Researching multilingually: Navigating multilingual databases in international educational research

• In Event: Crossdisciplinary Perspectives in Language Education Research

Proposal

Relevance & Context

International research on language issues in education has burgeoned in recent years, due to a combination of factors including growing recognition of linguistic diversity, critiques of using instructional languages that are foreign to learners, and emergence of a range of bi- and multilingual models that use learners’ own languages for initial and continuing literacy and learning (Heugh, 2011; Kosonen, 2017; Thomas & Collier, 2002). In the past 10-15 years, policy and practice in multilingual education (MLE) has become the focus of a great deal of research worldwide, resulting in scholarship that has been published in a number of journals in multiple languages. At the same time, international scholars rely heavily on the three biggest search engines: Web of Science, Elsevier’s Scopus, and Google Scholar to access research articles. While these search engines appear to be less US-centric and English-language centric than before (Anderson-Levitt, 2019), they do not fully capture MLE research published in other, non-English languages. Hence, scholars, practitioners, and policymakers are likely to access MLE research that is biased towards English-speaking contexts or that are written by English-speaking scholars at English-speaking tertiary institutions.

The purpose of this paper is to examine how a cross-institutional, multilingual research team investigated international MLE research across Chinese, English, French, and Spanish databases. Findings offer critical considerations for researching multilingually, while also challenging scholars to uphold social responsibility by “relearning” how to conduct research that accesses language scholarship in the context in which the research was conducted.

Inquiry

To examine international MLE research, the research team sought to document research related to the following lines of inquiry: theories that inform MLE policy and practice; MLE curricula; teaching and learning strategies; MLE assessments; and MLE student outcomes. Research was collected from 2007-2019 in Chinese, English, French, and Spanish databases.

Databases
Across all languages, Google Scholar was used as one of two search engines as it is freely available, accessible in multiple languages, and more comprehensive than Web of Science and Scopus (Harzing and Alakangas, 2015; Martín-Martín et al., 2018). A partner database was also used for each language area. Through discussions with MLE scholars in non-English speaking countries, the following databases were selected: CNKI, China Knowledge Resource Integrated Database (Chinese), ERIC, Educational Resource Information Center (English), CAIRN, an index of French journals and books, and SciELO, Scientific Electronic Library Online (Spanish).

Terminology, Inclusion, Exclusion, Reliability
To generate research in each database, Boolean operators were used in search engines for article inclusion. These terminology varied from database to database depending on the language context, which are discussed in the findings. Next, researchers applied exclusion criteria using titles and abstracts of articles, examining whether they addressed the lines of inquiry (above). The research team coded 10% of the sample until 95% agreement was reached. The remaining sample was coded independently.

Methods for the current study
Throughout the process of data collection, the research team took detailed notes about the process of researching multilingually. Meetings ranged from weekly to monthly occurrences over the course of two years where team members shared about challenges and solutions when researching multilingually. These conversations were instrumental in maintaining the reliability and validity of research in each language context. Afterwards, the team analyzed data collection notes for themes, which yielded important considerations for researching multilingually.

Findings

Selecting a database: When selecting databases for research in multiple language contexts, researchers should first consider using multiple databases to corroborate search results. Asking scholars from institutions in the context being researched will ensure the selected search engines yield representative results. Another consideration is to cross-check exemplar articles to ensure they are available in the chosen language database. For example, when using CARIN, we looked at three articles dealing with MLE in France and francophone African contexts to assess whether they were searchable.

Contextualizing search terminology: Terminology can vary from one language context to another. For example, while “mother tongue education” might refer to non-English languages in North America, the sociopolitical context of China suggests the “mother tongue” is Mandarin, the dominant language. Hence, “ethnic minority education” or “少数民族教育” more accurately captures this phenomena. Similarly, “pédagogie convergente” was included in the French database because it was salient in Francofone research in African contexts. Using highly cited systematic reviews of research in a particular issue will also provide insight on appropriate terminology to include.

Quantity of yielded articles: The number of searched articles varied dramatically between language contexts. In the Spanish language search, approximately 2.6 times more articles applied to the search criteria compared to English. While it is plausible there is more MLE research published in Spanish scholarship than English, scholars should carefully consider the studies yielded as some terms may be more inclusive within a specific language context.

Types of articles: Search engines provided peer-reviewed articles, book chapters, dissertations, master theses, presentations and research-based reports. The quantity and proportion of publication types varied by language database with the most peer-reviewed articles and book chapters available in English research. It is, thus, important for scholars to determine “what counts” as scholarship, especially considering how scholarship is primarily conducted in the specified context.

Contribution

The current study provides key considerations for maintaining the reliability, validity and integrity of a study when conducting research across multiple language contexts. Findings may also provide scholars, practitioners, policymakers, and stakeholders in international education with tools to research multilingually. Reflecting upon the Language Issues SIG call for proposals, this paper calls for a heightened awareness of how to be socially responsible when examining language education in diverse societies. Social responsibility in this case requires scholars to understand research in the context that it is explored and disseminated, and to “unlearn” ways in which we access and define such scholarship. Although English language research often takes priority in Global North contexts, the current study challenges us to “relearn” what we define as scholarship, to “reconsider” whose voices we are listening to, and to “reorient” the ways in which we access and listen to these voices.

Authors

• Kevin M. Wong, Pepperdine University

• Carolyn J. (Carol) Benson, Teachers College, Columbia University

• Erina Iwasaki, Teachers College, Columbia University

• Lajiadou Lajiadou, Teachers College, Columbia University