Qualitative data analysis occurs any time a researcher analyzes data this is not made up of numbers – images, audio, video and documents can all be considered qualitative data. In this case, your qualitative data is the text of your interviews. This type of qualitative data analysis can be a difficult and time consuming process, but it can take on different forms and different levels of intensity depending on how in-depth you are able and willing to go. For our purposes here, we will just outline a few of the major methods and ideas you can draw from when analyzing interview data. Your main goal in interview data analysis remains the same as any research – you are trying to figure out what is going on, trying to get some answers to your research questions. In interview data analysis, this usually means looking for major themes that emerge from your analysis.
The first thing you need to do is organize your data. Qualitative data can be very messy. Do you have transcripts of your interviews? Audio or video recordings? Your own notes? All of the above? Get these all in order in a way that makes sense for you. There is no one way to go about doing this.
The next thing you have to do is go through these interview data sources – by reading, listening or watching – and use codes as a way to identify major themes. Codes are ways to identify ideas, topics, themes, concepts, words, terms or phrases that appear frequently or seem to be important in your interviews. You might start your analysis with some codes already in mind – this deductive coding will be based on your experience or previous research and theory development others. Or, you may develop these codes as you go through your data, through a process called inductive coding.
Let's move back to our example from the Metamorphosis Project. If you remember, one of the things we were interested in was the extent to which the church strengthened ties between local African American and Latino residents. We were not quite sure what the answer would be since there was little previous research on the topic. As we read through the interview transcripts, we consistently saw that the pastors saw little interaction between African Americans and Latinos, for a variety of reasons.
How did we go about coding this topic? Well, each time we saw a reference to African-American and Latino interactions, we would make a note of it, giving it a name – for instance, Black-Brown relations. After reading through all of the transcripts, we then collected all of the quotes that touched on Black-Brown relations and read through them separately. From there, we could come to some conclusions about how pastors saw the Black-Brown relationship within the Church.
As you look through your own qualitative data, you should come up with a number of these types of codes, giving a codename to each different topic you come across. As you go through your data more than once, you will then see that some codes are similar and might be able to be combined. After going through your data several times (the number is up to you, but three or four times is a good minimum amount), you should be able to come up with a few major themes related to your research questions. Think hard about these major themes, what they mean for your research question and for your work. Write up your thoughts and include some representative quotes from your interviews that help explain how you came up with your themes.
Get Participatory! The analysis stage is another great way to get those under study involved in the research process, especially if they helped craft the research questions and methodology as well. Trained participants could help you code the data and develop major themes. Alternately, you might conduct most of the analysis and ask them to review your conclusions and make comments or criticisms before you proceed further in the research process,
Recently, several computer software programs have been developed that help people code their qualitative data. They usually cost money and can be a bit difficult to get the hang of, but many researchers find that they can be very helpful when dealing with lots of qualitative data that has been transcribed. We won't go through how to use them here, but a few of the most popular are: AtlasTI, Nudist and Ethnograph.