The process for analyzing in-depth interviews is usually based on qualitative analysis, or a review carried out without the use of statistical methods. Examples of non-statistical analysis techniques include content analysis, reviews of case studies, classification by common traits, and logical analysis.
Content analysis is one of the most common processes used to condense, synthesize, and analyze raw data into distinct categories or themes, based on inference and interpretation. It is particularly useful for taking unstructured data and creating a logical structure that can inform and provide deep insights into the “why” questions, which aren’t always evident when conducting basic statistical research.
While there are many valid methods for conducting an analysis of IDI data, the following steps have worked well at 4K Research.
1. Conduct an initial review of the data to identify key themes or categories. Many of these categories likely have already been developed during the formation of the interview guide, but any additional topics that have been uncovered during the research should be added.
2. Begin slotting comments from each interview into these categories. Each comment should have an identifying number and code(s) attached, so you can easily track it back to the person who said it, and quickly identify where they “sit” in the value chain.
3. As you’re conducting the review, look for an underlying structure, in terms of identifying which topics are broad themes, and which topics can be positioned as sub-topics. This will help ensure that the end analysis makes sense thematically and structurally (e.g. you don’t want to create 456 separate topic categories with equal weight).
4. Once all of the comments from the interviews have been categorized, look for patterns in the data, and key terms, concepts, or viewpoints that appear frequently. Pay particular attention to where the comments are coming from within the value chain. Look at the seniority or experience level attached to the comments. From these additional coding insights, you can then map out some overarching themes, sub-topics, and supporting information. Once these have been mapped out, some logical conclusions and insights should jump off the page. It’s often helpful to create separate documents for each topic, so that you can visualize the scope and depth of the insight, and compare them between categories.
The format of the end report likely will vary, depending upon the audience, the function of the research, and type of information required. But if the information has been logically organized and coded, it should be a fairly straightforward process for porting those insights to most types of reports.
No matter the type of report developed, a few final thoughts should be kept in mind:
· Make sure a logical, cohesive story guides the reader through the data, and then leads to a set of conclusions based on that data.
· Direct quotations should be used sparingly to show or support your analysis. Simply loading the report with direct quotes may muddle the underlying themes, since most people do not speak in a clear and consistent manner most of the time, sometimes contradicting themselves even within the same sentence.
· Any insights and recommendations should be drawn from not only what was said, but who said it, and taking into account any internal (such as limitations of the interview process itself) or external (including current events, the economy, etc.) factors or variables that may impact the responses.