You know I prefer Kelley, SnitgrrRl, Barbie and even Late for Supper to Prof K.
It's still not accurate to call what you do "statistically good." You can apply the same procedures to all sorts of things: words in texts, instances of violence or sex or whatever in television shows, to people's opinions, etc. The same logic applies. Pauls exammple of 'counting' who kills who on TeeVee is an example of this sort of logic. All I'm saying is this: Yes I agree that there is a way to claim some sort of validity and reliability regarding what you've found, but you've invoked the authority of statistical logic in order to make generalizable claims about what you see. But, that's just not what you're doing. But, I'd say that you can invoke the arguments deployed by interpretivists and critical theorists of social science and you're all set. Or, you could also claim that you've taken a 'theoretical sample' or a 'snowball' sample. Nonetheless, maintaining that your findings are *generalizable* based on statistical procedures is just plain wrong. They're generalizable, I think, on other grounds--I've already noted this.
A great discussion about this whole debate can be found in Michael Burawoy's _Ethnography Unbound_ Burawoy takes both positivist and interpretivist social science to task in order to reconstruct an approach to social science that enables ethnographers to speak to and make generalizable claims about social structure. Another good overview is Brian Fay's _Social Theory and Political Practice_ I think that you'd really like that book, short and sweet and insists on the practice aspect where the proof of the pudding is in the eating, not simply in the pudding.
Hey ask Frances about that one!! Anyhoo, another one you might like is Frank Hearn's _REason and Freedom in Sociologial Thought_ Hearn shows how the work of social scientists with the Solidarity Movment was an example of the engagement of critical social science with political practice.
Late for Supper