Most people that claim to be using grounded theory do their data analysis with tools such as atlas.ti or maxqda - therefore, I assume, they do some kind of content analysis.

Analysis in grounded theory seems to be foreshadowed by data collection, doesn't it? I mean you can't begin "data collection" unless you have ideas about what needs to be known, and therefore what might be important to collect. This "framing" of what data needs to be collected is preceded by thoughts about what questions are important to the researcher, and what implicit hypotheses are embedded in those questions. So, my intuitive take on how to begin data analysis when using grounded theory, is to observe what has been collected, and as Phillip has suggested, begin with counts of what has been collected. I think next, might come observing associations of what has been collected. Then characterizing the associations might follow. I would hope Dr. Glaser's work has evolved so that some semblance of a methodology has been formulated that is independently verifiable, and that doesn't require his personal consultation to qualify that work as work done using his grounded theory. Or is the method still "emerging"? http://www.groundedtheory.com/

There is a very good chapter in the Sage Handbook of Grounded Theory, describing the logic of inquiry. I used the pragmatic stance for my PhD research, as describerd by Jörg Strübing. See chapter 3 of my book Building an evidence based practical guide to Large Scale Interventions. Full text available here.

https://www.researchgate.net/profile/Tonnie_Zouwen/publications/?pubType=book&ev=prf_pubs_book

I think answers to this question depend on some of the context of the project e.g. is there a clear question or is it really exploratory? what type of grounded theory is preferred? how much time do you have? Some of the experiences and lessons I faced in my PhD are captured in this paper (attached).

From my experience of grounded theory I'd move away from doing something like content analysis or being overly concerned about counting how often particular phrases/words/concepts are mentioned - although this can be important. I think it is more important to try to keep track of how your understanding is developing between interviews, what assumptions are being revealed/challenged, what new questions are sparked by these interviews, etc. Perhaps keep a diary and memos of these developments. Like @EdwinHuff I believe the analysis starts before the start of the project, not when you start the first interview, and hopefully not after you've collected all your data as this misses the iterative nature of data gathering and analysis which is advised in grounded theory (although some claim to use grounded theory even after they have collected all their data).

If it's a big project I'd recommend QDA tool like Atlas.ti or NVivo, if it's small this might not be necessary. I like Charmaz's advice on having an initial coding phase and a focused coding phase - the first phases you don't really know what you're looking for, after some cycles of data gathering and analysis you at least have more direction for focused coding.

Conference Paper Confessions from a Grounded Theory PhD: Experiences and lessons learnt

I would say if you have a lot of data (I had roughly 60 interviews) what helps is to have your research question in mind and focus on material that provides answers to these question. I excluded a considerable amount of data this way and focused my analysis. Of course you can change your research question during the study as often suggested by hardcore GT people but pragmatically speaking only if your study context allows it (what your Prof. or funder supports, time left etc.)

Strictly speaking, Grounded Theory is "a SPECIFIC methodology developed by Glaser & Strauss in 1967 for the purpose of building theory" (Strauss & Corbin, 2008, p.1). It includes fairly clear instructions on data analysis (for a qualitative research design), which include at least the following three steps:

- open coding

- axial coding

- selective coding

Having used grounded theory exhaustively in my own research and being an editor of a journal that encourages qualitative research, the term "Grounded Theory" is often wrongly used, namely for research approaches/designs that are in fact not "Grounded Theory". This is not to say that this kind of research is worse than other, but it is simply not "Grounded Theory".

These techniques can, of course, be complimented (better say followed) by quantitative content analysis to provide additional insights (which results in a truly mixed-method approach) using software as mentioned above (nVivo, atlas.ti, etc.). Of course, these software tools support Grounded Theory approaches, since they support coding techniques.

A confusion that often seems to emerge with authors trying to get their work published is: Grounded Theory is not an "adventure" without a specific goal or target. Of course, grounded theory assumes that the researcher has no pre-assumptions about a specific social phenomenon that one looks at analytically. Ideally, a researcher pursuing a grounded theory approach should have no knowledge about the object of research (if we take it to the extreme). Still, you need to have an idea of what you are researching and why - most of the time (99.9%) resulting in a research question. Clear goals and purpose should always drive research (after all we are required to develop knowledge that is useful to society - hopefully :)!).

Even Corbin & Strauss (2008, p. 24) make it very clear that "all research inquiries necessitate a question of some sort to guide the inquiry." In short: Every research (even those exploratory in nature) should have a research question. Pragmatically speaking: If your paper lacks clarity in purpose, it will be rejected by reviewers/supervisors (no matter whether this is your master thesis, PhD thesis or journal article).

I have seen in the past that some authors tend to use specific paradigms as excuses ("I use an ethnographic approach so there is no structure to my research, because it is driven by the context and dynamic"). This is not true. Conducting research without a clear research question is neither efficient nor does it represent a rigour approach to research. All research has direction and purpose. Very good research always has it.

If you plan to undertake a study using Grounded Theory, I strongly recommend to read the following book (or any updated version):

Corbin, J. & Strauss, A. (2008). Bascis of qualitativer research: Techniques and Procedures for Developing Grounded Theory. Sage Publications.

I hope this provides a useful starting point to carry out your own research. If you are interested in another example of where grounded theory was applied together with several other research methods (mixed-methods), you might wish to take a look at the methods section of the following study:

http://ssrn.com/abstract=1743962 or http://dx.doi.org/10.2139/ssrn.1743962