Most foreign-led fieldwork in Vietnam runs through an interpreter. The arrangement is often treated as logistics — someone is booked, and the research proceeds.
It is not logistics. The interpreter sits between you and every piece of data you collect, and how that role is set up determines what you end up with.
What interpretation changes
Four effects, all of which show up in the data:
Compression. A three-minute answer becomes thirty seconds of English. What was dropped was chosen by the interpreter, usually without either of you noticing.
Smoothing. Hesitation, contradiction, a participant changing their mind mid-answer — these are often tidied away. In qualitative work they are frequently the most informative parts.
Terminology drift. A concept with no direct equivalent gets rendered by the nearest available word, and the difference is invisible to you.
Changed dynamics. Participants speak to the interpreter, not to you. Who the interpreter is — their age, where they are from, their apparent status — affects what people are willing to say.
Choosing an interpreter
Six criteria, and fluency is only the first:
Fluency in both languages, including the regional variety spoken at your site.
Familiarity with the subject matter. An interpreter who does not understand the topic cannot tell which distinction matters.
Appropriate relationship to the community. This cuts both ways: someone local opens doors, but if participants know them, candour may drop. In sensitive research, that consideration outweighs convenience.
No stake in the answers. An interpreter who works for the organisation being studied, or who is related to participants, is in an impossible position.
Willingness to be corrected and to say "I'm not sure how to render that".
Available for the whole study. Changing interpreter mid-study introduces a break in your data that is hard to account for.
Ask your partner institution to recommend someone. This is one of the things a local partner is genuinely better placed to judge.
Briefing
Spend two hours before the first session. Six things to cover:
What the study is about and why each question is being asked.
Key terms — go through them and agree how each will be rendered, consistently.
Translate everything, including the parts that seem irrelevant. Say this explicitly; the default professional instinct is to summarise.
Keep the speaker's register. If a participant is blunt, be blunt.
Flag untranslatable things rather than substituting an approximation.
Confidentiality — what they hear does not leave the session. This should be an explicit agreement, and in many settings a written one.
During the session
Five practices:
Address the participant, not the interpreter. Look at them, speak to them.
Short segments. Two or three sentences, then pause. Long stretches guarantee compression.
Watch the participant while they answer — tone, hesitation, body language reach you directly even when words do not.
Ask for clarification when an answer seems short relative to what you heard. "Did they say anything else?" is a reasonable question.
Record, with consent. This allows a separate check later and takes pressure off the interpreter to be perfect in real time.
Translating instruments
For questionnaires and structured protocols, three steps:
Translate into Vietnamese.
Have a second person translate back into English without seeing the original. Discrepancies show where meaning shifted.
Pilot with a few participants and ask what they understood each question to mean — this catches problems that both translations missed.
Report that you did this. It is a standard expectation in cross-language research and its absence is noticed.
Credit and payment
Three points:
Pay at the local professional rate, agreed in writing, and pay on time.
Budget for preparation and debrief time, not only session hours.
Acknowledge them in the paper. Where an interpreter has contributed to interpretation of meaning — as often happens in qualitative work — the question of whether that reaches authorship deserves an honest answer rather than an automatic no.
What to write in the methods
Four lines that readers of cross-language research look for:
That interpretation was used, and in which sessions.
The interpreter's relevant background and relationship to the community.
How instruments were translated and checked.
Whether analysis worked from original-language transcripts or from translations.
The last is the one most often omitted and the most consequential: analysis done on English translations alone is analysis at one remove from the data, and readers are entitled to know that.
How does interpretation change the data?
Through compression, smoothing away hesitation and contradiction, terminology drift, and changed dynamics — participants speak to the interpreter rather than to you.
What should a briefing cover?
The study and why each question matters, agreed rendering of key terms, an explicit instruction to translate everything rather than summarise, keeping the speaker's register, flagging untranslatable items, and confidentiality.
How should questionnaires be translated?
Translate into Vietnamese, have a second person back-translate without seeing the original, and pilot with participants to check what they understood each question to mean.
What is most often omitted from the methods section?
Whether analysis worked from original-language transcripts or from translations — analysis on translations alone is at one remove from the data.
