Why Political Polls Are Misleading


We are constantly being bombarded with political polls showing Obama way ahead or Romney ahead. Do these polls REALLY reflect anything meaningful?

The simple answer is "NO". But, the answer is more complex than that. The true answer lies in who is conducting the poll and whom do the pollster(s) support (or, how much bias is there?), what questions are being asked, and a whole host of other issues including how many persons were "polled".

Let's deal with the "what questions are being asked" first. A single, simple example should suffice for this one. Recently a poll reported that the majority of Americans were in support of allowing "illegal immigrants" to be in the US. The actual question asked was "Do you support immigration?". Not exactly the same issue, I feel. I know of no one who is against LEGAL immigration - the greatness of our country is based on legal immigration - to one degree or another. The actual questions used in any poll must be reviewed to ensure the validity of the reporting. More misrepresentation occurs than might be imagined.

Next, the number of persons being polled is an obvious one. The fewer people polled the smaller number of persons the resulting "numbers" can truly reflect. So, always check to see what the population size is that the poll is supposed to reflect AND check that against the number of people polled. If the poll is supposed to reflect the entire country, the sample being polled better be pretty large. Note: Many of these polls have around 1,000 people participating. One Thousand representing 300,000,000 ( i.e., .00033 % )? Not too very statistically relevant, I feel.

Lastly, let's look at "built in" bias in polling. This bias is what, to a large degree, is driving the current polls on the race for the Presidency. An easy way to "build in" bias is to skew the people being polled by those most likely to agree with the results you want. As we all should be aware, what is called the "mainstream media" (e.g., ABC, NBC, CBS, CNN, NPR, etc.) is heavily in favor of Obama's re-election. FOX seems to be the single holdout for Romney. The Pollister, Rasmussen, seems to be the only major polling firm that is "in the middle".

The below picture reflects a "scattergram" of polling sources and their biases:

From the above chart, every one of the Obama supporter news outlets have him ahead in the polls. But, look at the "skew" of poll participants. All have a significant higher number of democrats participating vs. republicans or independents - even FOX. Based on this "approach" is it any wonder that Obama shows up in the lead? Were this same "skew" to use republicans rather than democrats, Romney would show up as ahead. By the way, all polls that I have seen show independents favoring Romney by about 11%. Even in this, one has to ask "what were the questions?", etc.

Note: There is a website that "unskews" the poll data that is worth the "look" - (Click Here for the site).

So, without doing mucho research, poll results should be considered "suspect" - no matter who shows "ahead". If this is true, then what good are they? I believe they are nothing but political "tools" meant to be conveyed widely. The wider the better (for the one who is "ahead"). Why? Because, since most people will NOT do the necessary research, if the poll data is spread widely enough perception becomes reality - Or, another invocation of the "never let the truth stand in the way of a good story". The normal human condition is to want to be on the side of the winner. If one believes the polls, then that "human condition" will lead those undecided "to go with the flow". Might be good for "sheep" but not so much for the rest of us.

A point to remember: An educated populace is better than an ignorant one. Elsewhere on this web site is my "definition" of education - most applicable here ( Education Definition ).