A quitting interview was transcribed: check what the written version preserves
Trace speech-to-text choices and checking before treating a transcript or published extract as an untouched recording.
No Vape · Updated

Locate the stage called the transcript
A recording, a working transcript and an extract edited for publication can be different objects. Read the methods for the version the analysts actually used. A public quotation may be shortened for space while the working material was more detailed; neither state should be guessed from a smooth-looking sentence.
Bailey’s 2008 original methods abstract explains that transcription involves decisions about detail, interpretation and representation according to analytic aims. We use that concept only, not its example utterances or a claim that one transcription convention is universally best. The reader’s first task is to find what the study says it preserved.
Ask which details matter to the reported interpretation
An analysis of the topics people discuss may use a different written representation from one examining interruptions or pauses. Check whether the report explains its convention and its relationship to the question. The absence of pause notation in a published quote does not prove that a pause never occurred or that it was absent from the working transcript.
Consider an invented recording with an unfinished sentence followed by a self-correction about a meeting time. A cleaned extract might show only the final time. A reader assessing practical scheduling themes and a reader assessing how the correction happened would need different detail. This example illustrates a documentation question, not an accusation that any real study changed a participant’s meaning.
Keep production and verification distinct
Identify any stated human transcriber, software process and checking against the recording. Naming a transcription tool does not establish that every word was verified; employing a person does not prove error-free text either. Preserve the actual checking description and leave the rest uncertain. You do not need to request private recordings to note missing reporting.
Translation is another transformation, if it occurred, and needs its own account. Do not describe a translated English quotation as the participant’s original English words. Equally, distinguish a correction to the transcript from a participant later revising what they want to say. Those changes can have different provenance and should not be silently combined.
Finish with a short transformation record
Write the source medium, stated transcription approach, checking described and the status of the quoted extract. If one stage is unclear, ask a bounded public-methods question rather than treating the whole analysis as verified or invalid. An article can report useful interpretations while leaving a documentation detail unresolved.
This exercise does not independently authenticate a recording, recode interviews or measure a program’s effectiveness. It also does not justify transcribing someone else’s private support conversation without agreement. Keep ordinary personal notes factual and limited to their purpose; research transcription is a separate process with its own consent and methods requirements.
Sources and scope
Sources inform the guidance. The worksheets and examples are No Vape editorial suggestions, not tested treatments.
Written with AI assistance and checked against the cited sources. No clinical review is claimed. This guide covers practical planning, not individual medical treatment. How we research and write