Auto-Transcription and CAQDAS

Recent technological developments, the availability of speech-to-text APIs and of course the COVID-19 pandemic have combined to have a HUGE impact on a key and often time-consuming stage of qualitative and mixed-methods research: transcription.

I have been working with CAQDAS developers, colleagues and others to try and document what’s available and create step-by-step guidance for the leading packages to enable researchers to take advantage of these new opportunities.

The key posts are linked below, and updates will be flagged on this page.

Overview and details – creating and preparing auto-transcripts

The first one form the inspiring, fantastic and highly-recommended CAQDAS Networking project Seminar Series .

Working in specific CAQDAS packages with auto-generated transcript files (SRT or VTT):

Currently ATLAS.ti has the simplest, most robust and I consider most seamless integration for bringing in auto-transcripts and synchronising them with the media files to correct and code. it’s a joy to work with – detailed here.

MAXQDA also enables easy import of SRT files and these can then be synchronised for correcting, annotating and coding. MAXQDA is detailed step-by-step here.

NVivo has its own transcription service (which is a piaid-for service) which is in my testing less accurate than the APIs used by Teams or Zoom but has good and clear multi-language support. Importing has improved with Release 1 and support for tab separated values and some free engines help. however it requires patience and often needs some fiddly problem-solving to import if you get errors. NVivo is detailed here.

Working with Microsoft Word Online

You can also create transcripts – with speaker labels and timestamps – for free using Microsoft Word online. Working with Word is detailed here and includes information on importing into different packages. There are some limits (e.g. 5hrs audio per month) and again the process of importing is a bit varied.

Presentations and Seminars

Qualitative Transcription for the 21st Century: combining automated transcription with human interpretation using CAQDAS packages. 

Dr. Steven Wright, Learning and research technologist at Lancaster University, UK

The 17th CAQDAS networking webinar had me – Dr Steve Wright – discussing automated transcription with human interpretation using CAQDAS packages.

I also did a seminar for the Centre for Technology Enhanced Learning in the Dept. Educational Research at Lancaster University for their Show and TEL seminars:

Technologies for qualitative analysis, transcription and mixed methods

Discussion with Steve Wright

Overview: Speech-to-Text and CAQDAS (from Teams, Zoom and other systems)