Wednesday, April 23, 2014

I Was Part of the New York Times Flawed Poll On Louisiana Senate Race

I differ from most people that think telemarketers and pollsters should have a place in hell reserved for them where the flames burn a little hotter.

First I think political polling plays a critical importance in our nation.

 Second I have actually done what these folks calling are doing.. I cringe thinking of the verbal abuse I got from folks for a job that was not paying megabucks to say the least.. So when a poll taker gets me on the phone it's their lucky day. I also get that the poor guy  on the other end of the line does NOT get paid unless he or she gets complete responses. So no matter who is running the poll I always complete it no matter who is running it.

I got called last week to take part in a rather extensive poll .

I took particular interest in this poll since it was done by the New York Times. The results of this poll taken in several Southern States where Senate races are hot are in.

See Poll Shows Tight Senate Races in Four Southern States .

As to two of these southern states I think I have some familiarity with the current political dynamics . The first being Louisiana of course and the next to some degree Arkansas since I am just 30 minutes from the border.

 The poll  results today I think show while polling is needed it is still a very tricky business

The results of the polls on the Senate races drew some pretty quick reactions of disbelief in some Republican quarters. Pryor ahead in Arkansas by 10 points ? Mary Landrieu beating Congressman Cassidy 42 to 18 percent ? What ? From what I know this does not depict a reality any of these campaigns think are happening on the ground.

So what happened. First as the Weekly Standard  notes the voting sample used shows a voting public that does not match reality. They say in part :


In other words, the Times and Kaiser have produced a sample in Arkansas that reports they voted in 2012 for Romney over Obama--by one point. But Romney carried Arkansas in 2012 by 24 points. Similarly, the Kentucky sample is +3 Romney when reality was +23. The Louisiana sample is +3 Obama in a state Obama lost by 17, and the North Carolina sample is +7 Obama in a state he lost by 3.

Now how did this happen ?

First this poll had to be one of the longest polls I ever took. It ran close to 30 minutes and I was counting the minutes because the LSU baseball game was on at the time which I was of course missing.

The vast majority of questions were about Obama are / ACA and related Louisiana issues of Medicare expansion. Strangely these questions do not appear in the list of questions that were asked that is provided by the New York Times.

The questions on Obama care related issues seemed to take on a push poll like quality. A push poll is when the questions are phrased in a way to get a result. I am not sure that was the intention but that was I thinking .

I think this might have been an effect of just the cumulative nature of the questions . Further the questions were very detailed at times and the pollster seemed not interested in me putting things into a undecided or don't know category . The pollster was doing what is called a hard sell. He wanted you to make a choice.

I was starting to feel some aggravation at this as this poll went on and on. I can just imagine how some that are not pollster friendly were feeling. This becomes critical because  the who you were going to vote for Senate question came up fairly late.In fact it seem just a sideline of the overall poll itself.

I suspect a good many  of anti Obamacare , and conservative leaning voters bailed on this poll before it was completed and thus their responses were thrown out. . LIKE HEY YOU HAVE TAKEN ENOUGH OF MY TIME sort of thing.

On the flip side if you really are invested in the success of Obamacare /ACA I think it was more likely you would answer questions all day as to the way this poll was set up.Thus you got to the Senate question and had your entire poll answer counted.

So at the end of the day we get a voting sample that just seems so out of whack with reality. If the poll had fewer questions or perhaps more questions on a variety of topics we might have got a poll that seems to reflect in a Louisiana voting public ( the folks that actually vote )  reality.

Again polling is a tricky business.

That being said the fact that Mary Landrieu is only running in the low 40's in a voting sample of Obama + 3 can't be the best news of the day for her.


1 comment:

Undergroundpewster said...

I have yet to hear one of those telephone polls that was not biased in one direction or another. Some questions are subtle so maybe we need to call that "puff polling," or maybe that term should be reserved for the marijuana legalization polls coming to states near and far, or should I say, high and low.