In Case You Missed This... 'Wisconsin Posts Biggest US Job Loss, As Gov. Scott Walker Fights For His

[titaniumsalute]

14. No it isn't

Weighting isn't the cause of these results. While the Tuesday election IS what counts you cannot dismiss polls because of weighting. Weighting is done in every scientific survey and is necessary because it is a sample universe.

In order to have fairly accurate polling a company must randomly select individuals to ask questions. When you use a random selection process (such as random phone calls) the sample universe never quite matches up to the real universe. Let's say a county has an adult male population of 47% and a female adult population of 53%. But the random sample universe comes in 50%-50%. The male respondents are weighted down to 47% and the female respondents are weighted up to 53%.

They do this for age, ethnicity, gender, socio-economic categories, education levels, etc. They will also do it for Republican versus Democrat. Now what can be disputed is the actual "real" universe of Rep versus Dem percentages within a county or state. If those numbers are off then weighting can impact the results. But weighting happens in all polling...or it would be non-scientific.