Identifying initial concepts
We need to code the extracts of all our interviews so that we can begin to develop ideas that will form our conclusion. When we've coded all our examples of a particular theme begin to see which ideas help to describe these concepts. Remember we're looking for patterns, and our aim will be to explain the patterns. This is always an iterative process which involves to-ing and fro-ing between the data, the codes, the evaluation brief and our thinking and making sense of the data. It is very helpful to do this with someone else, a colleague who is helping with the evaluation, a supervisor or someone who has experience of analysing qualitative data.
It is important to try to identify links that explain the relationship between themes. This is one way to develop ideas that will help you to answer the evaluation questions.
e.g. In our example where we've coded for timing and financial cost to the patient we might check all the other examples in interviews or focus groups and explore a number of relationships between these two themes-any of the following might be possible relationships between the two codes.
Some timings (such as early clinic appointments, or lengthy waits) are more costly to some patients.
Some timings (such as clinic appointments shortly after regular bus route arrivals) are less financially costly to some patients.
Some patients try to reduce their financial costs by missing their timed appointments.
Car parking costs influence whether patients arrive on time.
No relationship between late opening of the clinic doors and financial cost to patient.
If we find examples that seem to stand out from patterns that we find, we can call these deviations and look for explanations that fit the deviations as well as the patterns.
