Critical Ethnography and Generalizability

I thought Hammersley's article was useful in the way that it unpacked what seems to be a very loaded term. He outlined the older, anthropological use of the term and contrasted it with modern interpretation, including:
-the ethnography may include qualitative and quantitative data (3)
-ethnography doesn't necessarily involve "living with" your subjects (4)
-ethnography is now more likely to take place over months than years, assisted by more advanced data collection technology like audio and video recorders (5)

He also brought up some of the risks inherent in ethnographic research, such as overgeneralizing results to assume they characterize "typical" activity (5). Later, he notes that critics of ethnography have charged it with "only documenting the surface of events in particular local settings, rather than seeking to understand the deeper social forces that shape the whole society, and that operate within those settings" (7). I was interested in the conflict between these two positions and how and when, whether in ethnography or other research methodologies, we can move from the specific to the general and vice versa.

This conflict is present in Canagarajah's article as well, in that he first acknowledges the risk of the teacher/researcher position and details why his particular subject position (including his status as a native Tamil, bilingual English/Tamil speaker, and progressive professor) might bear on his findings. After examining this a bit, he then moves to assert the relevance of his data by stating, "Although the uniqueness of each teacher/researcher-student interaction should not be slighted in favor of the generalizability of this study, we have to note that almost all Sri Lankan ESOL teachers are Westernized, middle-class, bilingual, native Lankans like me" (620-621). Thus even though he is cautious about generalizing, he wants the reader to know that this data might be common across the experience of many Sri Lankan teachers.

In approach the subject of critical ethnography, the questions then become: Can one design a study that is both "micro" in its execution (i.e. involving specific and close analysis), yet allow the researcher to make larger claims? Should "larger claims," as I've said clumsily here, ever be a researcher's goal? Does it do a disservice to your subjects to attempt to make larger claims or to shy away from them? In what ways is the subject able to assist in, contest, or question your focus and its "size"?