Phases of analysis*
| Phase | Description | Purpose | Timing |
| Initial coding | Line-by-line coding of each segment of data; use of gerunds as a heuristic device to describe data (and ‘to nudge us out of static topics and into enacted processes’) | Explore and interact with the data; begin to explain what it means; generate the ‘bones’ of the analysis | This coding is done as soon as possible after each interview |
| Focused coding | Coding with those codes that appear more frequently in the data | Sift, sort and synthesise large amounts of data; form the ‘skeleton’ of the analysis | After three or four interviews, patterns begin to emerge |
| Theoretical coding | Use theoretical sensitivity to conceptualise relationships between categories; determine the main category(ies) | Create hypotheses for the theory; begin conceptualising theoretical framework | After interviewing most of the participants and there are clearly significant patterns in the data, the theory begins to emerge |
| Theoretical sampling | Gather more data (further empirical data) on the main category(ies) and its(their) properties until no new properties emerge (to saturation) | Elaborate and refine categories; ensure robust categories; clarify relationships between categories; identify variations in the process | Final interviews validate and refine the theory |
| Theory development | Use theoretical sensitivity to raise categories to concepts; ask ‘what is the data a study of?’ | Develop the theoretical framework to explain the process; offer an abstract understanding of the relationships between the core concepts | ‘Member checking’ with participants and members of the same community adds rigour to the theory |
*Aubin D. Unmasking the self as a fallible health professional: a grounded theory study on the psychosocial process of mitigating the negative effects of shame due to mistakes. 2015. University of Alberta press.