Authour: Michi Komori
Edited by Christine Keene
Problem
Analyzing qualitative data from research is a challenging but necessary task in the innovation process. Data acquired during the Learn phase will reveal insights and challenges that lead to research question development and can often provide clues to potential interventions.
Solution
Thematic analysis can be used to make sense of seemingly unrelated material. It can be used to analyze qualitative information and to systematically gain knowledge and empathy about a person, an interaction, a group, a situation, an organization or a culture.
Background
While thematic analysis has been practised for some time, it has not been subject to a great deal of rigorous scholarly documentation. Typically, learning the process has been passed from one professional to a protégée.
Documenting the methodology was undertaken by Dr. Richard Boyatzis, a professor of organizational behavior at Case Western Reserve University who wrote the book, Qualitative Information: Thematic Analysis and Code Development (1998) which outlines the principles of thematic analysis. Dr. Boyatzis describes thematic analysis as a process for “encoding qualitative information.”
Used When
Applied during problem framing, thematic analysis helps researchers move from a broad reading of the data toward discovering patterns and framing a specific research question.
Used For
Researchers use thematic analysis as a means to gain insight and knowledge from data gathered. The method enables researchers to develop a deeper appreciation for the group or situation they are researching. By using thematic analysis to distill data, researchers determine broad patterns that will allow them to conduct more granular research and analysis. It is highly inductive: themes emerge from the data that is gathered and are not imposed or predetermined by the researcher. Using their findings, researchers can apply a statistical analysis to validate themes. In practice, depending on the context of the research study, thematic analysis could include a bit of grounded theory, positivism, interpretivism and phenomenology.
Process
1) Collect data – Data is gathered from sources that can include:
- field diaries
- observational data
- pictures/video*
- historical data
- questionnaire statements
- transcripts
- audio recording*
*Data from conversations on video and audiotape are transcribed.
2) Coding data – Researchers code the data by hand or through a software program. Typically, the researcher will be coding every two or three lines of text with handles that identify key words, concepts, images and reflections. Coding is an explicit and iterative process in which the researcher will alter and modify the analysis as reflected by the data and as ideas emerge. Coding skills improve with experience. According to Boyatzis, a “good code” is one that captures the qualitative richness of the phenomenon. A code should be clear and concise, clearly stating what it is, its boundaries and how to know it when it occurs. Codes become the foundation for the themes that are going to be used by the researcher.
A good thematic code should address five main elements.
For example:
What am I going to call it?
Authentic leadership
How am I going to define it?
Leaders who are true to themselves.
How am I going to recognize it in the data?
When respondents explicitly say they are authentic leaders who are deeply aware of how they think and behave and are perceived by others as being aware of their own and others’ values/moral perspectives and of high moral character.
What do I want to exclude?
Attributions of authentic leadership to external forces or socialization agents does not qualify.
What is an example?
Boards should choose authentic leaders for character, not charisma; for their values and ability to motivate employees to create genuine value for customers.
3) Code validation – To ensure the integrity of the codes—that is, that they have not been misinterpreted and are free of researcher bias—they should be developed and reviewed by more than one person. The researcher(s) read and re-read the data, double-checking the codes for consistency and validation. The integration of the codes from the data becomes the codebook from which themes emerge.
4) Themes/frameworks identification – From the codebook, the researcher identifies themes and sub-themes: patterns that have emerged from the coded data. Themes can emerge from patterns, such as conversation topics and vocabulary. Other factors could include the frequency of occurrence, occurrence only when certain factors are present, and time of the day, week or month. The researcher needs to be able to define each theme sufficiently so that it is clear to others exactly what the theme is.
5) Information consolidation, finalize theme names – The researcher finalizes the name of each theme, writes its description and illustrates it with a few quotations from the original text to help communicate its meaning to the reader.
Thematic Analysis: Information from Semi-structured interviews has been transcribed. Key quotes have been highlighted, coded and sorted into themes.
References
Boyatzis, R. (1998). Qualitative Information: Thematic Analysis and Code Development, Sage Publications.
Harvard Graduate School of Education. (n.d.). Foundations of Qualitative Research in Education. Retrieved from: http://isites.harvard.edu/icb/icb.do?keyword=qualitative&pageid=icb.page340897
Heading, G., & Traynor, V. (2005) Analyzing Qualitative Data – Coding. Retrieved from: http://www.nswphc.unsw.edu.au/pdf/ShortCourseResMetJun05/Tuesday%20Session%203%20-%20Data%20analysis%20coding%20GH%20&%20VT.pdf
Hernandez, J. R. (2009). Photo-ethnography by People Living in Poverty Near the Northern Border of Mexico. Forum :Qualitative Social Research, 10 (2).
Howitt, D., & Cramer, D. (2007). Thematic Analysis. Research Methods in Psychology. Prentice Hall (2nd ed.). Retrieved from: http://wps.pearsoned.co.uk/ema_uk_he_howitt_resmethpsy_2/77/19811/5071812.cw/index.html