Pollsters Are Intentionally Trying to Make Fools Out of American Voters

Posted by Tim Menius on September 21, 2012

It would be difficult for any American with a pulse to not be scratching their heads over the scatter shot results of presidential polls. Many of the recent polls, especially those who now provide the headlines to the unquestionably left leaning media, have been conducted by such political hack organizations as the New York Times, the Washington Post, NBC, ABC, CBS and Pew Research. Their “polls” are regularly scientific, mathematical, and statistical FRAUDS! Let’s explore, in as simple terms as possible, HOW and WHY they accomplish their mission for the Marxist Obama Regime. This information will lead us to one possible, very frightening theory as to why this kind of propaganda is not merely useful to the Left, but is actually a necessity.

The explanation that follows could be attacked by readers who have either an advanced degree in statistics or higher math, and justifiably so. Statistics is an extremely complex science and many of the formulae and equations that must be applied to sampling would look like a foreign language to most people. Accordingly, I’ll take some insignificant liberties with respect to leaving out some of the technicalities involved just to provide to non-statisticians an understandable working knowledge of this process.

Most Peoples’ Understanding of Polling (Statistical Sampling):

In general, when a reader sees the results of a political poll, he assumes facts that simply are false. Note that I didn’t say “not true”, I said false. For example, NBC/WSJ releases a poll that states that Obama leads Romney by a 48 / 44 margin, the sample included 996 adults (or registered voters, or likely voters) with a margin of error of 3%. Most observers would conclude, correctly, that if the poll is statistically valid, if the election were held today, on a final total basis, Obama would receive 4% more votes than Romney. The key here is whether or not the poll is statistically “valid”. Folks, I guarantee you that any poll currently showing either candidate leading by more than 4 or 5 points is statistically INVALID!

Factors Validating or Invalidating a Poll’s Results:

Definition of Sampling Population:

The paramount law of statistics relative to polling is that the population for the poll MUST be correctly defined. The populations in the example above would be either voters, registered voters, or likely voters. Trying to equate poll results for a presidential election to the outcome of that election dictates that the sampling population be comprised of respondents that will actually vote. So, any poll that purports to foretell the election results using anything other than likely voters in its sample is an exercise in idiocy! So, if we want to really gain insight into the temperature of the electorate, we should discount any poll not sampling likely voters. Readers who visit this website know this to be true, but for emphasis, these facts should be known to inquisitive readers. In the presidential elections beginning in 2000, only 54% of “adults” and 69% of “registered voters” actually casts ballots. Under these metrics, a poll sampling “adults” includes 46% of the respondents and a poll sampling “registered voters” includes 31% of the respondents whose opinions are no more relevant than the American political opinions of the citizens of Upper Volta! Diligent observers should take note of the number of left leaning polls, showing Obama doing swimmingly well, that poll anything BUT “likely voters”. Without question, that would include virtually all such “polls”.
Uncompensated Sampling Bias:

This is the area in which the explanation is over-simplified, but the relevant accurate conclusion is unaffected by that simplification. By definition, the sampling units (respondents) MUST be selected “randomly”, meaning that every member of the defined population has precisely the same chance of being selected as every other member of that population regardless, for example, of the party affiliation of any member of the population. Given the requirement of this statistical “commandment”, when the sampling units are selected there is absolutely no way for the pollster to obtain a truly “representative” sample. A representative sample, by definition, must be an accurate reflection of the population selected (as a whole). Theoretically, then, if the population is made up, for example of 37% Democrats, 34% Republicans, and 29% Independents, the sample must be comprised of an identical breakdown of respondents by political party.

Great concept, to be sure, but if the sampling units must be selected randomly, it would take an indescribable stroke of luck to produce a sample that properly reflects the party breakdowns as known in the overall population. And, the truth is, they never do. While this would appear to make the whole sampling process a farce, the statistical and mathematics principles associated with “polling” provide a specific set of (very complex) equations that result in a “bias compensation adjustment (“BCA”)” to the raw number results obtained from the respondents. Properly applied, the extrapolation of the BCA to the raw polling data produces polling results that truly are representative of the population as a whole. And with that, we are introduced to the most important person associated with any poll…the statistician or statistical analyst.

While there are multiple factors that enter into the calculation of the BCA, a number of these factors are significantly subjective. Among these are such things as i) exactly what is the party breakdown of the population as whole today (as opposed to four years ago), ii) what percentage of the various parties will actually go to the polls today (as opposed to four years ago), and so on. It should be intuitively obvious, even to the most casual of observers, that a biased (usually to the left) statistician could significantly alter the results of a poll by his own personal (or his employer, or the commissioner of the poll) biases surrounding the poll results. This reminds me of the old expression dealing with real estate appraisals, MAI, which is a high level designation in the appraisal industry. In practice, MAI was often whispered to mean “Made as Instructed”. This could easily be applied in the polling arena as well.

Although I cannot prove it to be fact, I suspect that at least some of the pollsters now showing Obama up by 6, 8, even 10 percent, are simply disregarding the bias compensation issue altogether. Lest we think this adjustment is not really that important, suppose we conducted a poll that was produced from a “failed sample”. In this particular poll, while we selected all sampling units randomly, through merely an inexplicable freak of nature, the respondents included 85% Democrats, 5% Republicans, and 10% Independents. Would anyone with a brain actually believe the Obama is leading Romney by over 60 points? Of course not, and this is an extreme example, but one that demonstrates the vitality of the bias compensation.

Why are Intentionally Biased Polls Beneficial to Either Party:

The simple answer, the obvious answer, is that such a poll might serve to energize and fire up the base of the “leading” candidate while also having the benefit of discouraging and tamping down the enthusiasm of the “trailing” candidate. Most engaged voters recognize this ploy and understand it is being used. This time…in this election, though, I suggest that there could be a more sinister factor at play. I pose this merely as a question to cultivate intellectual awareness. And, given that Obama, Jarrett, Axelrod and their merry band of Chicago thugs have proven time and again, their ends always justify their means.

Let’s imagine. First, let’s imagine that the likely voters in America actually do support Romney over Obama by 52 / 48. And, really using our imagination, let’s suppose that virtually all of the polls, immediately before the election, predict a Romney win. But, lo and behold, we wake up on November 7th to find, as a surprise to all, that Obama has carried over 50% of the vote. Were the polls ALL that wrong? Even the most naďve among us would likely scratch their heads and wonder aloud, “How did THAT happen”. Many will conclude that there must have been a huge turnout among the “unliving voters”, illegal aliens, and felons. None of these groups would have been represented in any of the polling results. An Obama win in the face of all of the polls predicting his defeat will elevate the mistrust in the election results to a level that even Obama and his spin doctors might not be able to control.

If, as I suggest, fraudulent votes will play a key role in re-electing Obama in November, the fact is that Obama NEEDS the voters to have gone into the election EXPECTING him to win. Fraudulent, misleading, and dishonest polling is absolutely essential in the Administration’s pre-planned defense against charges of gross voter fraud in November. Even he and his union thugs cannot stand against a Nation of voters who are incensed by the conviction that an election for the highest office in the land had been stolen from them.

Lastly, the only pollster whose work is both honest and accurate is Scott Rasmussen. His firm’s last polls before the last two presidential elections have been far and away the most accurate in terms of predicting the final outcomes and margins.

On an unrelated topic, why does the unelected, never coronated, Queen Valerie Jarrett deserve 24/7/365 Secret Service protection on the taxpayers’ dime, when Ambassador Stephens and his staff are devoid of any personal protection on 9/11 in LIBYA amidst terrorist threats apparently known by everyone in the Mideast EXCEPT the Obama / Hillary State Department?

The 912 Project