Actually, just writing that has made me think that I did have an advantage in not having too many signposts - it made my observations more intuitive and natural.  When I just went into a college to be a student and see what I noticed without any preconceptions (other than a feeling that there was more to it than deep and surface learning), I suppose I was engaged in a sort of grounded theory.  This was suggested to me at the time, in a casual conversation with a colleague.  When I looked into grounded theory, I decided it wasn't for me, particularly as the internal debates could have sidetracked me from my own debates with phenomenography.  It also seemed to demand a particular approach to data analysis that I wasn't sure about. But I still might have been doing something that could usefully be described as grounded theory.  In fact, the links that Robson makes between ethnography, case study and grounded theory (Page 190) might have reassured me.

It was very late in the day before I realised the significance of activity theory for my research (dangerously late!) And because of a reluctance to tell the story out of sequence, this meant that the role of "action" wasn't highlighted upfront early enough for the reader.  An earlier orientation to different approaches to inquiry and theoretical perspectives might have helped with this.

So perhaps I can now try a different type of study but with some similar themes to my last one.  I'm starting to think about taking myself out of the picture (as far as possible - but I'm not convinced that any researcher can do this totally!) I could use one or more of my earlier conclusions as a hypothesis and find a way of seeing what differences there might be online and face to face.  Perhaps I'd like to explore something around "learning outcomes" - and the relationships between intended, perceived, actual, additional and unintentional outcomes (and there'll be others perhaps).