How to Analyze Qualitative Data

Step-by-Step Guide

1. 1

   Step 1: Choose your approach.

   Hopefully, you chose your analytical plan when you were deciding on your methodology and well before starting data collection.
   
   You should have planned how you were going to analyze your data, how that would influence your methods, what data you would collect and how.
   
   See for an overview of choosing a qualitative research methodology.
   
   However, sometimes people change their approach in response to the data collection process.
   
   There are dozens of different approaches, and different disciplines use different terminology to describe what often appear to be the same technique.
   
   However, everyone’s approach is a little different, so read more about some of these broad types, and choose a technique or combination that is right for your research.
   
   Here are a few possibilities:
   Thematic / content analysis Generating ‘codes’ that describe themes in the text, such as ‘Anxiety’ or ‘Eating habits’.
   
   Discourse analysis Looking at the data in the context of a wider discourse, such as political, historical trends or a group setting.
   
   Narrative Analysis (Phenomenology) How people describe their own stories, using language, time and metaphor.
   
   Semiotic analysis (Hermeneutics) How participants (or the researcher) understand and interpret the world by their nuanced use of language and words.
2. 2

   Step 2: Develop your framework.

   There are two main approaches to choose from here:
   Grounded theory / emergent coding / inductive (data driven) This is where you don’t know beforehand what you are looking for in the data, and identify topics as you are reading it: you are creating theory on the fly.
   
   Framework analysis / structured / (theory driven) Essentially where you use pre-exiting theory and your research questions to set out topics you look for before you start the coding process. , Qualitative analysis is often an iterative approach, but it really helps to sit down with your data after you have collected it and read it all through before actually starting any analysis.
   
   Time constraints can make this feel difficult, but otherwise you may end up having to go back through your data because you noticed an important theme you hadn’t coded in the first few sources.
   
   If you’ve transcribed the data yourself (for example from recorded interviews) this really helps you be close to the data, but even if you had someone else do the transcription, read it through to check for misheard words. , This will also depend on the format of your data, especially if it includes multimedia sources like audio, video or pictures.
   
   Many people choose to use paper copies of their data, colored highlighters or markers to highlight parts of the data about particular themes.
   
   You can also use sticky notes, or cut out sections of data, and stick them onto charts on in files keeping all your relevant sections together.
   
   You may also choose to take one large sheet of paper (OSOP) and use this to explore and manage your themes.
   
   Dedicated software for qualitative analysis can help manage, code and explore connections in coded data, but can’t do coding for you.
   
   An independent overview of the different packages can be found at the CAQDAS Networking Project.
   
   However, many people choose to use standard spreadsheet or word processing software to help manage their data. , Now you should be ready to go through your data, source by source, line by line, and reduce it into meaningful codes.
   
   You should expect this to take some time, for example, as much as 1-4 hours for every hour of collected and transcribed interviews.
   
   Especially when using emergent coding, you may go through once and do very obvious ‘low-level’ coding, identifying things like ‘doctor’ and ‘treatment’.
   
   Then you could go through again, having broken down all the data into these codes and identify ‘higher-level’ codes, which with your research questions such as ‘fear of hospitals’ or ‘poor communication’. , Even before you have finished your coding, you might want to stop and examine the process.
   
   What themes are coming up a lot? What connections and trends are being shown in the coding? What is missing? Once you have finished, you will want to share your findings with others.
   
   That's a topic in itself, but you can share quotes, images, but mostly write from your experience with analyzing, what important factors you have discovered in the data.
3. 3

   Step 3: Get to know your data.

4. 4

   Step 4: Choose your technology.

5. 5

   Step 5: Code and recode.

6. 6

   Step 6: Explore and share results.

Detailed Guide

About the Author