George Rebane
We must always remember that more often than not the public mind is fickle, frantic, and foolish.
No one can escape the polls during this election season whose results will be deemed cataclysmic by at least half of the American electorate, no matter who wins. So all kinds of people from government, campaigns, and academia are attempting to peer into the collective public mind to determine its sentiment and portents. The instrument of choice for this endeavor is the poll. A poll is a type of survey which consists of applying a questionnaire to a sample of respondents drawn randomly from the population of interest.
Polls are abundant, there being at least twenty of them prominently quoted on a daily basis. To make sense of the plethora of divergent results, desperate people attempt to amalgamate them by various kinds of averaging each of which can stand about as much scrutiny as the legislature making laws, or worse, a bureaucratic brew drafting regulations. Here I want to disabuse the reader of giving much credence to any of the, say, ten polls to which he is subjected on any given day. The poll’s result is usually given by a set of percentages attached to alternative answers, percentages that are supposed to sum to 100% so that the waterfront is appropriately covered.
To add credibility to the poll, the media also often include the ‘margin of error’ and the number of people questioned or the sample size. As an article of faith the poll’s consumer is supposed to believe that the population was properly sampled and that the questionnaire was not loaded with agenda-driven queries meticulously wordsmithed. For the present exercise let’s all join that body of the faithful and concentrate only on the claimed credibility bestowed by stated sample size and margin of error.
[To continue reading this piece which goes on to contain numerical tables and graphics, please Download Polling Phollies ]
Not quite on topic about the science of polling (Statistics 202, IMHO), but I have always liked Michael Barone. Interesting data what he says about Hispanics. You would think as a whole the Hispanic/Latino/Chicano vote would be more engaged than ever in 2016.
Go figure.
http://www.washingtonexaminer.com/why-the-polls-are-tightening-up/article/2601206
Posted by: Bill Tozer | 09 September 2016 at 09:44 PM