Jacqueline Marcus: Tracking the National Election Polls: The Biggest Deception Played on Americans

A BUZZFLASH READER CONTRIBUTION
by Jacqueline Marcus

During election years, polls not only present voting percentages, but also they can be media tools of persuasion, depending on how voting polls are skewed. If you're baffled by recent polls that put the two candidates at an even 44% or by polls that put McCain ahead of Obama by 4 points, you may be equally baffled by the number of people who are asked to participate in the voting poll inquiries.

A mere 1,000 people represent the entire voting population in the United States. This is easy enough to discover by checking the footnotes at the bottom of most polls. For instance, this latest CNN/Time poll:

"The CNN/Time poll of 1,022 Americans was taken by Opinion Research Corporation and had a margin of error of 3 percentage points. The Hotline/Diageo poll of 924 voters had a margin of error of 3.2 percentage points." {Reuters; "McCain, Obama All Tied Up"; A CNN/Time poll showed the race deadlocked at 48 percent, and a Hotline/Diageo poll put the two candidates even at 44 percent.}

Are we to believe that 1,022 people represent the entire country of millions and millions of registered voters?!

Question: Where were these polls taken? Were they taken in Republican sectors in the South? Were phone calls made in the Midwestern states or Northeastern states? We don't know because the corporate-pollsters don't provide that information.

By analogy, suppose you were at a football stadium that holds approximately 125,000 people. You ask one person: "Name the team you're voting for." The person says, "Team X". You then send out a report to all the news and cable companies that Team X is the favorite winner based on your poll. And then you produce all sorts of colorful graphs that render the one vote into a national general percentage of voters.

This is precisely why mainstream polls are the biggest deception played on American voters on record. Furthermore, there is a huge contradiction between the polls and the rising numbers of Democratic registered voters: "An estimated 3.5 million new voters have been added to the national voter rolls. As the AP reported this week, 'these figures are up for blacks, women and young people. Rural and city. South and North.' In 17 of the first 24 primaries, voter turnout was the largest in four decades. {Associated Press, 5/5/08; USA Today, 2/29/08}"

At the same time, according to UPI: "The number of registered Republican voters in the United States fell as registered Democrats rose since the last general election, voting experts said. But, almost as often, people are registering to vote without declaring any party affiliation, The New York Times reported Tuesday.

While implications for the general election remain unclear, experts told the Times voting registration numbers may signal a movement away from the GOP that could affect local, state, and national politics for several election cycles.

In battleground states such as Iowa and Nevada, state officials reported more registered Democrats than Republicans, a reversal from 2004. No state officials reported a similar switch to more registered Republicans than Democrats for the same period." {UPI; "Report: Number Of Registered Republicans Falls" August 5, 2008}.

Perhaps the truth lies somewhere within these numbers:

The Polls Are "Dead Even" when you...

1) Subtract the millions of Black votes
2) Subtract the millions of Latino votes
3) Subtract the millions of new Youth votes
4) Subtract the millions of Gay votes
5) Subtract all the Democratic/Liberal votes

A BUZZFLASH READER CONTRIBUTION

Technorati Tags:

Novel concept...

Here's a novel concept, ignore all polling from any source and encourage everyone you know to do the same. I'd like to see a progressive push towards getting this message out to EVERY voter in the US. Maybe a website (www.pollsarebs.org?) or organization against polling in general??? People need to know how polls have ability to bias elections, how there is really no way to insure accuracy or fairness, plus the obvious opportunity for bias that is so often prevalent in so many of these "official polls". They need to accept that polling is nothing more than a tool used by the media as fodder for probably +50% of their pre-election noise, and both political parties to try to bias the vote.

For me it's simple. "Any and all polling results are cow crap, consume at your own risk."

Remember, the primary function of statistics is to process data to obtain the results you desire.

Stalin runs our elections

Whether the polls are right or wrong is irrelevent. It has been 12 years since the winner of our presidential election actually served in that office.

The Republicans aren't bothered by the niceties of morality or ethics or honesty. Their focus is on destroying government, not leading it. Their focus is on gaming the system in any way possible whether it is deleting registerred voters from the rolls, caging voters, riging election machine software, destroying voting records.

They follow Stalin's philosophy. It's not who votes who counts, it's who counts the votes who counts.

Accuracy of polling

Marcus should really inform herself of how a properly drawn sample of 1000 people can give a pretty good idea +/- 3% of public sentiment or knowledge or whatever. These pollsters are not stupid or lazy. A sampling frame has to be carefully drawn to have a good chance of being reliable. Basically the samples are stratified and then within strata respondents are randomly chosen.

There are caveats, of course. In a presidential election what matters in the end is the electoral vote. In 2000 no one disputes that Gore got about a half a million more votes than Bush but the Republicans manhandled the Florida election and the Supreme Court put all those electoral votes into Bush's column.

This year there may be new problems with the reliability of the polls. Many people, and not only young voters, have no landline phone. Many more have caller ID and never answer if they don't recognize the #. No sampling frame can take account of those changes in telephone behavior because the pollsters probably don't have good information about who are the increasing number of people who are not reachable by phone.

Whether some pollsters are also dishonest, I can't say, but the basic techniques of polling are well established.

Colleen Clark
Cambridge, MA

Liars, Damn Liars, and Statisticians

I have been involved in statistical studies, and the most informative thing I learned is that it is considered acceptable to delete data or groups from the tabulation until one gets the result one wants.

I remember when Bill Clinton was running against Bob Dole in 1996, and two weeks before the polls opened, the tabulation changed from "eligible voter preference" to "likely voter preference". This brought the numbers much closer together and made it seem like Dole even had a chance of winning - an example of dropping data to get the desired result for Dole.

It is no different today, except that the GOP is much more blatant about being open with their manipulations. They know that there is no one brave enough to call them on their evil acts. And if someone should, there is always the "right" of Bush to declare such a one an enemy combatant!

Truth within the lies within the truth

Well, yes, the polls subtract the newly registered black voters, and the newly registered young voters, and the voters with mobile phones and no landline.

But, we should consider that appropriate, since most of those voters will be caged by Republican dirty tricks and probably end up voting on provisional ballots that will not be counted.

Since the blacks and youth will not be counted this November, the polls are pretty accurate.

Oh, and if you're black and/or young and wanted to vote? Well, I would suggest you take advantage of the 2nd Ammendment.

Falsified figures standard operating procedure

Is anyone else convinced that the polls are being manipulated to serve their masters just as is everything else? Republicans have been very good at manipulating corporate earnings, CPI, and now accused of falsifying oil inventories (that was the root cause behind the explosion in oil prices). The office of the President means big money to powerful players and it would be easy for them falsify the data and convince the public that momentum is changing. The mantra is win at all cost and anyway possible.

The Validity of Sample Groups of 1,000

"A mere 1,000 people represent the entire voting population in the United States" It so happens that 1,000 - 1,200 is a decent enough sample group and accuracy improves if the polling universe is of likely voters. Why do I always see these kinds of "polls lie" articles in Buzzflash when our guy is losing and not when our guy is ahead?

Ok ya right, let's do some simple math...

126,000,000 - number of people who voted in the November 2004 election.
1000 - average sample size of polling data set.
.008% - sample set/number of voters in 2004.

Sorry but in my humble opinion, basing any results on 8 thousandths of a percent is ridiculous, I don't care how much statistical data you have to justify otherwise. My guess is that the studies that support this position come from the pollsters themselves.

If as an engineer I produced results based on .008% of a data set to justify my results, not only would I be a fool, but I'd be unemployed. And that's in a world of exact measurements and truly definable results, not in the whimsical world of human nature.

"All polls are cow crap, consume at your own risk."

Polls

When your guy is ahead, the polls are true and when behind they are not true,but having said that; I do not trust the polls either way.

poll error margin

I am sympathetic to the hope that Obama is actually better off than polls indicate. The whole idea behind sampling is that if one chooses the sample appropriately, then a small sample can approximate a large population. How closely the sample reflects the character of the population depends on the sample choice more than on its size. We need a truly representative sample to get reliable results. These concerns are hinted at in the article, but not made especially clear. In no case would a sample of size one be considered representative of a population of any size greater than one.