Software engineering is a social science, so we need social methods. Grounded theory has been getting more popular over the last decade:
But, do SE researchers use it correctly? Klaas-Jan studied 98 papers from 9 journals, that either mentioned GT or use their methods.
There were two groups as you can see in the figure, 46 are using “GT techniques”, and 52 are explicitly claiming the use of GT. A few findings.
Not everyone knows this, but there are different versions of GT and they for example prescribe different coding practices.
75% of papers do not mention what version they are using, which seems to indicate an unawareness of the existing difference. Also, GT is still evolving, and the focus of Straussian GT is still shifting.
What is a theory?
Klaas-Jan found that very few papers actually found a theory! Only 8 studies presented an actual study, with constructs and relationships. Several others had alternative answers to research questions.
Why should you care?
Well, authors should be way more rigorous in describing their methods. And, grounded theory is well-known, but not well understood. This may lead to a bad name for GT and quantitative research methods in general. Top help you, Klaas-Jan made a checklist:
A preprint is here.