Skip to Main Content

SOCIOL_490: Second Year Paper

An overview of key features in Atlas.ti.

Atlas.ti is software for coding and exploring qualitative data. It has a number of advanced features, but my recommendation when you're starting out is to ignore the bells and whistles and focus just on the core tasks of analyzing qualitative data.

This guide outlines those core tasks, with links to the relevant section of the user manual and a few tips and tricks I've picked up--mainly things I learned the hard way and want to help you avoid. 

Feel free to reach out to me as you're working on this project. I'm here to advise, consult, troubleshoot, commiserate, cheerlead, or whatever else you need via Zoom, email, or in person. 


Import Your Data

My Approach

1. Prepare your documents for import by giving them short, meaningful, unique filenames. Establish a filenaming convention that makes sense for your project. For interview studies like the ones you're all doing I might do something like this: 

  • CaseID_audio.mp3 (audio recording of the interview)***
  • CaseID_transcript.txt (transcript of the recording)
  • CaseID_jottings.pdf (scan of the handwritten notes I took during the interview)

2. Bring your data into Atlas.ti.

  • If you're using audio or video files, bring them in first so that you can associate the transcripts with them. If you're not using audio or video files, you can just import the transcripts as regular documents. 
  • For large multimedia files, Atlas.ti recommends linking to the files rather than importing them; be aware that if you do it this way the links will break if you move the files later. 

3. Attach transcripts to audio recordings. Double click on a recording to open it, then select Import Transcript from the top menu. From here you can either import an existing transcript (e.g., from an AI transcript service) or you can import a blank file which you'll use to manually transcribe the recording within Atlas.ti.

***IMPORTANT NOTE!!! I'm including audio files in this workshop so I can show you how Atlas.ti handles them. But there are some privacy concerns you would need to take into account--a person's voice can be personally identifying, or they may share information that can't be easily redacted. Depending on your project, it may not be appropriate to name your audio files like this or to associate them with other documents in Atlas.ti. 

Coding and Memoing

My Rules of Thumb

The mechanics of coding your data are simple. And therein lies the challenge--it's so easy to create and store new codes that they can proliferate past what you can reasonably use. Here are my strategies for avoiding that scenario: 

Open Coding

  • Keep your research questions front and center. That will help you focus your codes on what is relevant to answering your question. You can always revisit your data later with new questions. 
  • One concept per code. For example, instead of the single code "angst about career prospects," use two codes: "angst" and "career prospects."  You'll end up with fewer total codes because they're more flexible and reusable. 
  • There are two codes I include in every project: 
    • "Great Quotes" for tagging interesting quotes.
    • "Rabbit Trails" as an umbrella category for themes I notice that aren't relevant to the current project, but I might want to explore in the future.

Closed Coding Framework

  • Two levels of hierarchy is usually enough.
  • Five-ish top-level categories of codes (I often have more than this, but if so I don't try to code for all of them at the same time)
  • Five-ish codes per category.
  • One concept per code.
  • No duplicates. 


  • Memo as you go. Brains are for thinking thoughts, not storing them. 
  • Don't overthink it. A memo is just a note to your future self. They do not all have to be formal. If it turns out it was a dumb idea you can delete it.