How to transcribe an interview: A complete guide

Those who conduct comprehensive research know that administering and recording interviews is necessary for collecting project-related information. The best way to utilize interview findings is to create an audio-to-text transcription, which allows for a more focused analysis of the qualitative data.

Reading from a text document is less problematic than listening to lengthy audio recordings, and only transcribed conversations allow the researcher to efficiently sift through pages of dialogue.

Five steps to successfully transcribe an interview

Step 1: Maintain a log

Prior to the initial interview, the interviewer or an assigned individual should begin keeping a detailed research log. The log is used to record the contributions of every person involved in the interview and transcription, including those who edited or took part in publishing the final document. 

Step 2: Choose a transcription method

Transcriptionists who prefer not to manually convert the audio interview to text can use an automated transcription service such as Sonix.

You can also outsource the job to an experienced freelancer (consider the professionals at Upwork for example).

The method for transcribing an interview depends on the complexity of the audio file, the required level of transcription accuracy, project deadline, and confidentiality standards. Those who need a quick turnaround should use an online service or freelancer who can competently expedite the job.

Step 3: Edit and format the transcription

No transcription method will produce 100% accurate results, so when the initial audio-to-text is complete, the document must then be corrected of any errors in a final quality screening. Throughout the editing process, the transcriptionist should keep the following practicalities in mind:

• There are two transcription modes: naturalism and denaturalism. Naturalism is a precise transcription, completely verbatim, with every word and vocal nuance of the interview intact and transcribed in detail. The denaturalism mode is a considerably broader form of transcription, enabling the transcriber to remove errors in speech like stutters, verbal tics (i.e. ‘like,’ ‘uh’), and involuntary vocalizations (i.e., crying or laughter).
• Timestamps identify where particular areas of audio can be found in the text and should be utilized if the transcript will be used for analysis or converted to subtitles.
• Answers to questions like how to enter names, the correct use of margins, or how to recognize connections with other transcriptions, can all be found on the web. Websites like Quora can be reliable sources.

Step 4: Request an interviewee review

This step requires the transcriptionist to ask the interviewee to read the transcription and verify that it accurately documents their recorded conversation.

After reviewing the transcription, an interviewee may request the transcriptionist remove statements or entire portions of the interview. If this occurs, the transcriptionist should remind the interviewee that they need only check the accuracy of the document (i.e., to fill in any inaudible parts of the recording or to define acronyms), and not alter the dialogue. 

Step 5: Anonymize the transcript

Before publishing the transcript, all names involved must be anonymized; this ensures that the interviewee, as well as contextual names relating to them (organizations for which they’ve worked, towns in which they’ve lived, etc.), cannot be traced back to a particular individual. Actual names may remain on the transcript during the analysis stage but anonymized before publishing. 

There you have it! A complete guide on how to transcribe an interview.