Strange that some folks would give a rip what a pollster - a complete stranger - thinks of their views.
I agree with you. My take on the "Bradley Effect", if there is really such a thing, is not that the respondent LIES to the pollster, but that these (presumably racist) respondents are harder to get hold of than the typical respondent. Pollsters already have to use several filters to get the sample they want (eg. party affiliation, sex, age). But how do you have a filter for racism? Obviously, you don't. So if there is any correlation between racists and willingness to respond to pollsters, the poll will be skewed.
No kidding. And largely by interviews over the phone even. I think the "Bradley effect" is getting a rather large head. Regardless, I'm hopeful that this amendment will pass easily.