Why Polls Are Dangerous

Polling data or commentary on the results of polling data are a common component of news media outlets.

 Since the Republican primaries are under way, endless polls and forecasting winners and losers in the electoral races (based on polling results), makes up even more of news media content.

 Polls and election forecasting affect peoples choices and the political parties know this.

 Remember the 2000 Presidential election, where first the major network called the election for Gore, then it was discovered that Bush had actually won a couple of hours later? The major networks called the election before all of the polls were closed inFlorida.

 The panhandle ofFloridais in a different time zone, which meant the polls, in that part of the state, were open an hour longer than those in the eastern part of the state.

But the major networks couldn’t wait to be the first to call the race which announcements filtered down to radio stations.

 It was discovered later, after polls had closed, not only did Gore lose, but many Bush voters did not cast their votes thinking their candidate had already lost, so there was no point. The forecasting of the election, based on polling data created a lawsuit that never should have taken place.

 But it did, and the whole exasperating courtroom drama on television soon became a pointless argument about hanging chads, that weren’t punched properly by the voting machines. All of this could have been avoided had the networks reported the news instead of being in the business of forecasting.

 Polling and election forecasting turned out to be dangerous to the presidential elections in 2000. If election forecasting and polling had this much effect inFlorida, one can only guess the effects it had in the other 49 states.

 We are still bombarded with endless polls; which primary candidate will win in this state or that? Which Republican candidate will win against Obama in the general election? The polls are used to compile forecasts, and those forecasts are designed to influence your vote.

 The results of all of these polls are swayed one way or another. The results are subjective, based on who you ask and how you ask.

 Here are a few examples:

 This gallop Headline  

InU.S., Slightly More Want Obama to Set Course Than GOP

Based on this question asked last in a series of questions about Congressional job approval:

 Who do you want to have more influence over the direction the nation takes in the next year – [ROTATED: Barack Obama (or) The Republicans in Congress]?

 Since the Congressional approval rating is around 15%, who do you think is going to come out on the short end of the stick Obama or Congress?

 Let’s look at a Rasmussen example of an unlikely presidential match up;

2012 Presidential Matchups

Election 2012: Obama 48%, Santorum 38%

 Based on this question:

1* What if the Republicans nominate Rick Santorum? If the 2012 Presidential Election were held today would you vote for Republican Rick Santorum or Democrat Barack Obama?

Santorum                   38%

Obama                                    48%

Some other candidate          9%

Not sure                     5%

 How many people even know anything at all about Rick Santorum? Few people even know who Rick Santorum is; much less know anything about his views or political history.

Everyone knows who President Obama is. How can you not? He is on Television every day!  Name recognition alone would produce the same results this poll.

 And why was it necessary to include the party affiliation in the question? It wasn’t, but it sure swayed the results.

 Here’s a polling question I would like to see the results from;

“If a democratic primary were held this summer would you be more likely to vote for:”

Barack Obama

Hillary Clinton

Some other Candidate

Not sure

What would those results yield and how would that sway the electorate?

Think that it hasn’t been thought about?

How about this question:

“If the election were held today, and a candidate like ‘Joe the plumber’ was running against Barack Obama, would you be more likely to vote for”

Barack Obama

Joe the plumber

Not sure

What would those results show?

Let’s change out focus from the Presidential election to Congressional Approval ratings

There are some polls which show a congressional approval rating of less than 14%. And considering Congress is comprised of both the Senate and the House of Representatives, let’s see how they did separately. 

This question from real clear politics; 

“Please tell me whether you think each of the following political office-holders deserves to be reelected, or not. How about the U.S. representative in your congressional district?”

    Deserves to be       Does not         Unsure    

              53%                          39%                 8%    


How about this one from Pew;

“Would you like to see your representative in Congress be reelected in the next congressional election, or not?”

    Yes                          No                   Rep. is not running (vol.) Unsure  

      50%                         33%                             1%


 So Congress has a 14%-18% approval rating, yet when the Congressional houses are separated, and you are being asked about your Senator or Representative, the results are completely different. How is that?

Some of the discrepancy could be attributed to the fact that you are no longer dealing in generalities, but talking about a specific representative, but the greatest portion of the discrepancy has to be attributed to the phrasing of the question.

This makes polling dangerous, because it leads to forecasting, and forecasting leads to influencing the electorate.

Here is a nice polling summary; but notice how and to whom the questions are posed (which the selection of the actual participants is not always available to see).  

The people who actually participate in these polls are often referred to as “likely U.S. voters.” Just what exactly is a “likely U.S. voter?” Is that someone who will probably vote, or is that someone who is registered to vote?

 Why is that important? When was the last time we had a registered voter turn-out of more than 65% on a statewide or national election? And how many of the likely voters were in the 35% who didn’t vote?

 Of course the major polling companies post the methodology (which contains who, what, where, and when of the participants) but the screening is still vague at best.

 Maybe the methodology should contain a little more detail of people who participate in these polls; maybe, the pollsters should ask these participants “when was the last time you actually voted?”

 And the most important question; when was the last time, if ever, you contacted your Senator or Congressman about any issue?

 In the final analysis you must not let polling data influence your decision on who to vote for or let it deter you from voting. Polling and projections influence votes; you are less likely to vote for a candidate if you think they are going to lose. That would be un-American, so you don’t bother. And why would you think they are going to lose?

 By the polls of course!

 The 2012 presidential election is too important to the future of this country, and the polls are constantly trying to convince you that Obama cannot be beaten. But he can, and should be, if he runs on his record. But he cannot be beaten if you are swayed by the polls and stay at home.


Think about it!