Name:   MAJ USA RET - Subject:   Don't overextend here - 'splained Date:   8/10/2012 9:33:40 AM

A poll is a statistical sampling of a representative number of people in order to accurately describe the current average condition.  SO, why can you not rely on polls?

a) They are not so random – the very methodology for selecting phone numbers immediately precludes randomness (i.e. home phones versus cell phones).  The polling industry is also compromised by the slick, call center, fake pollsters who sell free cruises and free dinners.  People who DO NOT HANG UP are not the brightest political thinkers. The rest of us folks hang up and go back to supper with the family. Does this not fly in the face of randomness?

b) They fail to be precise - (What good is +/- 4%?) The pollsters simply do not have enough time to contact enough people to “bring their shot group in tight”.  They have to “go to press” so quickly that they just can’t do a truly representative sample.  Polling is all about media.

Rasmussen could do this right if they only reported once a month.  RealClearPolitics further muddies the polls by adding noise and then averaging the results (as a geophysicist, I’d be fired for doing that).

c) They don’t sample enough voters (or the respondents are not truthful) See b) above. 
 

d) They don’t ask a valid question – If the question is couched in innuendo, and the respondent's preferred response is not available in the choices… it isn’t a valid question.

e) The responses don’t measure the condition being surveyed. – If folks are asked whether they like big furry brown dogs, you cannot conclude that they do not like Shetland sheepdogs. Lies are not winnowed from truths.  If my dog could talk, I’d let her answer the pollster’s questions.

BOTTOM LINe: Martini is RIGHT! The polls benefit the media... not you and me. Polls are of no use to us… none. Unless of course, they make you feel good about something.