Why You Shouldn’t Trust Polls

October 22, 2009

Poll watching is a great way to see which politicians are winning what races, what the country thinks of the President, and other various topics of interest.

However, one is often left with several questions if you are looking at polling data on the same subject that are taken during the same time frames.

For example, on 10/14 Fox News reported that President Obama had an approval of 49% of respondents and a 45% disapproval.  A difference of positive 4.  However, on 10/18,  ABC News/WaPo reported that President Obama had an approval rating of 57% and a disapproval of 40%.  A positive 17 percent difference.

What can explain the difference? 

“… do you support candidate A’s common sense approach to problem x or candidate B’s reckless and irresponsible policies?”

Most likely the biggest difference is that Fox news interview sample was 41% Democrats, 37% Republicans, 16% Independent, and 5% other/don’t know.  The ABC/WaPo interview sample was 33%  Democrat, 20% Republican, 42% Independent, and 5% other/don’t know.  The independents were roughly a wash with 20% leaning Democrat and 19%  leaning Republican.

As you can see in the WaPo sample, the vast majority of people interviewed were Democrats or lean Democrat.  Adding the percentages together, the ABC/WaPo sample was 53% left, 39% right, and the rest center or don’t know.  ADD moment: what is up with all these “don’t know” people?  It turns out that Democrats support President Obama more so than Republicans.  I am sure that if you polled a vast majority of Republicans, the approval would plummet and the disapproval would skyrocket.

According to a story at HotAir.com, a recent Gallup poll put the difference between Democrats and Republicans at 6 points and not the 14% represented in the ABC/WaPo poll.

Another interesting tactic to skew the results is the words used when asking a the question.  For instance, I once received a robo-call asking me “do you support candidate A’s common sense approach to problem x or candidate B’s reckless and irresponsible policies?”  Um, what was that common sense one again?

There is a great post over at RealClearPolitics.com that shows how the language used can boost or hurt perceived support for the “public option”.  Support shoots up if the poll uses words like “choose” and “compete”.  However, when words like “government controlled” or “taxpayer funded”, the support drops.

I’ve become a big fan of Real Clear Politics as you can click on a poll and often get the details on who was asked what.  A reputable pollster will at least include some statistics that can help you gauge any bias in the poll.

By the way, 100% of those surveyed strongly approved of Conservative Patriot HQ.

Tags: , , , , ,

Leave a Reply

Your email address will not be published. Required fields are marked *

Sponsor Links

Follow Us on Twitter