Grounded theory is an approach to qualitative research in social science which posits that the theoretical constructs should be grounded in the observational data and emerge from them rather than be imposed on them. Thus hypothesis formulation takes a more iterative and formative tone compared with the classical scientific paradigm (at least if viewed simplistically).
It is felt to be particularly useful in areas where there are no convincing prior theories of "why things are" (or are not) - and many would feel that this applies to the area of e-learning take-up in universities. where many theories have come to grief, often at considerable cost.
Grounded theory emerged in the 1960s from the work of sociologists Glaser and Strauss; but has since split into several schools of thought which diverge and then sometimes later converge. A partial synthesis of recent thinking has been undertaken by Charmaz. An element of most of the approaches is the systematic use of coding of observational data, with impassioned debate about the extent to which coding schemes should themselves emerge from the data or be derived from more general considerations.
Grounded theory approaches have been used in a number of evaluative studies in library science and e-learning. In particular Google records nearly 500 hits for "grounded theory" AND e-learning within the UK academic community.
Within benchmarking Phase 1, grounded theory was most explicitly visible in the work of the team at the University of Glamorgan - part of the MIT90s cohort - where the research approach was led by Virendra Mistry (formerly of JISC). However, there are other links between benchmarking and grounded theory - and between the lively debates in the two domains.
The systematic approach to codes and tagging of benchmarking entities within this wiki could be taken by some as an hommage to grounded theory - but readers should remember that the credibility of theories can go down as well as go up.
For further reading see the Wikipedia article on Grounded theory.