If news reports are to be believed, 54% of Republicans believe that President Obama is a Muslim [and 100% of them, I assume, consider being a Muslim an especially bad thing.] When I read reports like this, I despair for my fellow homo sapiens. The scores of millions of Americans presumably represented by the poll respondents hold critical jobs -- as traffic policemen, as bus drivers, as doctors, as lawyers, as chicken pluckers. If the polls are to be believed, a sizeable fraction of the cars approaching me on Interstate 40 at a combined speed of 160 miles an hour are, especially here in North Carolina, driven by motorists completely unhinged from reality. Is it safe for me to drive?
Thus troubled, I looked within for reassurance. Deep in the far recesses of my memory I found a faint trace of an article written almost seventy years ago by two of the great figures of mid-twentieth century American sociology, David Riesman and Nathan Glazer. I am sure those names are completely unknown to you, although you may be familiar with some of the terms they gave to our conversation about public affairs -- "other-directed, "inner-directed," "inside dopester."
With remarkably little effort, I located this essay by means of Google and a few key words: "The Meaning of Opinion," by David Riesman and Nathan Glazer, Public Opinion Quarterly, Volume 12,m No. 4 [Winter 1948-40], pp. 633-648. Read it! It is so far superior to anything written by sociologists and public opinion pollsters today as to take one's breath away.
How can it be that 54% of Republicans think Obama is a Muslim? The answer -- not simple at all -- is that public opinion polling is a socio-psycho-dynamically complex interaction between the poll-taker and the respondent in which the manifest content of the question and answer are a very imperfect representation of the latent interactive processes taking place in the polling.
In the simplest terms possible, I suggest that the answer to my despairing question is this: When a pollster asks a respondent the manifest question "Is President Obama a Muslim?," the respondent at some level experiences this as the quite different latent question, "Do you like President Obama?" The respondent understands quite well, even if not consciously, that to give the patently true answer "No" to the manifest question would actually be to give the answer "Yes" to the latent question. So the respondent answers "Yes" to the manifest question, not wanting to be trapped into expressing any sort of support or sympathy for Obama. The poll taker dutifully records this as a "yes" to the manifest question rather thasn what it really is, a "No" to the latent question.
I am quite confident that if a polling organization were to ask a statistically representative sample of Republicans "Does President Obama have horns?," a significant percentage of respondents would say "Yes," even though all of them have seen Obama on television many times and know quite well that he has no horns.
Riesman and Glazer would have understood this perfectly.