Abstract
This metadata-based literature study provides a comprehensive mapping of China’s music education by analysing 3,257 Chinese research articles published during 2007–2019. The selection of the articles is based on careful inspections of the China National Knowledge Infrastructure (CNKI) database and rigorous coding of literature attributes. The analysis of descriptive and analytical attributes showed (a) dominance of qualitative over quantitative/mixed-method research paradigm, (b) a strong preference for conceptual over empirical studies and (c) a policy-reinforced research orientation. The prime inference from the findings also raised concerns on several ‘neglected’ research areas and quality issues in the selected literature.
Introduction
In Chinese literature, studies of music education became popular in the early 1980s and were increasingly influenced by educational movements worldwide. In contrast to a booming wave of publications in Chinese, research about China’s music education seems to be underrepresented among a vast body of educational studies in English academic literature. Using ‘music education’ and ‘China/Chinese’ as title keywords, the Scopus online journal database returned 21 matching results to 12 February 2020. No systematic literature review was found. In the China National Knowledge Infrastructure (CNKI) database, 50 articles (out of 62 results) were found to be literature reviews of music education studies from 24,000 articles about music education. Most of the reviews adopted either chronological or selected thematic approaches, whereas only three of these referenced over 50 separate pieces of literature (see details in Supplemental Appendix). Given that most of these reviews did not specify the method by which the review was undertaken, it is difficult to piece together a comprehensive picture of music education studies in Mainland China. Therefore, a systemic literature review in this area is essential and urgent, and requires three consecutive steps: comprehensive literature coverage mapping, robust selection criteria development and a systematic review procedure. The aim of this study was to profile and analyse prominent features of music education research articles that were published in Chinese from 2007 to 2019.
Rationale
In response to the calls for evidence-based practice, recent literature studies used new ways, such as systematic review, to conceptualise research developments and trends in a range of disciplines (Moher et al., 2015). In the systematic approaches to literature study, both metadata analysis and content analysis, either quantitative or qualitative, were used to enhance research reliability by providing specific descriptions on data analysis processes (Campbell et al., 2011). Metadata, as a type of attribution data, is an essential component of contemporary scholarly communication and information management (Borgman, 2012). When the use of different languages sometimes separates research communities, metadata-based studies can simultaneously address the technical, social, and organisational aspects of the development and trends in local communities as well as in global arenas (Bowker et al., 2010). Therefore, metadata schemes can facilitate the documentation, exchange, archiving, and reuse of research information across languages and disciplines, when individual studies can be associated with both cross-disciplinary and discipline-specific data repositories (Fathalla et al., 2018). In many recent literature reviews (Lavranos et al., 2016; Miksza & Johnson, 2012; Nichols, 2013; Young, 2016), thematic reviewing and content analysis were often used as primary methods. These approaches are able to provide either in-depth discussion or descriptive study on the chosen research topic and data source to some extent. However, it is technically difficult to (a) present a comprehensive picture of the reviewed research discipline using a limited number of articles based on ‘expert selection’ or (b) justify the captured characteristics of a large volume of research studies using word frequency mapping. From an information theory viewpoint (see Figure 1), the two approaches, thematic and content analysis, potentially undermined the reliability of any comprehensive literature study that is tended to profile a full view of the body of knowledge studied. As a result, a complete systematic review always demands clarity in (a) research discipline specification (location), (b) data sources description (span), (c) literature selection criteria (index) and (d) the ‘representativeness’ of high-impact literature (magnitude).

Methodology model for systematic literature review (adapted from a base model by Finfgeld-Connett, 2014, p. 69).
Due to the huge volume of literature to be reviewed, it would have been technically challenging and methodologically problematic to use either content analysis or thematic analysis as the starting point for reviewing. Instead, metadata analysis was first conducted to clarify the location, span, index and magnitude of the literature data source. Metadata of academic literature are the fundamental component of any information-based knowledge system, such as journal resources. Since these sources are represented systematically, goal-driven analysis, such as systematic literature review, can be achieved using repeatable measurements and observational/statistical data (Willis et al., 2012). In general, the metadata-based literature analysis is a powerful method to study the developments in research methodology, interests, impact, network and community collaboration in a broader academic entity. It strengthens the reliability and quality of a literature review by (a) consolidating the foundation for the other two methodological approaches and (b) deepening the reviewer’s inquiry down to the database level (see Figure 1). Metadata-based analysis occurs less frequently in music education review studies than in other academic areas, such as public health research (Khan et al., 2016) and business management (Bhosale & Kant, 2016). Similar to studies in these other areas, the metadata analysis in this study is guided by the Greenberg’s MODAL model (Metadata objectives, Principles, Domains and Architectural layout) that was designed to ensure the robustness of overall goals, scope and output of metadata analysis (Greenberg, 2005). The current study sought to profile an up-to-date systematic literature review on recent developments in China’s music education research using the CNKI database.
Objectives
As a part of a comprehensive systematic review that incorporated all three methodological approaches previously described, this metadata-based literature analysis was intended to (a) provide a detailed profile of music education studies during 2007–2019 and (b) establish a concrete foundation and framework for future systematic reviews, using either content analysis or thematic approaches. The four research questions were as follows:
Research Question 1: What is the general overview of music education research studies in Chinese literature?
Research Question 2: What are the prominent characteristics of the academic journal publications in this research area as a whole?
Research Question 3: What were the research trends during 2007–2019?
Research Question 4: Is there any sub-research area that was ‘neglected’ or ‘left behind’?
Method
Article selection criteria
Database
CNKI is the biggest state-funded national database for academic publications that covers Chinese newspapers, dissertations, proceedings, yearbooks and reference works published as early as the 1900s (see details in Table 1). Due to the differences between types of literature, as well as for quantity and quality reasons, the current review used only two key sources in the CAJD database: Chinese Social Sciences Citation Index (CSSCI) provided by the Institute for Chinese Social Sciences Research and Assessment and Core Journals of China (CJC) provided by Peking University Library. Although journals indexed in these two sources are regarded as prestigious academic publications, only CSSCI specified its criteria for selection, such as academic standard, cross-reference and impact factor. To make valid comparisons, only articles published after 2007 were chosen for metadata analysis, as by then the 2005 Descriptive Rules for Bibliographic References had been fully implemented.
CNKI database coverages by 2017.
The database used for this study.
Search protocol
Using the syntax below, the CNKI database returned 45,755 results for the period 2007–2019, of which 5,326 items were from the core journals (CSSCI and CJC).
In total, 3,257 articles (7% of the total) were found eligible for metadata analysis, excluding journal notices, editorial, biography, conference reports, book reviews, and musicological and ethnomusicological studies.
Data analyses
Attribute mapping
Two sets of attributes, Descriptive and Analytical, were developed for literature coding. The descriptive attributes were retrieved directly from the data source to capture most physical characteristics of the articles, whereas the analytical attributes were abstracted from the article contents, which present research characteristics of the literature (see Table 2). The metadata analysis comprised two steps: attribute coding and statistical analysis of attributes. The latter included both descriptive statistical analysis of each attribute category and a combined analysis of all attributes.
Operational definitions for attribute coding.
Coding procedure
Two general attribute categories were applied in the article coding: (a) descriptive attributes that were retrieved from metadata of articles and (b) analytical attributes that were labelled by 20 trained research students under the supervision of three expert researchers (see Figure 2). The overall procedure comprised two major phases: coding and checking. In the coding phase, 80 example articles were first coded by the experts for coding training, and then the remaining 3,178 articles were evenly distributed to eight teams of two research students. In the checking phase, (a) each team completed 398 coding items that were checked by both students, of which (b) 40 items were then randomly chosen for subsequent swap-coding with another team, then (c) 15 out of the 40 swap-coding results were again selected randomly for coding consistency inspection and (d) upon the completion of coding, all items were assigned to the three expert researchers for the final check.

Coding procedure workflow.
Result
Descriptive attributes analysis
The overall coding consistency (intercoder reliability) rate was above 94% in this category. The 3,257 articles distributed almost evenly across the focus years with the exception of a sudden increase during 2012–2013, which is suspected to be the result of a core journal list extension in the Chinese Core Journal Directory Sixth Edition (Zhu et al., 2011). Due to the exclusion of 77 review articles, this was a considerable drop in publications in 2019. Given that 83.4% of these articles were contributed by single authors (Table 3), teamwork seems to be rare in the current research community dominated by higher education academics (including Teachers’ University, Comprehensive University and Music Conservatory). In terms of format, 32.8% of articles were less than 3,000 words in length, which seems to be a very challenging threshold for developing adequate discussions in Chinese writings. Although 39.7% of articles were in the range of 3,000 to 5,000 words, similar studies in other education subjects often devote 5,000 to 7,000 words to elaborate reasonable accounts of the research. Referencing was found to be a major defect in writing, in the sense that 16.8% of articles had no references and half of the articles provided no more than five references. From the remaining 83.2% of articles with references, 39.2% failed to make correct matching between in-text citations and the corresponding reference list. Excluding translated works, very few articles (8.3%) used non-Chinese literature in writing. In total, 25.3% of articles were associated with state-funded projects.
Coding results by descriptive attributes.
Analytical attributes analysis
The coding of analytical attributes largely depended on the professional judgements of the coding teams and the three experts. The coding consistency rates on five attributes were above 80%, while the rates on ‘Research area’ and ‘Field in education’ fluctuated between 52% and 76%. For the Research Category section, working definitions for educational research and each sub-category were drawn from established work by Johnson and Christensen (2014, pp. 5–12, 38–52) with further reference to works by Laverty (2018, pp. 16–17) and Gall et al. (2010, pp. 9–12) on critical-theoretical and conceptual research. Conceptual studies, which are related to abstract concepts and ideas, often adopt a methodology whereby research (a) is conducted by analysing existing information on a given topic and (b) does not involve any observational activities or practical experiments, as is the case with educational studies. As a result, research fieldwork was used to set a clear division between conceptual and non-conceptual studies, while the nature of research data collection and analysis determined whether the study was categorised as (more) qualitative or quantitative. Experimental research was used in the conventional sense in educational studies. In contrast to a tiny proportion (3.8% in total) of quantitative research, the overall dominance of conceptual studies (60.1%) appeared to be different from the current practice of music education studies in English academic publications (see Table 4). Only 32.0% of the articles provided original data or reported fieldwork data. Among these, tertiary education seems to be the major focus of study, while 29.6% of articles discussed music education holistically (or fail to specify the educational contexts). Given that 88.4% of articles have not provided sufficient explanation on their research method(s) or methodology, it was difficult to determine the research area and focus for some articles. As a result, analytical attributes with coding consistency rate under 80% were used as supplementary information in subsequent analyses.
Coding results by analytical attributes.
Combined analysis
Combined analyses on descriptive and analytical attributes revealed that only 313 (9.6%) of the selected articles provided both fieldwork data and research method information, whereas over three-quarters of qualitative empirical studies had not reported their methodological design. Further breakdown of these data by authors’ affiliation suggests that academics from tertiary education are very active in undertaking research across all range of educational contexts, while school-based researchers tend to focus on their specific education sectors only (see Table 5). However, apart from studies on their working environment (372 in tertiary with fieldwork data), 53.8% (433 out of 805) of publications provided very limited information about fieldwork in studies ranging from pre-school to special education, which seem to be ‘distant’ from their areas of ‘expertise’. This, unfortunately, implies a serious concern in relation to the credibility of these studies. On the contrary, voices from school-based researchers were much less in the community due to possible limitation of their employment conditions (time and resource).
Articles that provide fieldwork data and research method(s) by authors’ affiliations.
Pre: pre-school; Pri: primary; Sec: secondary: Comp: compulsory; Ter.: tertiary; NFE: non-formal education; SPE: special education.
Supported by quantitative content analysis methods, metadata-based attributes were used to categorise 3,257 journal articles by research networking, academic standard adherence, bibliography, methodology approaches, research topics and contexts, and fieldwork employment. Results of the metadata analysis also revealed that
Teamwork and connection with international communities were not very common in the selected articles, in the sense that 83.4% of the articles were single-authored, which is three times the typical number of single-authored papers (27%) in current social science studies (Macfarlane, 2017).
Academic writing standards in research studies were under-developed, particularly in regard to referencing and citations not receiving appropriate care in many cases. Although this appears to be a matter of writing accuracy, it may potentially lead to concerns of plagiarism.
A vast majority of studies failed to provide any specifications on the fieldwork and/or research method(s). This reflects a challenging transition from the old standard to a more ‘international’ academic standard, as has been reported in other non-English research communities (Kelemen, 2014).
There is a clear domination of conceptual (60.1%) over empirical studies, while quantitative or mixed research approaches are marginal. In education research areas, it is surprising to find that most of the research works have concentrated, primarily, on theoretical topics.
In comparison to tertiary education (52.2%), research articles focused on pre-school, primary and secondary music educators are underrepresented as to the population concerned. This goes against the general developmental trends under a nation-wide reform of quality education since 1999 (Ministry of Education).
Music education research for/about adults and senior people is extremely rare, whereas music course resources have been rapidly developed to serve 8 million students in 70,000 universities for the elderly (Mu, 2019).
An additional word frequency analysis of article keywords in the selected literature suggests that curriculum reform (343), aesthetic education (147), traditional music (98) and education philosophy (64) seem to be the long-lasting, trending themes across formal music education. This echoes the trends in other educational disciplines and general social and cultural policy movements in China during this period.
Five articles were found to be either complete or partial replications of other authors’ previously published articles, having only minor changes in their titles. Another 28 publications seem to have replicated either the discussion structure or findings from other published studies, with minor modifications to research contexts and data sources only. These articles seem to suggest a practice of intentional plagiarism that could be mostly related to (a) the lack of regulated academic practices in music education previously and (b) urgent project publication requirements. Fortunately, the loophole for similar academic misconducts has been fixed after the release of the CNKI digital Academic Misconduct Literature Check (AMLC) system that is governed by strict state rules (Ministry of Education, PRC, 2016).
Discussion
In general, music education research in Chinese literature appears to have been dominated by conceptual studies from 2007 to 2019. Research paradigm was not adequately defined in a vast number of studies that claimed to be ‘applied’, such as comparative, historical, narrative and phenomenological. The fusion of these paradigms commonly appears in comprehensive discussions about music education, which are often regarded as ‘educational research’. One reason for this misunderstanding might be the ambiguity of the endocentric phrase in Chinese ‘音乐教育研究’, where two types of interpretations are both valid, namely ‘research about music education’ and ‘educational research in music’. The confusion between these two seems to be inevitable because music education is currently a sub-discipline under musicology in Chinese academic taxonomy (China National Institute of Standardization, 2009). Most articles in this review may not be formatted as typical ‘educational research’ in relation to methodological aspects. A possible reason for this could be that interdisciplinary studies (music psychology and education) are technically mutually exclusive in this tree-structure taxonomy, while articles and degree theses are very likely to be reviewed by musicology experts, rather than education researchers, in most cases. This attempted fusion of education and musicology research paradigms not only hinders effective dialogue between domestic and international academic communities but also prevents music education from being further developed into a separate and specialised discipline – as for the case of physical education that has been recognised as a legitimate academic domain in the taxonomy.
Second, the usage of academic terminology by individual researchers is sometimes inconsistent and ambiguous, which is particularly obvious for abstract concepts, such as ‘ethnomusicological perspectives’, ‘aesthetic education’, ‘praxical philosophy’ and ‘postmodern’. In cases of different academics, potential conflicts exist among multiple interpretations of these terms, which were often quoted out of their original contexts. A personal flavour or ‘favour’ (in the sense of common interest and agreed viewpoints) was sometimes added to certain ‘imported’ scholastic concepts and theories from ‘the West’. This potentially made some of these writings unfamiliar to Chinese readers and even less recognisable to experienced scholars in the English-speaking research community. Therefore, standardised usage and application of academic terms, as well as research norms, should be developed and practised in Chinese academic publications.
Third, in intensive discussions on the revival of Chinese traditional music in music education, several philosophical strands, such as historical-materialism, postmodernism and postcolonialism, are constantly incorporated into all aspects of music practices, suggesting an urgent need to alter the ‘Westernised education system and mindset’ in teaching and learning. If these three statements were valid, the proposed ‘research paradigm shift’ towards a ‘Chinese music ecology’ requires not only methodological adaptations to a new academic culture but also fundamental transformations of modality and mentality, which will very likely be under the constant challenge of logical relativism. Considering the lack of fieldwork and background information among many selected studies, the observability (in a general sense) of ‘the alternative research philosophy’ is still contested (Li, 2016). The constant sudden increase of ‘hot’ research topics after the release of state policies seems to suggest that most educational studies were ‘riding with the tide’, rather than ‘leading the change’. This partly explains the intensive criticism (or hindsight) of every ‘old’ curriculum when each was abolished.
Finally, the research interests in the music education community seem to be distant from ‘real world’ issues, in the sense that, in either the CNKI database or articles selected for inclusion in this study, no empirical study ranked in the top 20 most-referenced articles or highest citation rate index (refer to the full coding data source). Instead, policy and philosophical studies attracted the most attention from contributors across all ranges. Although it is understandable that conceptual studies may have better compatibility with different educational scenarios, empirical studies can provide contextualised discussions on practical issues and possible solutions. The top-cited article is an extreme case where no reference was provided. Therefore, the music education world in the research community seems to be a self-sufficient system, which casts a mirage of trending topics.
Limitation(s)
Working with a considerable amount of literature, this study has limitations in sample selection, coding reliability and validity, and reviewing strategies. First, only articles from the ‘core journal lists’ (high impact index) were used as high-quality literature in the initial attribution analysis, leaving the remaining 89% articles ‘untouched’. Although it would be overly optimistic to believe the official impact indexes without question, the core journal lists offered an economical option between quantity and quality. To reduce data selection bias, search protocols for article retrieval were designed to be more accommodating to cover most possible combinations of the applied keywords. Second, because attribute coding was based on the professional judgement of a team of trained research students and three experts, the study applied multi-level coding and randomised double-checking methods to minimise any error in individual differences. Although these efforts will have greatly enhanced the reliability of attributes analysis, the criteria that were used for attribute coding may not have covered all standards in the quality evaluation of academic writing. Questions may be raised on the direct adoption of Johnson’s categorisation and definitions for educational research and then applied to research studies conducted in a Chinese context (validity). Given that an agreed ‘Chinese paradigm’ is not available yet, a ‘temporary’ solution is to use the ‘Western paradigm’ as a reference. Accordingly, analytical attributes are used as additional parameters to support the data output from descriptive attributes.
Reliability of data retrieval methods can be a challenging issue in any metadata-based analysis using different databases or languages. A prudent examination of the database source is essential to ensure the data retrieving protocol is compatible with the system and optimised for the research purposes. Any omission or inclusion of metadata strings, such as keywords and abstract, will have significant impact on the retrieval results. Meanwhile, subtle semantic variations and cultural norm difference may generate ambiguity and contradiction in data retrieval. To resolve these problems, combining multiple sets of retrieved data along with manual metadata checking and coding are crucial in order to filter out redundant data and to better understand the data source as a whole. To further examine the reliability and validity of coding, full data coding results are available for public access upon publication. It would be worthwhile to compare the original coding data with public inputs again after some time.
Summary
While this study generally supports most findings from previous literature reviews, both Chinese and English, on Chinese music education, it is evident that many of these (a) presented certain parts or sections of a comprehensive landscape, (b) have less specification on quantity and quality issues in literature selection, and (c) provide limited analysis of research attributes. The application of metadata analysis makes the interpretation of results and comparison of findings possible and meaningful, as each piece of literature corresponds to a set of common attributes that represent its ‘coordination’ and ‘magnitude’ in the academic knowledge system. These set up a robust foundation for further systematic review and inquiry into Chinese music education research and pave the way for cross-disciplinary studies in an international arena.
Supplemental Material
sj-pdf-1-ijm-10.1177_0255761420988923 – Supplemental material for Mapping music education research in Mainland China (2007–2019): A metadata-based literature analysis
Supplemental material, sj-pdf-1-ijm-10.1177_0255761420988923 for Mapping music education research in Mainland China (2007–2019): A metadata-based literature analysis by Yang Yang, Aiqing Yin and Tao Guan in International Journal of Music Education
Supplemental Material
sj-xlsx-2-ijm-10.1177_0255761420988923 – Supplemental material for Mapping music education research in Mainland China (2007–2019): A metadata-based literature analysis
Supplemental material, sj-xlsx-2-ijm-10.1177_0255761420988923 for Mapping music education research in Mainland China (2007–2019): A metadata-based literature analysis by Yang Yang, Aiqing Yin and Tao Guan in International Journal of Music Education
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship and/or publication of this article.
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
