Portal de Periódicos da CAPES

Translation of research interviews: Do we have a problem with qualitative rigor?

2022; Wiley; Volume: 32; Issue: 1 Linguagem: Inglês

10.1111/nae2.31

1750-4910

Lisa McKenna,

Innovations in Medical Education

Qualitative research has been widely adopted in nursing and other health professions to illuminate and understand the realities of work, patient experience, and education through subjective experiences and observations. Increasingly, cross-language data collection, analysis, and reporting is being performed, adding translation complexity to the research process. However, to date, this aspect has received little attention. This paper seeks to prompt discussion of the important, but frequently overlooked, aspect of reporting cross-language qualitative research and ensuring rigor of findings and associated responsibilities of researchers, reviewers, and editors. Much has been written about the concept of rigor in qualitative research. Trustworthiness is commonly referred to as a goal to achieve “quality, authenticity, and truthfulness of findings of qualitative research,”1 (p. 254) while some opt to refer to validity and reliability, even though use of these terms in qualitative research is contested by others.1, 2 Despite the individual stance, various approaches have been described to promote qualitative research rigor. Member checking or validation is widely recognized as a process whereby interview participants are invited to confirm either translations or interpretations drawn by the researcher.2 Peer review involving external review of audit trail activities and use of multiple sources of data to facilitate triangulation are also recognized means for managing rigor2, 3 and expected processes reported by researchers. When reporting findings, it is expected that researchers provide details on the strategies employed to ensure the rigor of their qualitative research processes and findings. Internationally, there is a push towards publishing in highly ranked international journals, and this includes researchers in many non-English speaking countries. Hence, authors are more often seeking to publish in English language journals. Increasingly, as journal editors, we are receiving manuscripts describing qualitative research where the data have been collected in a language other than English and are being reported in English, but without description of the processes required to ensure accuracy and rigor of translations. Accepted processes for ensuring rigor all have a role where research data are being translated from one language to another. However, this can be extensively more complicated. For example, member checking can be challenging when working with translations as the researcher would need to ensure the participant speaks and understands both languages if they are to confirm translations. Similarly, external audit may be more difficult if the reviewer is not bilingual. Furthermore, unconscious bias may arise in translation of concepts that are not directly translatable from the source to the target language.4 From personal experience, often authors do not declare the language in which their qualitative data were collected, resulting in oversight of the issue by editors and reviewers. This may not be an intentional situation, as currently reporting frameworks, such as Consolidated Criteria for Reporting Qualitative Research5 and Standards for Qualitative Reporting,6 do not require this to be reported. As a reviewer, I have experienced many negative reactions from authors who have declined to amend manuscripts because the reporting frameworks do not require them to address this aspect. It is likely that accepted processes for ensuring quality of translations probably have not been undertaken in these cases. This raises the possibility of poor integrity of research findings, and potentially negative impacts for the journal itself around the quality of its publication rigor. The translation of qualitative data has received intermittent attention in the literature, yet few if any, journals appear to have implemented processes for evaluating quality in this regard. The greatest risk in data translation relates to potential loss or alteration of original meanings and emphases in translated text. Metaphors, for example, may not have the same meanings in other languages, while some aspects may not have easily translatable equivalents,7 particularly jargon, slang, and regional linguistic characteristics.8, 9 Furthermore, elements of body language captured in field notes may have different interpretations across cultures.10 From the available literature, it is clear that appropriate translation processes add a high degree of complexity to the research process and its reporting. Translating research interviews can be challenging on a variety of levels. Elements of grammar vary across languages, while preserving meaning in metaphors and capturing the semantic and sociolinguistic nuances of individual languages may lead to poor equivalence in translations, and hence, incorrect translated meanings and incorrect findings.9 Al-Amer et al.8 recommend researchers maintain “a trail of atypical words”8 (p. 1160) encountered, such as local idiosyncrasies, jargon, and slang. It has also been suggested that transparency of translations in research reports could be enhanced by providing participant quotations in the original language as well as in English11; however, this is currently not normal practice in nursing or other health professions. Furthermore, research texts are largely silent on this issue meaning there is little guidance for researchers. WHY? The reason for translating WHEN? The time for translating WHAT? The data or content for translating WHO? The person(s) doing the translating HOW? The mode of translating WHERE? The location for translating BY WHAT MEANS? The means and tools for translating (p. 3)12 Regardless of the means, there is a need for stronger reporting of research data translations, and guidance for all key research stakeholders. Translation of qualitative data is not only about language, but also reflecting the culture in which the data were collected.13 Data analysis and translation of interviews can be undertaken at different points in the research process. It is common practice for entire transcriptions to be translated into English, then analyzed as an entire dataset. On the other hand, some researchers analyze in the source language and translate the data components (quotations) that are to be used in reporting findings. Santos et al.14 recommend that translation is done early in the study to facilitate access to the dataset for other members of the research team. However, that will depend on the overall available language proficiency of research team members. The concept of back-translation receives mixed critique within the available literature. Chen and Boore13 assert that back-translation needs to be performed well to ensure truth of reported findings. However, Ozolins et al.15 undertook a review of back-translation in health research, finding it to be a poorly implemented, often flawed, and outsourced process. They suggest having language experts on research teams and translation and back-translation issues directly considered within the team. This approach is reinforced by Al-Amer et al.8 who recommend carefully selecting bilingual and bicultural team members with sufficient qualitative research expertise, and use of translators who have similar culture, language, and discipline expertise. However, this may not be practical for many research teams. Issues surrounding translation raise the additional question of who should perform the translation, and back-translation if utilized. For the purposes of rigor, it is recommended that different people perform each of the two processes in a study.13 Both of these translators need to be truly bilingual and understand not only the languages, but the two cultures.13 Furthermore, van Ness et al.7 suggest that translating across languages entails some interpretations as well, which provides researchers with better understandings of the phenomenon under investigation.8 According to Abfalter et al.12 ideally those engaged in translations know the field of research and act as co-researchers. Qualitative researchers undertaking research in one language and reporting in another need to understand the ways in which to ensure, and report on, the processes undertaken to ensure the rigor of their work. The transparency of transcriptions is key to demonstrating sound procedural and ethical decision-making and ensuring quality in research reporting.11 In reporting on their research, authors should seek to provide specific details of how translations were handled from data collection to reporting. Similarly, editors and reviewers need to be mindful of this issue, observant, and seeking to ask questions when manuscripts are submitted from predominantly non-English speaking countries, or that report on research carried out with particular cultural groups. Finally, and most importantly, there is a clear need for work to revise existing, or develop new, qualitative research reporting frameworks to ensure that translation rigor is appropriately reported. Qualitative research is commonly conducted and important to nursing knowledge. Rigor of qualitative approaches is crucial to ensuring research quality. Increasingly, data are collected in one language and reported in another, yet rigor of translation processes is not widely reported. This paper aimed to bring this issue to the attention of authors, reviewers, and publishers in an attempt to work towards better quality reporting of qualitative research in nursing. It is recommended that authors maintain auditable records of their decision making to minimize associated risk to study rigor and are encouraged to report this in describing their research methods, and ideally, reporting guidelines are modified to incorporate this aspect. Overall, this will increase the quality and trust in qualitative research reporting and knowledge generated. Lisa McKenna, RN, PhD, is Professor, Dean, and Head, School of Nursing and Midwifery, La Trobe University, Melbourne, Australia. She is also Editor-in-Chief of Collegian: The Australian Journal of Nursing Practice, Scholarship and Research.