Themes Don’t Just Emerge — Coding the Qualitative Data

ProjectUX
Jul 23, 2018 · 5 min read

Written by Erika Yi, Ph.D.

Interviews are one of the best methods for qualitative research. Most qualitative UX researchers are familiar with building rapport and conducting interviews, but that is not the end of a qualitative research. The analysis process of the interview data is not only vital to render useful research insights but also essential to build your credibility as a responsible qualitative researcher. We don’t have time to explore all the nuance of qualitative data analysis in this article, instead, we will focus on the coding aspect of data.

Why Do You Need Code Your Qualitative Data?

Coding the qualitative data makes the messy scripts quantifiable. How are you going to convince the stakeholders that the insights you collected in the interviews actually reflect users’ needs and wants but not something you just eyeballed five minutes before the meeting? Simple, show them your data.

Codes in qualitative research are as important as numbers in a quantitative study. Your codes give you credibility when presenting them to your teams, your clients, and your stakeholders. With proper coding, you can say with confidence that these findings are in fact, representing the majority of user feedback.

Coding gives data structure. (Photo by Rachel Lynette French)

Coding the qualitative data creates structure. “But we already have an interview script!” you say. Yes, a structured interview protocol can help researchers to locate the questions, but not necessarily the answers. As we know all too well, interviews don’t always run as expected. Conversations can take an unexpected turn and open up a new area for a researcher to explore. This means that the same interview questions might be addressing different aspects of the problem. Coding the data gives you a way to organize your scripts in such that you can pull the scripts from the same code effectively without looking through the entire interview questions again.


A Quick Guide To Qualitative Coding

Codes are the smallest unit of text that conveys the same meaning (for the purpose of your research). Codes can be a word, a phrase, or a paragraph, you are in charge of choosing the forms of your codes and sticking with your choice for data consistency.

1. Determine The Type Of Coding Method You Want To Apply Before The Data Collection

There’re two types of coding methods, deductive and inductive.

Deductive coding is the coding method wherein you have developed a codebook as a reference to guide you through the coding process. The codebook will be developed before your data collection starts, usually in the process of researching the existing field. Usually, if you have a general direction in mind, you will be able to develop a rough codebook. Of course, the codebook changes as you code on, new codes will be added and categories re-organized. In the end, your codebook should reflect the structure of your data.

Inductive coding method is used when you know little about the research subject and conducting heuristic or exploratory research. In this case, you don’t have a codebook, you’re building on from scratch based on your data.

The two types of coding method have their own pros and cons, but the end result should be similar. The majority of your data should be coded and be able to form a narrative.

Is it deductive or inductive? (Photo by Artem Sapegin)

CODING IN ACTION

Once you know what types of text you’re coding for, the action of coding is fairly simple — you select text, and give it a code name that captures the essence of the text. Next time when you encounter a text with same meaning, you give it the same code name. Here is an example:

Participant A: I had chicken and rice for lunch.Participant B: I had beef lasagna for dinner and drank some wine.

Depending on what you want to find out, you may code both these two sentences “meals”, in this case, each sentence is a code. You may also code “chicken and rice” and “beef lasagna” as “food” and “wine” as “drink”, notice how this time each phrase is a code and you have two different codes? The detail of codes completely depends on your research question and what you’re trying to get out from the data.

2. Initial Coding

The initial coding process is fast and relatively easy. You just need to read through your data and get familiar with it. At this point, you don’t have to develop sophisticated codes for the data, but rather just an idea of what the overall data looks like. You can try to code sections with a broad code name for future reference, writing down notes as you read is also a good idea.

3. Line-By-Line Coding

As the name suggests, in this stage, you comb through your data with a closer eye. Your codes now should have more details. Try to code everything, even if you know certain codes are not going to make it in the endgame. Your analysis of the data will become more profound as your codes become more detailed.

4. Categorization

When you’ve finished the line-by-line coding, you’ll usually have a messy collection of codes. This is when you want to put similar codes into the same categories and move them around in order to find out a way that reflects your analysis the best. By analyzing and sorting your codes into categories, you will be able to detect consistent and overarching themes for your data. And within the themes, you can tell the user story.

5. Determine Themes

What story are your codes telling? (Photo by Pedro Nogueira on Unsplash)

The categorization of codes reflects themes. The bigger categories are the overarching themes while the sub-categories supporting themes. This is where you can engage in storytelling from your data. The themes can tell the same story from different perspectives, or several different stories that connect with each other. With great narratives created from the themes, the messy qualitative data are now in a meaningful order.

In a nutshell, coding is the data analysis process that breaks the text down into the smallest units and reorganizes these units into relatable stories. As Christians and Carey suggest, the poetic resonance in a story is what every qualitative UX researcher strike to achieve.

“Qualitative research in its best form seeks through naturalistic observation to set up a poetic resonance with the native interpretation.”

— Clifford G. Christians and James W. Carey

Reference

Christians, C. G., & Carey, J. W. (1989). The logic and aims of qualitative research. Research methods in mass communication, 354–374.

ProjectUX

Written by

ProjectUX

ProjectUX is the UX Show for startups. We help startups improve their products through a heuristic review of their UX and a round of usability testing.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade