Media Hiding The Truth About Polls

Polls.  Polls.  Polls.  Every day a new poll surfaces on the presidential race.  The mainstream media report the poll results as if the data is factual information.  Reporters and editors make no attempt to warn the public of the errors inherent in the methodology of most political polls.

First, let's dismiss the idea that polling results represent facts.  They do not.  Polls are a snapshot in time of the sentiment of a narrow slice of Americans.  Most polling organizations interview 1,000 or fewer people and then extrapolate the results to produce a purported representative sample of adults.

To put that into perspective, there are about 200 million adults in the United States, according to the latest U.S. Census.  That means the typical poll of 1,000 adults is representative of .0005 percent of the adult population.  The media never mention that fact in reporting on polls.

The Roper Polling organization estimates that an adult's odds of being called in any given year for a political survey are more than 100 to one. How many friends and family do you know who have participated in a telephone poll in any presidential election?  Crickets.

Secondly, the vast majority of presidential polls are conducted by land line telephone. Ask yourself: How many people with caller ID even answer a call from a polling organization?  How many adults accept the call then refuse to answer the questions?

Pollsters have tried other methods, including online surveys and text messages to elicit opinions about politics.  However, both political parties have learned how to influence the results by encouraging their supporters to cast votes to sway the poll.  That renders the data worthless.

It is a disservice and dishonest for news organizations to keep the public in the dark about all the variables implicit in the polls.  Don't take our word for it.  The ex-chairman of one of the most-respected polling organizations in America has scolded the media for misleading the public.

"But they (the media) would be better off assuming--as most of the readers surely do--that all surveys and all opinion polls are estimates, which may be wrong," warned Humphrey Taylor, the former chairman of Louis Harris And Associates, Inc., writing in an article in 1998.

Taylor pointed out that the wording of questions, the order of questions, the refusal rate, the non-availability of people and inadequate weighting are factors that make polling results subject to "substantial error."

For that reason, Harris included a "strong warning" in all its polls.  "It is difficult or impossible to quantify the errors that may result from these factors (cited above)."  So-called journalists are not the least bit interested in these caveats, concluded Harris.

Despite all the admonitions, the media persist in deceitful reporting of results. Unfortunately, too few Americans understand political polling, thus many continue to put stock in the results.

Although political polls enjoy public trust, these samples have a checkered past.  In the 1948 presidential race, the Gallup poll had Thomas Dewey ahead of Harry Truman, 45 to 41 percent. Truman won the White House with 50 percent of the vote to Dewey's 45 percent.

In 2008, a consensus of seven polls taken just before the New Hampshire Democratic primary showed Barrack Obama had an eight-point margin over Hillary Clinton.  Instead, Ms. Clinton won by three-percentage-points, leaving pollsters red-faced.

American pollsters could learn a few things from the British.  After opinion polls were grossly wrong in predicting the outcome of May's general elections in Britain, some smart folks decided to do a face-to-face survey of voters by going door-to-door.  

The sample included 3,000 people who fit the profile of those who regularly showed up and voted at the polls, instead of just surveying adults in general.  The results were eyeopening.  The voters' preferences in the poll accurately reflected the actual vote in the general election.

That begs the question:  Why don't more polling outfits employ that same methodology if it is more accurate?  The simple answer is money. Telephone surveys are infinitely less costly than door-to-door polling, especially when the research is conducted monthly or even weekly in the U.S.

The big media cabal understands the flaws with U.S. presidential polling. That's why their reporting is a deliberate deception.  There is never an attempt to explain the limitations and fallacies of research. Reports on the latest polls are presented as unassailable truth.  

Shawn Parry-Giles, a political communications professor at the University of Maryland, argues the media should stop treating polls as if they are authentic.  "This is about what the voters say and do, and the media has to be very careful about how they frame the polls," she notes.

In other words, what voters tell researchers often may not be the same as their decision in the voting booth.

Her advice has gone unheeded by a mainstream media intent on using polls to advance their narrative about the race.  When polls showed Hillary Clinton leading, it was front page news.  Now that Donald Trump has narrowed the gap, there is less media enthusiasm for showcasing the results.

With Ms. Clinton sliding in the polls, the media use data to strengthen the case for Ms. Clinton.  When Trump began climbing in the polls, the media quickly noted that he was losing with women, Hispanics and African-Americans. The media snickered that Trump was winning only with white men.

Of course, these conclusions were drawn from polls, too, subject to the same distortions as the presidential samples.  In general, people trust polls way too much.  The only thing that matters is the actual votes cast on November 8.  Everything else is political chatter.