Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department


Major Professor

Jennifer Bugos, Ph.D.

Committee Member

Victor Fung, Ph.D.

Committee Member

Janet Moore, Ed.D.

Committee Member

Jeffrey Kromrey, Ph.D.


mixed methods, FaceReader, familiarity, audiovisual stimuli, formal training, making sense of the context


The purposes of this study were to explore the preference of the Chinese undergraduate music majors (N = 27) for Chinese Xi-Qu and Western opera audiovisual examples, the reasons for preference, influence of familiarity on preference, and the relationship between preference ratings and the emotions as detected by FaceReader. The mixed research method, convergent parallel design, was used to explore this topic in depth. As Xi-Qu and opera integrate multiple art forms, eight audiovisual examples (Xi-Qu, n = 4, opera, n = 4) were selected as the stimuli to show the characteristics of the two genres. The participants watched the audiovisual examples individually and responded to a questionnaire meanwhile their facial expressions were recorded for FaceReader analysis. The semi-structured interviews were administered to collect qualitative data pertaining to participants’ general opinions about the musical examples, familiarity, reasons for preference, and the emotions encompassing when watching the musical examples. Descriptive and inferential statistics were used to analyze the data obtained from the questionnaire. The facial expressions video files were analyzed by FaceReader. The qualitative data obtained from interviews were coded to find themes.

The quantitative findings suggested that the operatic examples received higher mean preference ratings than the Xi-Qu examples. The top three preferred examples were all operatic pieces while the three least preferred examples were Xi-Qu pieces. Results of one-way ANOVA showed that the difference among the preference mean ratings showed the statistical significance, F (7, 208) = 14.15, p < .01. The operatic examples also received higher familiarity ratings than Xi-Qu examples. The difference among the familiarity mean ratings also showed the statistical significance, F (7, 208) = 2.99, p < .01. The preference and familiarity ratings showed a modest but statistically significant relationship (r = .45, p < .01). A statistically significant relationship was found between the preference ratings and tempo (r =. 23, < . 01). Furthermore, singing was always among the top three most liked elements in the operatic examples, but singing was always among the top three most disliked elements in the four Xi-Qu examples despite that singing was also among the top three liked elements in two Xi-Qu examples.

Numerical FaceReader results showed a strong negative relationship between “angry” and preference (rho = -.976, p < .01). The moderate relationship was found between “sums of negative emotions” and preference (rho = .741, p < .05). No statistically significant relationship was found between valence and preference and between arousal and preference. The results of temporal FaceReader analysis showed that the participants’ emotional response to the audiovisual examples changed with the unfolding visual and audio information.

The qualitative analysis revealed a model of Xi-Qu and opera preference. The model contained the factors influencing preference for Xi-Qu and opera, including personal factors, cultural and environmental factors, visual factors, musical factors, and musical response. Formal voice training was the most reliable indicator of preference for operatic examples. Familiarity gained through guided listening instead of random repetition was positively related preference for Xi-Qu examples. The unexpected findings were the influence of religion and static perspective on preference for music. Implications and recommendations were discussed, and the suggestions for future research were included.