What is interview analysis?
Interview analysis is a qualitative research technique that involves analysing data collected from one-on-one or focus group interviews.
The goal is to identify patterns and themes in the responses of the interviewees, usually to inform and guide the development of a product or research.
It is widely used in various fields, such as clinical research, market research, social science research, UX research, educational research, and many others.
In this guide, we will look at how to systematically analyse data from interview transcripts.
Coding of data in research
Coding of qualitative data is a fundamental step in the analysis process. It involves systematically organizing and labeling segments of content to identify patterns, themes, or concepts.
How to code interview transcripts
The coding process is often iterative and goes through multiple stages to progressively move from raw and unstructured data to meaningful insights.
Open coding
Open coding is the initial phase and involves identifying and labeling different ideas, concepts, or phenomena that emerge from the interview transcripts.
Axial coding
Axial coding is a subsequent step that involves organizing and connecting the initial codes into broader categories. It helps to establish relationships and connections between different codes.
Selective coding
Selective coding is the final step and involves identifying overarching patterns or themes that emerge from the categorized data to develop a coherent narrative or explanation.
A coding framework or codebook is often created to guide the organisation and analysis of the data. It can be developed inductively, based on the data itself, or deductively, using pre-existing codes from previous research or theory.
Inductive coding
Inductive analysis involves developing themes and categories from the data itself, without any pre-existing theoretical framework or hypothesis.
It is a bottom-up approach to analysis, where the labels, categories, and themes emerge from the data.
This method is best suited for exploratory research, where the objective is to gain an in-depth understanding of a topic or phenomenon.
How to do inductive analysis of an interview
Transcribe audio
Open Gems, click on the icon, and upload your interview recording to get it transcribed.
Extract key ideas and quotes
Start the coding process by reading the transcript line by line or paragraph by paragraph.
Click on the icon to get AI-generated summaries or the icon to see the key sentences highlighted.
Summaries and key sentences offer a convenient and time-efficient way to look for meaningful units of information, such as specific ideas, concepts, or themes, that you want to assign codes to.
Just select the part of the text you want to extract and click “Copy to New Note”
Label key ideas and quotes with a descriptive word or phrase
Since you are doing inductive analysis, codes have to emerge from the interview data.
Click on the icon to ask AI to label the extracted content. This can speed up the labeling process or just give some ideas to get you started.
Try to apply the codes consistently throughout the transcript and across different transcripts so that it's easier to see major themes emerge in the following steps of the analysis. But don't worry about being too rigorous because AI can help you with grouping too.
Group coded data into broader categories
Click on the icon. AI gets the nuances and semantic similarities within the data and automatically groups the notes of your interview by topic, so you can start seeing the main themes of the conversation.
If you already have some notes on the board from previous conversations, you can also click on the icon. AI groups the notes of your entire research!
If you want to establish relationships and connections for a specific note, you can select the note and click on the icon. AI gives you a list of notes semantically similar to the targeted one, ordered by a similarity score. From there you can manually draw connections.
Of course, none of these groups and connections are set in stone. These are just ways to sift through your codes faster and refine them with an iterative approach in collaboration with AI.
You can move notes and groups on the whiteboard, create new ones, and nest them.
Find the main themes for each group
Finally, select a group and press '?'. AI can describe you what grouped notes have in common and make the underlying themes more explicit to you.
Thematic analysis and narrative analysis are inductive analysis methods for identifying themes and patterns in the data to generate new insights and understandings. Let's look at some use cases.
Thematic research examples
Understanding the voice of employee for a better workplace
You conducted interviews with employees to understand their experiences of workplace communication.
As you analyse the interviews, you discover themes such as:
"ineffective team communication"
"communication barriers across departments"
"need for transparent communication"
"importance of active listening"
These themes are derived directly from the data, allowing you to understand the communication challenges within the workplace and inform strategies for improvement.
Make sense of user interviews and build the right product
You conducted user interviews to explore users' experiences with a mobile banking app.
As you analyse the interviews, you discover themes such as:
"ease of navigation"
"security concerns"
"desire for personalisation"
"clarity of transaction history"
These themes emerge from the users' feedback and experiences, allowing you to gain insights that can guide improvements in the app's user interface, security features, and overall user experience.
Narrative research examples
Investigating the impact of a new policy on a particular population
You have interviewed individuals who have experienced homelessness.
Through narrative analysis, you explore their personal stories, paying attention to the plot, character development, and overarching themes.
You identify common themes such as:
"loss of stability"
"resilience"
"struggles with societal judgment"
"journey towards finding housing"
By examining the narratives, you gain a deeper understanding of the complexities and experiences related to homelessness, which can inform policies and interventions aimed at supporting this population.
Provide tailored care to patients with chronic diseases
You collected patient narratives about their experiences with a specific chronic illness.
Through narrative analysis, you explore the stories and accounts shared by the patients. You identify common narrative themes such as:
"journey of diagnosis"
"emotional impact of the illness"
"challenges in treatment adherence"
"support systems and coping strategies"
By analysing these narratives, you gain insights into the lived experiences of patients, their needs, and the impact of the illness on their daily lives.
This understanding can inform healthcare providers to personalise care delivery.
Deductive coding
Deductive analysis, on the other hand, involves testing a pre-existing hypothesis or theoretical framework by analyzing the data collected.
It is a top-down approach to analysis, where the labels, categories, and themes are predetermined based on a theoretical framework.
This method is best suited for confirmatory research, where the objective is to test a hypothesis or theory.
How to do deductive analysis of an interview
Define the research objective
Clearly establish the research objective or research question that will drive the coding process.
Develop a theoretical framework
Create or select a theoretical framework that aligns with the research objective. This framework should provide a set of predefined codes or categories that will be used to analyse the data.
Identify and code the data
- Click on the icon to identify the key sentences of the transcription
- Select the part of the text that fit the codes and click “Copy to New Note”
- Label it with one of the predefined codes from the theoretical framework
Be consistent and accurate in applying the codes to ensure reliability and validity.
Group and synthesise the coded data
Once the data has been coded, you can still use Gems' AI features to explore the relationships between different notes, but make sure to interpret the findings in relation to the research objective and theoretical framework. Consider how the findings align with existing theories or literature in the field.
Deductive research example
Prove detrimental effects of social media on adolescents
You are conducting a study on the impact of social media on adolescent self-esteem.
Using a deductive approach, you start with the hypothesis that excessive social media use negatively affects adolescent self-esteem.
You define specific categories related to social media behaviours, such as:
“time spent on social media”
“comparison with others”
“feelings of inadequacy”
Then, you conduct interviews with adolescents, asking structured questions aligned with the predefined categories.
During the analysis, you systematically code and analyse the interview transcripts, examining how participants' responses align with the predefined categories and hypotheses.
By consistently finding examples where participants describe the negative impact of excessive social media use on their self-esteem, you validate the initial hypotheses.
The findings can inform interventions and educational programs aimed at promoting healthy social media habits.
How to present interview findings
In this article we have seen how to systematically analyse your interviews following inductive and deductive coding. It is important to note that they are not mutually exclusive and that researchers often use a combination of both approaches in qualitative data analysis.
Once you have identified your key themes, patterns, and insights, you need to effectively communicate your findings in a logical and coherent way.
Support with evidence
Provide vivid and detailed descriptions of participants' experiences and include direct quotes or excerpts from their responses to illustrate and validate the themes or patterns identified. This helps to provide credibility and authenticity to your findings.
Use visual aids
Consider using charts, graphs, or diagrams to present the findings. Visual representations can help highlight patterns, comparisons, or relationships among the data.
Think about your audience
Tailor the presentation or write-up to meet the needs and interests of your audience, focusing on key takeaways, practical implications, or recommendations that may be relevant to them.