I've also been told I have little tact, so if this offends you simply ride on.

Tuesday, May 10, 2016

Polling failures

Just a few thoughts on polls before I take Big Seig to the vet for his shots.  They are very good for pointing towards trends but very bad depicting reality.  Republican polls had shown Romney smacking the shit out of Obama and winning the presidency... didn't happen.  Interpretation is always biased, not matter how much they tell you the opposite, pollsters want specific answers.  I know this because I was polled about two months ago.  Even though I was told it would only take a few minutes, I ended up enduring questions for almost 45 minutes.  At one point the pollster actually said, "well, you're pretty noncommittal," that's because I didn't hate either Bernie or Hilary.  My strongest answers were against the Republican party, and while they were strong, they weren't hate, hate, hate.
I bring this up because this morning I was reading about a hypothetical match up, who would fare better against Trump, Hilary or Bernie, and, of course, the hypothetical match up was based on polls.

For those who have never taken a Poll, here's an example of a question:

What do you think of Donald Trump:
          A.  I love him
          B.  I like him
          C.  I dislike him
          D.  I hate him
          E.  I have no opinion

The next question depends upon how you answer... and it's always multiple choice.  You never get asked why you chose that specific response.  So if you say you hate Donald Trump they never give you the opportunity to air your grievances, what it is about him that so turns your stomach.
The pollster actually asked me who I would vote for in a General Election.  My response:  "We are guaranteed a secret ballot in this country, so why would I tell you?"  Anyway, that wasn't the response he was hoping for.  This is life.


  1. Part of the fun/foolishness of polls is that if you ask people if they trust polls, you are in fact polling them. If I do not trust polls, and I answer "No", would that make your poll trustworthy? Or would "Yes" make the poll unreliable and thus confirm my bias? Am I overthinking this? Under-thinking? (The latter seems more common among both poll designers and poll respndents.)

    At any rate, when I have been asked questions on the phone in political polls, the questions are sometimes as absurd as "Do you support the kind and loving John Doe or that vicious rabid hatemonger Ellen Bellen?" All too often even when questions are more reasonable, after I answer the first question, they thank me and hang up. As you point out, they ask no follow-up questions to find out what my opinion is based on, if it is a matter of religious belief or mental aberration or misogyny or whatever. The information they obtain is essentially useless but will provide fodder for endless political pundits to predict what will happen and then to explain away why they were wrong. And none of them are likely to say, "We relied on polls that asked stupid questions and obtained useless data and thought we knew what the answers meant and we didn't. Therefore we are quitting our jobs and recommending that you spend you time more fruitfully by visiting with your grandchildren."

  2. Whats that quote about lies and statistics