The right editing level depends on how closely the written record must reproduce the recording.
Unless agreed otherwise, TP Transcription normally supplies Level 3 — Intelligent Verbatim. Tell us before work begins if you need another level, timestamps, anonymisation, a house style or specialist notation.
| Level | Treatment of speech | Typical use |
|---|---|---|
| 1 — Fully edited | The transcriber restructures and condenses speech into polished written English. Incomplete sentences can be completed, repetition removed and grammar adjusted while preserving the intended message. | Reports, publications, articles and material where readability is more important than the speaker’s exact words. |
| 2 — Lightly edited | The speaker’s words are retained but pauses, non-meaningful hesitations, verbal tics and obvious false starts are removed. Grammar is only lightly corrected. | Presentations, professional meetings, podcasts and readable records of spoken material. |
| 3 — Intelligent verbatim | Meaning and characteristic phrasing are preserved. Repeated false starts, excessive “um” and “er”, stutters and repetitions that add no meaning are omitted. Significant reactions can be noted where required. | Research interviews, focus groups, consultations, business meetings and most general transcription. |
| 4 — Strict or true verbatim | Every spoken word is included, together with fillers, repetitions, false starts, stutters and agreed non-verbal events. It records what was said without tidying the language. | Detailed qualitative analysis, evidential review, speech research and circumstances where delivery matters as well as meaning. |
| 5 — Discourse or Jefferson-style | Special symbols mark features such as pause length, overlapping speech, emphasis, pace, intonation, cut-offs and audible breathing. The notation key and required precision must be agreed in advance. | Conversation analysis, linguistics and specialist academic research. |
What does not change between levels?
All levels still require careful listening, speaker identification, consistent formatting and appropriate handling of unclear audio. The editing choice does not correct factual errors made by a speaker or authorise a transcriber to guess missing words.
How to choose
- Choose Level 3 when you need an accurate, readable record of what participants meant.
- Choose Level 4 only when fillers and false starts form part of the evidence or analysis.
- Choose Level 5 only when your methodology requires formal interaction notation.
- Choose Levels 1 or 2 for material intended to be read as prose rather than analysed as speech.
Project instructions
For consistent output, supply speaker names or pseudonyms, spelling lists, terminology, timestamp frequency, anonymisation rules and a sample document. For multi-file research projects we can agree a written project specification before the first recording is transcribed.