Survey as problematic methodology

It surprises me how problematic surveys are.  Of course, I have a hard time in accepting the objectivity of any description of so-called universal truths.  I think that a researcher can do everything she can to randomize samples and still not end up with an untainted, laboratory situation.  I think it's impossible to do much more than generalize the thoughts of an individual and that it is an ethical imperative to note and explore the vagaries of producing research texts. I have trouble accepting that any essay does more than provide me with an opportunity to create an author.  Of course, then, I'm going to have trouble with surveys.  

I'm not saying, however--or at least trying not to--that surveys aren't valuable.  I do think that a conscientiously crafted and disseminated survey can lead to some interesting observations.  However, like I said above, human beings have a complex relationship with text and a survey is no different.

In more specific response to Jen's prompt, I think that the above is one of my pet ideas that I would have to take into account were I to try for positivistic survey research.  In essence, my belief in the ultimate subjective relativity of all meaning means that I can make survey responses say whatever I want them to say.  However, I think I might combat this with a more qualitative approach to surveying--giving the respondents more space and voice in the results of the survey.

I don't know if this makes any sense at all.  I think my biggest shortcoming with a survey is that I probably won't ever conduct one.