Grounded theory is an approach by which theory is extended from qualitative analysis (Charmaz, 1990; Walsh, 2014). It began nearly 5 decades ago (Glaser & Straus, 1967) and has since developed and diversified (Heath & Cowley, 2004). This article outlines a process of thematic analysis directed by the grounded-theory approach and discusses the conditions under which this process is most suitable, using examples from my work with a research team on my master’s thesis about gender-role conceptions among Latinas (Heydarian, 2016).
The use of thematic analysis driven by grounded theory is particularly informative for this area of cultural research. The prominent literature of Latina gender studies in the social sciences promotes a stereotypical image of Latinas as submissive and dependent; the grounded-theory approach to thematic analysis allowed me to explore the detail and nuances of how Latina women themselves describe the Latina experience. From my own analyses, I found that Latinas view the experience of being a woman in Latina culture as a complex identity beyond stereotypes. The study participants noted that their identity changes and evolves in different situations and across the lifespan. These findings have implications for how Latinas are viewed and treated in social-science research, setting the stage for future directions in sociocultural and clinical studies.
Grounded Theory in Data Collection
Grounded theory is an approach whereby the researcher refers back to the literature relevant to the research topic and to qualitative observations throughout data collection and analysis. Review of the literature and qualitative data can help shape subsequent data collection and analysis according to new perspectives that arise from reference to previous research and participants’ observations. During the data-collection stage, the researcher may realize previously unanticipated characteristics of the construct by analyzing participants’ responses and consequently refine subsequent data collection.
Grounded Theory in Thematic Analysis
The grounded-theory approach also may be applied to the data-analysis stage of a study. This process involves the critical review of responses to determine appropriate coding and the formation of themes from those codes. Researchers can conduct thematic analyses on the transcriptions of participants’ responses to interview questions, other dialogue, or responses to open-ended questions (Braun & Clarke, 2006; Pope & Mays, 1995). I examined responses to the question “What it is like to be ‘feminine’ and ‘motherly?’” from a semistructured interview.
Preparing and Revising the Codebook
The researcher first develops a preliminary codebook — a predetermined set of constructs and their associated definitions and characteristics. (This codebook will be refined throughout analysis.) This is determined a priori from the existing literature, the proposed research questions, and consultations with experts familiar with the constructs of interest. For example, one construct that emerged in my study of Latinas’ perspectives of gender roles was familismo — prioritizing, providing for, and taking care of the family (Castillo, Perez, Castillo, & Ghosheh, 2010; Guzmán, 2011; Heydarian, 2016; Lugo-Steidel & Contreras, 2003). Codes initially assigned to one theme may be moved to another theme during later stages of the analysis.
The researcher then will need to select certain themes to report. This selection is based on what the researcher determines to be the smallest-sized theme of interest for answering the research question (i.e., what is the smallest number of people who gave a response that fits within that theme?) and the practices of the field. The researcher also may choose to highlight themes of particular theoretical interest.
During this final stage of coding, subthemes may be identified. These may emerge when several participants give similar detailed descriptions of a characteristic of the theme. For example, a subtheme of the theme familismo may include taking care of children (Heydarian, 2016).
The Coding Process
The coding process entails reading through the data and list of response codes, referring back to the original interview transcriptions, and reassigning response codes to different themes that best represent them. Ideally, the researcher should analyze the data with a team of two or more research assistants familiar with the codebook and coding procedures.
After response codes are identified, researchers can sort them into themes. Both theory-driven (deductive) and data-driven (inductive) stages of analysis can be used to generate themes from the response codes. The researcher and research assistants independently examine the response-code data for theory-driven themes according to the codebook, then meet to resolve coding discrepancies and identify quotes that did not fit within the theory-driven themes. Then the members of the research team independently can examine the response codes that did not fit within the predetermined deductive themes and identify new, inductively derived themes. It is important for the raters to carry out this stage independently so that their interpretations of the data are not influenced by others. The team constructs new themes that are not described by previous literature, with corresponding definitions to capture the prevalent characteristics described by the participants. For example, one previously unidentified construct associated with marianismo — the constellation of stereotypes associated with women in Latina culture — is empowerment (Heydarian, 2016). Our research team identified an internal empowerment theme and an external empowerment theme. Internal empowerment refers to the sense of a strong identity and self-confidence; external empowerment refers to the desire and self-efficacy to make a positive change in one’s own life and in the community.
The research team will meet again following the second stage of independent coding to consult on the quotes that were not assigned to either the deductive theme or the inductive theme. After the discussion of possible inductive themes, the primary researcher reviews all of the coding and arrives upon a final codebook.
Limitations and Strengths
The grounded-theory approach to qualitative data analysis is heavily directed by the primary researcher. This element of the approach can introduce bias into the analysis. The primary researcher must carefully consider the perspectives of the
research-team members and the research participants by revisiting the data several times when revising the codebook. The research team that I worked with for the study on gender-role perspectives of Latinas contributed greatly to shaping the codebook and findings of the study, and ultimately helped contribute to the field.
The grounded-theory approach is useful when the area of study is new. It also is helpful for identifying details of constructs. In addition to themes and subthemes related to familismo and empowerment, we discovered themes capturing perspectives about beauty, interpersonal manners, and human qualities (e.g., being loving and caring). When the researcher carefully considers other perspectives and is well versed in the existing literature related to the research topic, the analysis can make a great contribution to shaping theory.
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3, 77–101.
Castillo, L. G., Perez, F. V., Castillo, R., & Ghosheh, M. R. (2010). Construction and initial validation of the Marianismo Beliefs Scale. Counseling Psychology Quarterly, 23, 163–175. doi:10.1080/09515071003776036
Charmaz, K. (1990). ‘Discovering’ chronic illness: Using grounded theory. Social Science Medicine, 30, 1161–1172.
Glaser, B., & Strauss, A. (1967). The discovery of grounded theory. Hawthorne, NY: Aldine Publishing Company.
Guzmán, C. E. (2011). Toward a new conceptualization of marianismo: Validation of the Guzmán Marianismo Inventory (Unpublished doctoral dissertation). Retrieved from ProQuest. (3534136)
Heydarian, N. M. (2016). Perspectives of feminine cultural gender role values from Latina leaders and community residents (Unpublished master’s thesis). Retrieved from ProQuest.
Heath, H., & Cowley, S. (2004). Developing a grounded theory approach: A comparison of Glasser and Strauss. International Journal of Nursing Studies, 41, 141–150.
Lugo-Steidel, A. G., & Contreras, J. M. (2003). A new familism scale for use with Latino populations. Hispanic Journal of Behavioral Sciences, 25, 312–330.
Pope, C., & Mays, N. (1995). Reaching the parts other methods cannot reach: An introduction to qualitative methods in health and health services research. Journal of Behavioral Medicine, 311, 42–45.
Walsh, I. (2014). Using grounded theory to avoid research misconduct in management science. Grounded Theory Review, 13. Retrieved from http://groundedtheoryreview.com/2014/06/22/using-grounded-theory-to-avoid-research misconduct-in-management-science/