To the Editor:

Polls are based on responses and statistical data using smaller, random samples to draw conclusions about the opinions of much larger populations. This explains how polling companies can use a sampling of 900 people to justify their assertions as to how 130 million people will vote on Nov. 6.

To make the poll statistically representative, a process called “weighting” is applied, which factors a large range of variables, such as party identification, age, race, gender, etc. The various subgroups are assigned a number or value (based on current data), and the polling result for each subgroup is multiplied by its value and then combined to produce the final result that is reported by the media.

• In 2008, Democrats had a 3 percent voter turnout advantage over the Republicans, where 40.3 percent voted as Democrats and 33.3 percent as Republicans.
• In 2012, as reported by Rasmussen Reports, 33.3 percent of the country aligns itself with the Democrat Party and 37.6 percent with the Republican Party.

In this year’s election cycle, however, the integrity of the national media’s polling has become suspect.

According to their own data, the latest New York Times, Quinnipiac, and CBS News polls are oversampling Democrats and using statistical data from 2008, ignoring the 2010 elections and recent polling data.

Consider the following…

In this year’s election, who is more likely to re-consider their vote—people who supported the Democratic Party in 2008 and 2010, or people who support the Republican Party? This year’s election is NOT about any one person or party, it’s about the direction in which the country is being led!

Mark DeVol

Anderson County