Translation in multilingual research

Why do I write this post❓

Cross-cultural and cross-linguistic studies in applied linguistics are in an enviable - yet precarious - position. For a successful outcome, they are reliant on the effective navigation of a treacherous linguistic divide, where the slightest misstep could cast doubt on the validity of their findings. Researchers in this field need to undertake a careful balance, ensuring that the data they collect and analyse is accurate and untainted, while also being mindful of their growing obligation to make it readily available for public scrutiny. The challenge to remain transparent is only further complicated when multiple languages come into play. Flawed translations can easily become the source of skewed research results, while at the same time, a lack of clarity around the procedures used to translate the data means that its trustworthiness can never be fully assured. To ensure success, researchers must be both meticulous and diligent in the way they handle and report their data if they are to keep their work from becoming an unwinnable game of chance.

Translation is no longer simply a nice-to-have skill set, but rather a necessary requirement for the contemporary academic field. With a global society raising and relocating at unprecedented rates, research has taken a turn, embracing a "multilingual turn" (May, 2014). The monolingual bias of traditional scholarly research is giving way: now, data is not just collected, but analyzed and reported across borders, creating a need for translation fidelity. Universities, meanwhile, compete in a global ranking system; as such, international collaborations and faculty modifications are common. Translation, it seems, is gradually becoming "central" to research design and methods in applied linguistics—and, in fact, necessary for the modern academic.

With a background in Chinese-English translation, I post this learning note to take a deep dive into the complexities of translation procedures within the field of applied linguistics. It is fundamentally a guide for researchers working in multilingual contexts, outlining the aims of translation and introducing the reader to a variety of techniques, while critically assessing its strengths and weaknesses. With an eye to the nuances and challenges of translation, it also takes a broader look at the considerations surrounding it, providing an invaluable checklist for researchers in this area to ensure accuracy and integrity.

Why is translation important?

Researchers became increasingly aware of the significance of language decisions in achieving an effective equivalency between a source and target text. Guidelines for translating survey research specified the use of simple sentences, noun repetition, and the exclusion of colloquialisms, passive voice, or subjunctive mood. The concept of “back-translation,” in which the target text was translated backward to the source, was promoted as a strategy for ensuring accuracy and comprehensibility in both texts; however, this practice was met with criticism as there were concerns that a text may be accurate in its structure yet undergo mistranslation, potentially leading to confusion when it is back-translated to the source. It became ever more apparent that successful translations required more than thoughtfully chosen words; accuracy extended to conveying the depth and nuance of the original source text.

Semantic equivalency

The importance of equivalence of meaning in translation is well illustrated in the oft-quoted phrase “I am happy”. This seemingly simple structure can contain multiple nuances that may be lost with a rigid adherence to linguistic accuracy when attempting to translate the phrase into another language. Across contexts, it can mean anything from joyous and smiling to simply being confident with a given proposal. Without an acknowledgement of this semantic equivalency, the meaning of the phrase can be fundamentally altered and make instruments of research or transcripts unreliable.

Idiomatic equivalency

Though typically best avoided, idioms play an unavoidable role in research data. They possess a unique power to convey underlying emotions or social realities in qualitative studies, often calling for substitution with alternative terms in the target text. For instance, the expression “I am feeling on edge” is translated as “I have my nerves outside my skin” in Italian and “I feel nervous, tense” in French. This transition from one language to another, using multiple variations to capture the intended emotion of the source phrase, exemplifies traditional translation standards. That said, there may still be some circumstances in which no equivalent exists, necessitating the direct quoting of idiomatic expressions from the original. An example of this can be seen in a Mandarin language interview, featuring a participant reflecting on the differences between international and domestic students: “Perhaps it is like the Chinese saying ‘You walk on your sunny path, and I go on my single-planked bridge’ (你走你的阳关道，我走我的独木桥). That kind of feeling.” Here, the traditional connotations of the idiomatic phrase provide a vivid and nuanced account of the participant’s unique point of view.

Empirical equivalence

Accurately translating a questionnaire or interview protocol to different cultures demands an understanding of the advantages and challenges of cultural translation. A key element of successful cultural translation is using language that resonates with its target population. While adapting questions to different cultural contexts may appear a daunting task on the surface, educational researchers must remember that making use of cultural translation ensures their reach is maximized and the data collected is relevant to the intended audience. To ensure successful research outcomes, researchers must be aware of the need to consider the cultural contexts of participants when creating questionnaires and interview protocols; isolating them from the same will likely result in data that is skewed or irrelevant.

Conceptual equivalence

Establishing conceptual equivalence when translating is undeniably a complex and multifaceted process, as illustrated by the example of the term 'remedial' in the 2009 PISA survey. Depending on the source and target cultures, the same word might have different connotations and width of usage, leading to potentially serious difficulties when it comes to interpreting research results. In PISA 2009, for example, the same term was reimagined as “learning support” in New Zealand, “catch-up lessons” in England, and “tutoring class,” “rescue course,” and “guidance class” in China, Taiwan, and Macao, respectively. These translations all contemplate a different concept of remedial outside-of-school support, making it clear how navigating the waters of cultural references can be tricky. To protect the reliability of research outcomes, researchers must dive deep in order to make sure that translations fully capture not only the semantic meaning of words, but also the full scope of their cultural implications.

Some basic translation techniques 🐟

Forward-translation

Forward-translation is often seen as an imperfect endeavor – time-consuming and dependent on the skill of individual translators. Yet, in the realm of cross-cultural studies, it can be a powerful tool. To properly evaluate the implications of cross-cultural findings, the initial translations of the work must be conducted with care and precision. Despite the perceived challenges, forward-translation can be mastered, allowing scholars to bridge language gaps in meaningful, valid ways. With this method of translation, researchers can gain insights into previously unknown cultures, uncovering complex narratives that embody both a people's history and their aspirations for the future.

Back-translation

Back-translation often serves as a method of quality assurance in research: a forward-translation is produced by an independent translator, followed by a translation "back" into the original language for a comparison to the source item or text. To ensure accuracy, the forward- and back-translations must be undertaken by separate translators, and it is sometimes suggested that researchers employ the same number of translators for both processes. This approach is further strengthened by utilizing multiple translators working independently, so that the results of one have the rigor of being checked against those of another. By doing this, researchers can ensure that their results provide a true reflection of the original material. When implemented properly, this technique can provide a reliable, precise, and valid translation of any source material.

Back-translation can provide insight into the equivalence of the source and target versions, allowing for comparisons across groups. While this technique offers an initial assessment, there are weaknesses to be aware of. Poor forward-translations which utilise the same words and phrasing as the source language can lead to inaccurate back-translations, creating weak target language versions. On the other hand, forward translations may not be the same semantically though they are easily back-translated. This could be due to the choice of the translator, resulting in differences in the source and back-translation that were never actually present. Overall, back-translation can offer an important step in the systematic translation process, yet more measures are required to ensure the validity of any research which draws upon it.

Bilingual comparison and testing

To ensure maximal inference of meaning between source and recipient, there exists an additional step in which bilingual individuals are consulted to judge any discrepancies between both interpretations. This method allows for further translation, if necessary, and further guarantees thorough understanding between two languages.

Survey research using questionnaires often involves eliciting responses from bilinguals in both their native and additional languages, allowing the assessment of discrepancies between responses. Still, this approach poses a threat of recall bias, prompting some experts to suggest deploying cross-over designs instead. In cases where one language is significantly stronger than the other, it is recommended reversing the traditional order of assessment, and having bilingual participants begin with the weaker language. This strategy particularly applies to applied linguistics, where many bilingual test subjects may be asymmetrically proficient in both their languages.

Bilingual testing is a powerful tool, but it comes at a cost. It is a time- and resource-consuming process, and not always available or achievable to small-scale, unfunded studies; a predicament that many graduate students can easily find themselves in. Global research projects have become increasingly popular, but often involve collaborations between institutions located in different countries that may lead students to endeavors with multi-lingual components. If bilingual test participants are not available, researchers can still apply monolingual testing and examine the response patterns of two distinct language groups.

Review

Ultimately, researchers must incorporate a review component into their approach. Depending on the goal of the study, teams may choose from a variety of review scenarios, from self-referencing, to enlisting independent third-party reviewers, to forming a joint collaboration between a panel of experts and the research team, to involving a panel of experts without the research team. In some cases, one or more members of the research team may be second-language proficient and should be an active part in the review process. If a discrepancy between the source language and translated target language still lingers, researchers may need to conduct additional translation rounds — though with translators completely independent of the project.

Conducting a thorough review of a translation is an essential infrastructure of systematic translation practices. To ensure accurate and comprehensive results, a number of approaches — such as rating scales and including members of the target population in the review committee — can be implemented and incorporated into a review plan. However, the implementations of these approaches require adequate resources, such as funds and a substantial amount of time, as well as cooperation between all the involved parties. Thus, depending on the scope of the project and its desired accuracy, the review process can shift from a single reviewer to a team of independent reviewers with structured response questionnaires and multiple cycles of review.