 11 Match of Lyrics 0.01.99 5.36 1.57 11 Match of Mood.05.16.88 3.18 1.40 11 19

Table 6 Spearman Correlation between ratings of overall preference and music/performance elements for the first performance of Don t Stop Believing, video and audio-only conditions combined. Ratings r t p (2-tail) M SD n Overall Preference 2.56 1.53 27 Musical Skill.54 3.18.004 4.22 1.58 27 Singing Skill.43 2.38.03 3.85 1.98 27 Sound Quality of Voice.41 2.27.03 4.00 1.69 27 Guitar Skill.46 2.59.02 4.26 1.23 27 Expressiveness.48 2.74.01 4.41 1.39 27 Sincerity.33 1.76.09 4.04 1.32 27 Confidence.26 1.36.19 4.93 1.33 27 Match of Rhythmic Style.30 1.58.13 3.26 1.48 27 Match of Singing Style.59 3.70.001 2.33 1.36 27 Match of Accomp. Style.41 2.22.04 2.78 1.40 27 Match of Lyrics -.13 -.65.52 5.96 1.13 27 Match of Mood*.62 3.92 <.001 2.56 1.37 27 Table 7 Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of Don t Stop Believing, video and audio-only conditions combined. Ratings r t p (2-tail) M SD n Overall Preference 3.67 1.36 27 Musical Skill*.73 5.19 <.001 3.89 1.21 26 Singing Skill*.75 5.59 <.001 3.35 1.20 26 Sound Quality of Voice.57 3.47.002 3.33 1.24 27 Guitar Skill*.72 5.15 <.001 4.37 1.04 27 Expressiveness.57 3.45.002 3.44 1.58 27 Sincerity*.66 4.44 <.001 4.00 1.14 27 Confidence*.61 3.80 <.001 4.19 1.30 27 Match of Rhythmic Style.47 2.67.013 4.22 1.31 27 Match of Singing Style.35 1.88.072 3.22 1.40 27 Match of Accomp. Style.44 2.45.022 3.56 1.40 27 Match of Lyrics.33 1.73.096 6.12.99 26 Match of Mood*.64 4.12 <.001 3.30 1.49 27 20

Table 8 Spearman Correlation between ratings of overall preference and music/performance elements for the first performance of Firework, video and audio-only conditions combined. Ratings r t p (2-tail) M SD n Overall Preference 3.50 1.51 16 Musical Skill*.79 4.87 <.001 4.38 1.26 16 Singing Skill*.83 5.62 <.001 3.13 1.50 16 Sound Quality of Voice*.82 5.45 <.001 3.81 1.60 16 Guitar Skill*.78 4.68 <.001 4.94 1.65 16 Expressiveness*.82 5.16 <.001 3.73 1.49 15 Sincerity.69 3.54.003 3.75 1.44 16 Confidence*.84 5.71 <.001 4.00 1.37 16 Match of Rhythmic Style.31 1.22.24 3.88 1.63 16 Match of Singing Style.70 3.69.002 2.94 1.57 16 Match of Accomp. Style.28 1.07.30 3.25 1.44 16 Match of Lyrics.01.04.97 5.94 1.44 16 Match of Mood.67 3.33.005 3.13 1.36 16 Table 9 Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of Firework, video and audio-only conditions combined. Ratings r t p (2-tail) M SD N Overall Preference 3.06 1.57 16 Musical Skill.70 3.65.003 4.00 1.37 16 Singing Skill.66 3.30.005 2.63 1.15 16 Sound Quality of Voice.47 1.97.07 3.19 1.28 16 Guitar Skill.51 2.20.05 4.88 1.54 16 Expressiveness.62 2.98.01 4.25 1.48 16 Sincerity.47 1.98.68 4.19 1.38 16 Confidence.54 2.43.03 4.81 1.68 16 Match of Rhythmic Style.27 1.03.32 4.88 1.54 16 Match of Singing Style.69 3.59.003 3.13 1.31 16 Match of Accomp. Style.15.56.58 4.06 1.53 16 Match of Lyrics -.02 -.07.95 5.94 1.39 16 Match of Mood.35 1.42.18 4.19 1.87 16 21

Table 10 Spearman Correlation between ratings of overall preference and music/performance elements for the first performance of Hey Jude, video and audio-only conditions combined. Ratings r t p (2-tail) M SD N Overall Preference 2.74 1.29 31 Musical Skill*.70 5.25 <.001 3.13 1.12 31 Singing Skill*.67 4.87 <.001 2.29 1.35 31 Sound Quality of Voice.39 2.27.03 2.77 1.38 30 Guitar Skill*.65 4.66 <.001 3.29 1.04 31 Expressiveness*.62 4.24 <.001 2.97 1.28 31 Sincerity.21 1.18.25 3.29 1.79 31 Confidence.26 1.43.16 3.55 1.52 31 Match of Rhythmic Style.14.77.45 4.13 1.28 31 Match of Singing Style.41 2.45.02 2.68 1.38 31 Match of Accomp. Style.36 2.05.05 3.26 1.29 31 Match of Lyrics -.23-1.27.21 5.81 1.35 31 Match of Mood.52 3.28.003 3.00 1.39 31 Table 11 Spearman Correlation between ratings of overall preference and music/performance elements for the second performance of Hey Jude, video and audio-only conditions combined. Ratings r t p (2-tail) M SD N Overall Preference 4.09 1.30 32 Musical Skill*.78 6.79 <.001 4.31 1.23 32 Singing Skill*.81 7.55 <.001 3.66 1.49 32 Sound Quality of Voice*.56 3.73 <.001 4.00 1.44 32 Guitar Skill*.64 4.57 <.001 4.22 1.24 32 Expressiveness*.68 5.03 <.001 4.28 1.37 32 Sincerity.54 3.47.002 4.16 1.37 32 Confidence*.62 4.29 <.001 4.10 1.40 31 Match of Rhythmic Style.31 1.81.08 3.22 1.54 32 Match of Singing Style.46 2.81.009 3.55 1.29 31 Match of Accomp. Style.34 1.95.06 3.41 1.56 32 Match of Lyrics.06.31.76 5.72 1.37 32 Match of Mood*.57 3.76 <.001 3.38 1.66 32 22

0 2 4 6 8 Musical skill Skill as singer Sound quality of voice Skill as guitarist Musical expressiveness Sincerity Confidence Highest Correlation to Preference Lowest Correlation to Preference Match of rhythmic style Match of singing style Match of accompaniment style Match of lyrics Match of overall mood Figure 2: Frequency with which different musical elements were found to have the highest three or lowest three correlation scores (r), across eight different performances. Data Analysis for Research Question Two Will listeners musical backgrounds and music experience levels have a significant effect on their rated preferences for familiar popular songs performed by music therapists? Data were analyzed using a One-Way Analysis of Variance (ANOVA) to determine if participants with different levels of music experience had significantly different ratings of overall preference for the performances they heard in this study. ANOVA results showed no significant difference in ratings of preference between participants with differing levels of experience in private music lessons; F (4, 166) = 2.12, p =.08, or differing levels of experience in music ensembles or groups; F (4, 166) = 0.2, p =.94. Means and standard deviations for ratings of overall preference for differing levels of music experience can be found in Tables 12 and 13. 23

Table 12 Means and standard deviations of overall preference for participants with differing levels of experience in private lessons. Number of Years in Private Lessons n M SD None 42 3.67 1.36 0-2 Years 55 3.20 1.52 3-5 Years 30 2.87 1.60 6-10 Years 26 3.27 1.25 Over 10 Years 18 2.67 1.46 Table 13 Means and standard deviations of overall preference for participants with differing levels of experience in music ensembles or groups. Number of Years in Music Ensembles or Groups n M SD None 48 3.33 1.55 0-2 Years 35 3.17 1.29 3-5 Years 36 3.25 1.48 6-10 Years 36 3.14 1.75 Over 10 Years 16 3.00 1.75 Data Analysis for Research Question Three Will listeners familiarity with alternate renditions of a familiar popular song have a significant effect on their rated preference of music therapists performances of the same song? A One-Way Analysis of Variance revealed that listeners self-reported familiarity with alternate renditions of the familiar popular songs (see Table 14) had a significant effect on their overall preference for the performance; F (3, 167) = 2.98, p =.03. Those who reported familiarity with only one alternate rendition rated higher overall preference (M= 3.64) than those who reported familiarity with four or more renditions (M= 2.61). Analyses of variance were not significant for different levels of familiarity with the original recording of the song, F (3, 167) = 24

.45, p =.72, or for different levels of preference for the familiar original recorded version of the song, F (3, 167) = 1.15, p =.33. Participants reported levels of familiarity and preference for the original recorded version of the song can be found in Figures 3 and 4. Means and standard deviations for overall preference for listeners with differing levels of familiarity and preference for the original recorded version of the song can be found in Tables 15 and 16. Table 14 Means and standard deviations for overall preference for participants with differing levels of self-reported familiarity with alternate versions of a familiar popular song. Number of Alternate Song Versions Known by Participants n M SD None 41 3.37 1.51 One other version 46 3.61 1.39 2-3 other versions 66 3.00 1.39 4 or more other versions 18 2.61 1.58 50.00% 45.00% 40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% 1.16% 0.00% 1- Not at All 5.81% 5.81% 16.28% 25.58% 45.35% 2 3 4 5 6 7- Very Much Percentage of Responses Figure 3: Participants response to the question How familiar are you with the original recorded version of this song? 25

60.00% 55.81% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 20.93% 8.14% 4.65% 6.98% 1.16% 1.16% 1-Dislike 2 3 4 5 6 7- Like Percentage of Responses Figure 4: Participants response to the question What do you think of the recorded version of this song? Table 15 Means and standard deviations for overall preference for participants with differing levels of self-reported familiarity with the original recorded version of a familiar popular song. Likert-Scale rating of Familiarity n M SD 4 or lower (Not Familiar) 22 2.95 1.36 5 27 3.11 1.45 6 44 3.18 1.32 7 (Very Familiar) 78 3.33 1.58 Table 16 Means and standard deviations for overall preference for participants with differing levels of self-reported preference for the original recorded version of a familiar popular song. Likert-Scale Rating of Preference n M SD 4 or lower (Dislike) 26 2.88 1.21 5 12 3.75 1.48 6 37 3.08 1.55 7 ( Like) 96 3.28 1.48 26

Data Analysis for Research Question Four Will listeners who see videos of music therapists performing familiar popular songs have significantly different ratings of overall preference from those who hear an audio-only version of the same performance? A One-Way Analysis of Variance showed no significant difference between audio-only and audio-video conditions F (1, 169) = 0.01, p =.92. Table 17 shows the means and standard deviations for participants ratings of preference for audio-only and audio-video conditions. Table 17 Means and standard deviations for overall preference for participants who rated audio-only or audio-video performances of a familiar popular song. Audio/ Video Condition n M SD Audio-Only 87 3.22 1.58 Audio-Video 84 3.20 1.34 Correlations between preference ratings and all other music and performance elements can be seen for the audio-only condition in Table 18, and for the audio-video condition in Table 19. Table 18 Correlation between ratings of overall preference and music/performance elements for all performances of all songs, audio only condition. Ratings r t p (2-tail) M SD n Overall Preference 3.22 1.58 87 Musical Skill.66 8.12 <.001 3.76 1.35 86 Singing Skill.65 7.85 <.001 3.16 1.51 87 Sound Quality of Voice.53 5.70 <.001 3.56 1.44 87 27