Posted on 10/7/2016, 2:34:02 PM by Signalman
“Put not your trust in princes” as Psalm 146 advises… nor put such trust in princesses or polls. For “princess,” awaiting a queenly coronation, read Hillary C and for “polls” include any of those focused on the current presidential race.
In point of fact, as the Brits would say, the proper way to read the current presidential polls may be with a modicum of cryptographic decoding and guided by a few essential considerations.
The first is the obvious fact that a respondent answering the pollster may not rank toward the top of the ingenuousness scale. That is to say, he (she) may be directly lying about, or incorrectly predicting, his (her) ultimate vote… and that includes, as well, the possibility of not voting at all.
The second pertinent fact is that such inauthentic self-reports are not randomly distributed so as to generate mere statistical “noise” of the sort reflected in the reported “error limits” of the particular poll. To the contrary, the majority of those who misinform the interviewer may go in one direction rather than the other, whether the choice is between classical or country music or between Hillary Clinton or Donald Trump.
From what precedes there follows this crucial assertion: considerably more pro-Clinton respondents in the current polls will, in ultimate fact, vote for Trump than will “Trumpers” vote for Clinton. A much smaller number of others will vote for the Libertarian candidate or won’t vote at all. Does this prediction have a basis in any reality known to pollsters, sociologues, and social psychologists or is it mere wishful conjecture? It does indeed have a very potent basis. To use the jargon of the social psychologist, that basis is now called “Evaluation Apprehension.”
(Excerpt) Read more at spectator.org …