All research is "grounded" in data, but few studies produce a "grounded theory." Grounded Theory is an inductive methodology. Although many call Grounded Theory a qualitative method, it is not. It is a general method. It is the systematic generation of theory from systematic research. It is a set of rigorous research procedures leading to the emergence of conceptual categories. These concepts/categories are related to each other as a theoretical explanation of the action(s) that continually resolves the main concern of the participants in a substantive area. Grounded Theory can be used with either qualitative or quantitative data.
Source: Grounded Theory Institute http://www.groundedtheory.org/what-is-gt.aspx
What is it?
An inductive form of qualitative research, introduced by Glaser and Strauss, where data collection and analysis are conducted together. Constant comparison and theoretical sampling are used to support the systematic discovery of theory from the data. Thus theories remain grounded in the observations rather than generated in the abstract. Sampling of cases, settings or respondents is guided by the need to test the limits of developing explanations which are constantly grounded in the data being analysed.
Grounded theory is an approach that develops the theory from the data collected. Rather than applying a theory to the data. This can be a popular approach for people exploring a new area of research. The theory developed from the data can then be tested by further research.
Strauss and Corbin (1990) suggest there are 3 stages in analysis in grounded theory: open coding, axial coding and selective coding. During open coding the researcher reads the text and asks questions to identify codes that are theoretical or analytical. What is going on behind what the person interviewed says rather than just coding literally what is said.
Constant comparison
This involves various methods of constant comparison. Previously coded text also needs to be checked to see if the new codes created are relevant. Constant comparison is a central part of grounded theory. Newly gathered data are continually compared with previously collected data and their coding in order to refine the development of theoretical categories. The purpose is to test emerging ideas that might take the research in new and fruitful directions.
Coding Line-by-line
Another approach used in grounded theory is line-by-line coding. Which literally means coding each line of an interview. This approach is intended to keep the researcher close to the data while forcing them to be analytical. This means the researcher is really having to think about what the person being interviewed is saying and hopefully stop their analysis being influenced by their preconceived ideas or just accepting the point of view the interviewee.
The next step is to check the codes against the text again and see how they can be improved. The codes are also linked with each other and with more general codes.
The next step after this initial line-by-line coding is to refine the actual codes and to link code together in a meaningful way according their importance. So there may be main code with sub-codes relating to that topic.
Source: Online QDA http://onlineqda.hud.ac.uk/methodologies.php#Grounded_theory