The role of ethnographic studies in empirical software engineering
Helen Sharp and Yvonne Dittrich and Cleidson R. B. de Souza
IEEE TSE 42:8, August 2016
Pages: 786 - 804

If you just skim this paper, you won't get much from it.  It is written in
a wordy and vague style; it does not give concrete examples; and it does
not give actionable suggestions.  The suggestion that permeates the paper
is, "Researchers should do more ethnographic studies," but because of the
way it is written, this paper is not a very good advertisement for that
suggestion.  There isn't much guidance, either; what does one learn from
"Making such compromises is not easy or straightforward, and there remain
many challenges to designing an ethnographic study for software practice."?

The paper gives 8 examples of ethnographic studes (in Table I), but even
the "implications for empirical software engineering" column is not very
specific in terms of the actionable outcomes of the studies.

However, if you take the time when reading the paper to let each of its
points prompt your own thinking about that topic, and you do your own
synthesis of conclusions, then the paper's wide-ranging discussion can be
useful.  So, think about the paper, what it provides, and what you like and
dislike about it.

The purpose of ethnography is to study a culture, describing it from the
point of view of its members rather than from an outsider's perspective.
It tries to understand not just what is done, but why things are done the
way they are.  The main outcomes are a comparison between the foreign
culture and the researcher's home culture.

Questions for discussion:
 * what are the costs and benefits?
 * how can one use ideas from ethnography in other types of studies?
   That is, how can one achieve similar benefits without completely
   buying into the full ethnographic methodology?
 * When should ethnography be used?
   Typically, ethnography is a "fishing expedition", where there are not
   concrete questions.  The researcher does not yet know what is or is not
   important.  However, there can be topics on which the ethnographer
   focuses.  The advantage of this non-focused perspective is that it may
   be easier to discover the real issues (such as problems in an
   organization) and innovative or unexpected solutions to them.  The
   alternative would be to come in with one particular point of view,
   aiming to solve one specific problem (which might be a real problem but
   might not be the most important problem).
 * How should one perform an ethnographic study?
   What are common mistakes?
   (The paper isn't helpful regarding this.)
 * What is the relationship between ethnography and requirements
   elicitation?  and market research? 

Many people think of an ethnography as a very time-intensive activity
(scientists embedded in a culture for months or years at a time), but this
paper emphasizes that it can be done more quickly -- "a couple of weeks is
enough time to produce good results" and maybe even just days or hours.

The paper says that both rigor and flexibility are required, but doesn't
give metrics for determining rigor.

===========================================================================

LocalWords:  Dittrich Cleidson de Souza TSE