Using Codes for Qualitative Research

Qualitative research deals with non-numeric information - the kind of information that can’t be measured by numbers. Compared with qualitative data, quantitative data is easier to analyze.  Qualitative data contains opinions, attitudes, behavior towards certain phenomena and requires finding trends in responses to open-ended questions or interpreting reviews, interviews, social media posts, etc.  How is it possible to identify recurring themes in a possible “wall of text” that is qualitative data?  By coding it.

Coding is a process of analyzing the data by assigning different codes (labels, tags) to it. The codes can be numbers, words, and short phrases. In other words, anything that lets the researcher easily examine the data. That data includes interview transcripts, documents, photographs, artifacts, emails, etc. For instance, let’s check this sentence:

“I would like to have a police officer guarding the apartment I live in.”

This respondent wants to live in a safe environment. Therefore, we can code the entire sentence as “SAFETY.”

By coding the raw data, researchers can organize and categorize the content to draw conclusions. This process allows them to get comprehensive insights into the data they can’t acquire by merely reading it. It helps to distinguish between the essential data and the information that is not relevant to the research.

There are two approaches to coding:

Deductive coding

It means researchers use a predetermined set of codes and assign them to new information.  The source for these codes can be the previous researches in the same field. Using deductive coding is also possible if the researcher already familiar with the research topic. By taking this approach, the researcher predicts the possible answers the respondents will provide.

Inductive coding

In this case, researchers set the codes during the coding process. These codes in inductive coding are not preconceived.  Inductive coding is especially useful when your research topic is new and has never been covered by the other researchers. To put it another way, you are the first person researching that subject.

Researchers often use a combination of these two approaches to cover all available information because it is impossible to forecast all reactions they may get in the process. Thus deductive coding alone won’t work.

There are also different coding concepts. They include:

In Vivo coding

Researchers use terms or phrases participants provide in their feedback, which means the codes are derived directly from the data. 

Values coding

The codes are used to describe values, opinions, motivations, emotions, beliefs, etc. By using these codes, researchers can determine the participants’ worldviews.

Process coding

This method uses gerund (words with “-ing” suffix) to display specific actions.

Descriptive coding

In this case, code is often a noun and summarizes the idea expressed in the text.

Simultaneous coding

The method involves applying several codes to part of the text when it contains different meanings.

Coding enables analyzing the data in a structured way. It is easier to examine connections between the codes than the long passages.  It is possible to code manually by using different-colored pens and highlighting parts of the data. Using the software is often beneficial because it offers features that can’t be replicated by manual coding. Tools such as MAXQDA, Atlas.ti, NVivo, Transana, QDA Miner are worth checking to ease the coding process.