English has been the most dominant language used to communicate research and M&E findings, even in countries like Afghanistan, where the language is barely spoken. The chief reason for the given practice is that enterprises and donors who fund these studies are international organizations adopting English as their official language. While data collection is primarily conducted in local languages, the collected data ends up being translated into English. Two of the most common data collection methods are Key Informant Interviews (KIIs) and Focus Group Discussions (FGDs) which mainly gather qualitative data; this article intends to discuss why qualitative data translation is crucial and should be conducted with a quality-centered approach.

Qualitative research deals with the interpretation of textual, verbal, or visual data. The data are usually generated through KIIs and FGDs, and the researchers systematically sense meaning from them. As mentioned earlier, before the analysis stage, the data is collected in the local languages and then translated into English. For the data analysis to be accurate and consistent, the quality of translation should be cross-checked and improved. A poor quality data translation will directly reflect low quality, misinformed, and inconsistent data analysis. In other words, poor quality translation in qualitative research can lead to misinterpretation as qualitative research analysis is interpretative.

Unlike quantitative research in which the analysis is done based on some standardized measures and statistical operations with the help of previously defined formulas, the researcher should represent his understanding of the data in the research findings. Therefore, the validity of the research will be jeopardized if the researcher does the analysis based on some poor quality or inaccurate data resulting from low-quality translation or any other misconduct during the research. Moreover, the worst possible scenario can be if the subsequent data is used for decision-making based on these findings. Therefore, to ensure ethical, unbiased, and accurate analysis, it is stipulated to conduct a careful and high-quality translation of the research data.

The Not-so-simple Stage of Transcription and Translation

Before collected data is translated, there is another phase called transcription, which can affect the translation and the whole research findings. Transcription in the linguistic sense is the systematic representation of language in written form. The source can either be utterances (speech or sign language) or preexisting text in another language. (Transcription-Linguistics, 2020) However, in our context, it involves a native speaker listening to recorded data and representing it in a written document. It may sound like a simple task, but for anyone involved in such a process, it is obvious that there are hidden challenges behind the seemingly simple instruction to “transcribe” the data. For instance, we should know that a recorded document of 30 minutes can take several hours to be transcribed. It could be exhausting and overwhelming even for an enthusiastic researcher and fast typist.

It is suggested that qualitative researchers should have the opportunity to experience for themselves what it feels like to do transcription. It is recommended that the person who does the transcription be present at the interview or at least engage with some parts of the research team. This helps ensure the transcriber has a good sense of the study objectives and can understand the discussion context. (Sophie Chabeda, Jane Kahindi, Manyan Van Ryneveld, 2018)

2. Translation

The translation is the communication of the meaning of a source-language text utilizing an equivalent target-language text. (Translation, 2020) In simple words, translation is the process of creating a written document from a transcript from one language to another language. The most crucial aspect of translation is maintaining the integrity (meaning and significance) of the document (Sohpie Chabeda, Jane Kahindi, Manyan Van Ryneveld, 2018). Therefore, it has to be done by someone who has a complete understanding of both the original language source in which the data is collected and the target source, which is usually English. The quality of the translation will significantly rise if the translator has background knowledge about the research topic or field. For instance, a translator with a degree in sociology can provide a more quality translation in social science research as s/he is already familiar with the jargon and technical expressions.

3. What Makes a Translation Acceptable?

What makes a translation acceptable or not is a narrow line of being or not being “Fidel” to the source text. Merriam-Webster Dictionary defines “fidelity/faithfulness” as the quality of being accurate, loyal, and exact. According to some critics of translation, fidelity in translation is just a word-for-word transmission of a message from the source text to the target text. Meanwhile, some believe that fidelity to the source text is adopting the free, idiomatic method in passing on the message. (Kolawole, S.O. Salawu, Adewuni, 2018). The second opinion fits more than a literal translation mostly does not work mainly because linguistic structures differ from one language to another and it might be difficult to understand unless one is a bilingual person. Therefore, a sense-for-sense translation of the message is a better approach since it lets one maintain the integrity and stay faithful to the source text.

Some Tips to Ensure Quality Translation:

• Making sure that one is familiar with the research topic, content, and objectives. For instance, before one starts translating a focus group discussion or a key-informant interview, one should know what the research is looking for in general. This will dictate and direct the focus on the main parts of the data. This tip comes in handy when the respondents provide a lot of irrelevant insights in an in-depth interview.
• Avoiding literal translation. As previously discussed, not only does word-for-word translation sound of low quality to the readers but also it does not make sense in most cases. It is not acceptable unless the material for translation is that of a technical, i.e., medical prescription. While it can be an exhausting task to avoid being literal with a seemingly difficult sentence, changing the word order or reorganizing it can provide a clue to decoding it. In other words, sometimes it is necessary to split or combine sentences.
• Contacting the data collection officer or transcriber, the transcription or parts of it are vague, senseless, or difficult to understand. Or listening to the audio/video file for a better understanding of those parts can also work
• Resisting the temptation to tidy up people’s comments. There should be no effort to ‘clean’ or ‘clarify’ what people are saying. (Sophie Chabeda, Jane Kahindi, Manyan Van Ryneveld, 2018). Any effort to clean or shorten the sayings of respondents is an act of involving one’s judgments in the translation. For instance, one’s attempt to not include parts of a source text into the translation means one thinks it is unworthy or unnecessary. Neither one has the right, nor is it ethical that one involves their judgments in translation.
• Giving a final review of the text before one closes the document. It will take a couple of minutes, but it is fine. A small typo one may consider unimportant would destroy all of the quality steps one has carried out until now.
• Remember to include notes or comments for the analyst about any things one may think are important to be considered during data analysis.

Written By Rasool Ahmadi, Senior Translation and Data Verification Specialist at AACS Consulting.