Blog
3-point plan for producing perfect transcripts
3-point plan for producing perfect transcripts of interviews and focus groups:
- Create a template
- Produce the 1st draft
- Proofread the transcript.
Before we get started: I’ll assume you’ve got a good audio transcription program,
headphones and footpedal. I’ll also assume you’re a touch-typist, have an excellent
‘ear’ and excellent grammar, spelling and punctuation.
1. Create a template
If you are producing a transcript for a research project there are probably a number of transcripts involved. If so, the first thing to do is prepare a template to use for all the transcripts. This will save you a lot of time and ensure that all the transcripts are presented consistently, and that you don’t accidentally omit important information from some of the transcripts.
The template may include headers, page numbers and line-numbering (for
referencing). It may include information that applies to all the
interviews or focus groups such as the name of the interviewer or
facilitator, the topic, the venue or institution, and
the type of interview or focus group. It may also include ‘prompts’ to be
filled out for each interview or focus group such as the time, date, name of interviewer/facilitator etc.
Depending on how familiar you are with the subject matter of the transcripts, you may also wish to create a glossary for the project, or you may already have a glossary that you’ve used previously for a similar type of project. If you have a glossary of relevant terms and words, you refer to it as you work to make sure you spell unfamiliar words (such as specialist words) correctly.
2. Produce the 1st draft
How to produce the 1st draft
Transcribe the audio file on either ‘normal’ or ‘slow’ speed, depending on how quickly the people are speaking in the recording. I usually place the speed at one notch slower than
‘normal’ during the 1st draft stage. Use a 0.5 second, 1-second or 1.5-second
auto-rewind, depending on your preference, although if the audibility is poor you might like to try using a 2-second auto-rewind.
Now, there
are usually some words in a recording that, the first time you hear them, sound inaudible. The trick
when producing the 1st draft of a transcript is to not worry too much about trying to get those seemingly inaudible words right, because it will always be easier to decipher them the second-time round, rather than during the 1st draft stage.
So when you come across a word you can’t make out, just leave a gap in the
transcript (to show that a word is missing) and type on. You’ll be in a much better position to correct those mistakes during the next stage. Similarly, don’t worry about trying to get the punctuation right during the 1st draft stage. It will be far easier to fix it up the punctuation during the proofreading stage.
Accuracy – 1st draft
- If the recording is of good audio quality – that is, if most of the comments are clearly audible – then you won’t have many ‘gaps’,
mis-heard words or typos in your transcript. It will be of quite good accuracy, although there will be some mistakes and it won’t be perfect. - If the recording is of reasonable audio quality – that is, if most comments are reasonably audible but there are a few patches where people are
talking at the same time, or they’re very quiet, or there are interruptions, or there’s background
noise – then you’ll have quite a few ‘gaps’, mis-heard words and mistakes sprinkled throughout your
transcript. It will be generally accurate but very far from perfect. - If the recording is of poor audio quality –
that is, if most comments are not very clear and many comments inaudible due to people talking at the same time or too quietly, or loud background noise, or static within the recording – then the transcript will be full of ‘gaps’, mis-heard words and
mistakes. If the audibility is very poor then the 1st draft transcript may have so many ‘holes’ that it may not even make any sense. The accuracy of this transcript will reflect the audibility: poor to very poor.
3. Proofread the transcript
It is during the proofreading stage that you have the opportunity to perfect your transcript, or at least make it as accurate as possible.
You will be listening to the audio recording a second time whilst correcting the draft transcript. You will be filling in the ‘gaps’ in the transcript, correcting mis-heard words and mistakes, and correcting punctuation to ensure the transcript is as easy to read as possible. If the recording is reasonably audible, you should be able to produce a perfect transcript. If the recording is of poor audio quality, you should be able to produce a transcript that is as accurate as possible given the inferior audibility.
How to proofread the transcript
When proofreading a 1st draft transcript against the recording, if
people are speaking at normal conversation pace, set the speed on
‘normal’. If they are speaking more slowly than usual and the
audibility is very clear, you may set the speed at one notch faster. The only exception to this rule is when the audibility is poor in which case you may need to proofread at a slower pace.
During the proofreading stage, because you are travelling far more quickly through the transcript and audio, and are listening to the audio for the second time, you can ‘hear’ the context and meaning of the discussion more clearly than during the first listen. You are also now familiar with the speakers’ accents, manners or speech and the types of words they use. This makes it easier to transcribe accurately and improve the punctuation. You should use this stage to:
- Correct and/or insert punctuation in such a way that it makes the transcript as
easy to read as possible. - Conduct any necessary research to ensure unusual words and acronyms are all spelled correctly.
- Insert or italicise any necessary ‘transcriber’s notes’ (such as ‘Laughing‘ or ‘Phone ringing‘) within the transcript.
In regards to punctuation: I always insert more commas in transcripts than is the usual practice in corporate writing because people being interviewed or discussing issues in focus groups often speak in very long sentences, or train-of-thought non-sentences, that would not make any sense if not punctuated by numerous commas (or if appropriate, semicolons). You should note, however, that to do this correctly you need very good knowledge of grammar and good comprehension of what the speaker is trying to say.
Accuracy – proofread transcript
If the comments in the audio recording are audible you will be able to produce a truly perfect transcript. If the comments in the recording are inaudible or only partially audible you’ll be able to produce a transcript that is not perfect, but as accurate as possible given the inferior audibility of the recording.
- If the audibility of the recording is poor (that is, much of it is only partially audible and many comments are inaudible), slowly proofreading the transcript against the audio allows you to fill in some of the ‘gaps’, attempt to correct the many mistakes that will be present in the 1st draft, correct the punctuation, and research to correct any unusual words or names. Result: the best case scenario is a transcript that is reasonably accurate, although it may contain some misheard words or ‘gaps’ where comments could not be deciphered.
- If the audibility of the recording is reasonable (but not ‘good’), proofreading allows you to fill in most of the ‘gaps’, correct the mistakes that will inevitably have occurred due to comments being only partially audible, correct the punctuation, and research to correct any unusual words or names. Result: your transcript should be highly accurate except for some ‘gaps’ (missing words) where comments could not be deciphered.
- If the audibility of the recording is good, proofreading allows you to correct any mistakes (most of which are usually minor) and the punctuation, and research to correct any unusual words or names. Result: your transcript should be perfect.
For more information about audio transcription, go to:
Photo by S.W.Kane.
Back To Blog