sleep comes before statistics....and so i will sleep now, for an hour. tomorrow i will get back to posting for real (tomorrow i will also force myself to sit through the state of the union, eeew.)

seminar last night was interesting...a fight between pluralists and third-face-of-power folks over how much power "elites" really dominate in terms of public opinion. i'm getting depressed on this score, because it seems less and less likely that quantitative analysis (especially over only one country...like the US) can ever yield results that differ from the pluralist account of who controls political power (no one, for very long, they say). i just don't believe it, and i've spent all my idle brain time (right, because it's not as if i have stat seminar tonight or anything) since then thinking of empirical questions that might implicate the third face. are there measurable differences between the ways that people perceive policy effects and the way they actually play out?

and then i said to myself, doesn't the discussion we had about survey sampling last night say something about the power of words and symbols to shape opinions? it's well-known that asking the same questions with different words and word orders may yield different results. how is that *not* a proof that language matters, that people, even macro populations, are not so rational after all? when you ask the same question, worded three ways, you don't have to assert that the questions are actually different to get three different results. add active obfuscation to the equation, and i don't see how you can avoid the implication that americans, by and large, can be quite swayed by the manner in which things are said.