Recording Interviews in the Field Without Losing Data

Plan for what goes wrong in the field

In an office, a failed recording can usually be repeated. In the field, it often cannot. A participant may agree to one conversation only, a village meeting happens once, and travel budgets rarely allow return visits. The recording plan should assume that batteries will run down, cards will fill, a device will be dropped and the quiet room you were promised will be next to a generator.

Think in terms of single points of failure. If one device, one card, one charger or one bag holds everything, a single mishap loses data. The rest of this article reduces each of those risks.

Equipment and spares

A practical fieldwork kit for interviews:

• A primary recorder. A dedicated handheld audio recorder is the usual choice for its battery life, physical controls and reliable storage. A phone is a workable backup or, for some projects, the primary device; phone-specific technique is covered in the guide to recording audio on a phone for transcription.
• A second recording device, ideally running at the same time as the first during important interviews.
• External microphones suited to your setting: clip-on lavalier microphones for seated interviews, a small directional microphone for group settings. The guide to choosing a microphone for speech explains the types.
• Foam windscreens and, for outdoor work, furry wind covers.
• At least twice as many memory cards as you think you need, labeled.
• Spare batteries of each type, plus a charger that works from the local supply or a power bank.
• Headphones, closed-back if possible, for checking levels.
• A small tripod or stand, and a soft cloth or pad to set the recorder on so it does not pick up table knocks.
• A waterproof bag or case, and silica gel packs in humid climates.

Test everything at home under realistic conditions: a full-length recording on batteries, a file transfer to your laptop, playback on headphones. Learn the recorder's menus before you leave; fieldwork is a poor place to discover what a setting does.

Settings deserve a few minutes of thought. An uncompressed or lossless format such as WAV keeps the archival master in full quality; sample rate, bit depth and levels matter less than microphone distance once they are sensible. Whatever you choose, use the same settings for every interview so files are consistent.

Power without mains electricity

Power is a common failure point in rural and remote fieldwork. Plan for it explicitly:

• Know how long your recorder runs on one set of batteries at your chosen settings, from your own test, not the packaging.
• Carry enough spare batteries for at least a full day of recording, plus a margin.
• Use a power bank for phones and USB-powered recorders, and recharge it whenever mains or solar power is available.
• Turn off features you do not need, such as wireless connections and screen backlights.
• Check the battery indicator before every interview, not just each morning.

Rechargeable batteries are economical but lose charge over days; label charged and spent sets with tape or separate pouches so you never load a flat set by mistake.

Backups: the daily routine

The most important habit in fieldwork recording is copying every file off the card on the day it was made. A simple routine:

1. At the end of each day, copy all new files to a laptop or tablet.
2. Copy them again to a second storage device, such as a portable drive kept in a different bag.
3. If you have reliable internet and your data management plan allows it, upload a copy to your institution's approved storage.
4. Check that each copy opens and plays before deleting anything from the card. Many researchers keep files on the card until it is full, as a third copy.
5. Log each file in your fieldwork register.

Encryption of laptops and portable drives is commonly required for identifiable research data. Your institution's data management guidance will say what is expected; check before you travel, because setting up encryption in the field is harder.

Noisy, outdoor and shared settings

Fieldwork rarely happens in a studio. Common problems and responses:

• Traffic, generators, livestock, markets. Move a few metres away from the source, put a wall or vehicle between you and it, and bring the microphone closer to the speaker. Distance from the microphone matters more than almost anything else.
• Wind. Use a furry windscreen, shelter behind a building or vehicle, and avoid open hilltops. The article on wind noise explains why wind overwhelms microphones and what can and cannot be fixed later.
• Echoing rooms. Hard concrete or tiled rooms make speech harder to understand. Soft furnishings, a smaller room, or recording outside in a sheltered spot can help.
• Other people joining. Decide in advance whether you will pause, continue or note who joined, and make sure your consent arrangements cover bystanders' voices.
• Television, radio, music. Ask politely whether it can be turned down; if not, move the microphone closer and note it in your log.

Do a short test recording in each new location and listen on headphones before starting. Thirty seconds of checking can save an unusable hour.

Metadata at the moment of capture

Recordings lose context quickly. A file named ZOOM0047.WAV means nothing three months later. Capture metadata as you record:

• Start each recording with a short spoken slate: date, location, interview code, interviewer, and any other people present. Avoid saying participants' full names if your data will be anonymized.
• Keep a fieldwork register, on paper or in a spreadsheet, with one row per file: original file name, new file name, interview code, date, place, duration, language, consent status, and notes on conditions.
• Rename files to a consistent scheme as soon as you back them up, such as project code, date and interview code.
• Photograph handwritten notes and consent forms where your ethics approval allows, and store the images with the same codes.

The register becomes the link between your recordings, transcripts and analysis. It also makes data management and archiving much easier later.

Recording consent

How you record consent depends on your ethics approval. Some studies use signed forms; others, especially where literacy or trust in paperwork is limited, use verbal consent recorded at the start of the interview. If you record verbal consent:

• Read the approved wording, not a paraphrase.
• Record the participant's answer to each consent item clearly.
• Consider recording consent as a separate file, so it can be stored apart from the interview content if your data plan requires.
• Note in the register that consent was given and how.

Explain what will happen to the recording, including any transcription by a person or a service and where files will be stored. The specifics depend on your approved protocol; check with your ethics board if your plans change in the field, for example if you decide to use an automatic transcription service you had not mentioned.

Common mistakes and how to avoid them

• Recording on a single device with no backup, for an interview that cannot be repeated.
• Leaving the microphone on the table a metre away from a quiet speaker.
• Forgetting to press record, or pressing it twice and stopping. Check the counter is moving.
• Filling a card mid-interview because yesterday's files were not cleared after backup.
• Relying on memory for which file is which interview.
• Assuming a noisy recording can be cleaned up later. Some noise can be reduced, but it rarely recovers what was masked. The five-minute audio quality check is worth doing on a sample from each location.

From field recordings to transcripts with mydubly

Once files are safely backed up, mydubly can produce draft transcripts from them. You choose a recording from your device, such as a WAV master or an MP3, M4A or FLAC file, or video files such as MP4 and MOV, up to two hours each, and it returns a plain transcript, a timestamped transcript and subtitle files. It detects the spoken language automatically and supports 21 languages, so check that yours is among them before relying on it; many regional and minority languages are not. The audio to text page lists formats and outputs.

Drafts reflect the recording. Wind, distance and crosstalk produce more errors and gaps, and there are no speaker labels. Files with several audio tracks use only the default track. Whether research audio may go to an external service is a question for your ethics approval and data management plan; check with your institution first. Transcription costs one credit per minute, so twenty one-hour interviews cost 1,200 credits ($1.20).

Next step

Assemble the kit, then do a full rehearsal at home: record a mock interview on batteries with both devices, copy the files twice, rename them, log them in your register and listen back. If any step feels awkward, fix it now. If you work with archival or life-story interviews, the article on oral history transcription covers the later stages of that kind of project.