Monday, April 04, 2011

POlling problem after all?

Sure enough, just days after I write that the Civic Platform's (PO) polling problem was conquered, a poll comes out showing a polling problem. The PO's support fell to 31% in a weekend survey from the pollster Homo Homini. Jaroslaw Kaczynski's Law and Justice (PiS) was just 5 percentage points back at 26%. Considering the margin of error is at least +/- 3 points, the race would appear to be very tight.

But wait a second. Didn't the PO just receive 46% in a survey done by GfK Polonia, putting it 22 percentage points above PiS? Isn't Prime Minister Donald Tusk's PO clearly over its early-year problems and well-positioned for the October parliamentary elections?

In short, something smells worse than a batch of rotten prairie oysters.

What is going on? Well, basically, Poles are schizophrenic. Yes, all of them. One day they like the PO and the next they don't. Problem solved.

The true answer could perhaps be more involved. The pollsters themselves would invariably say polls are merely snapshots of opinion at a particular time and not a 100% view of what many think or what will happen on election day. There is an error margin. The question matters. What form of poll is another issue.

Respondent selection is one more major problem. The 1,000 or so sample size can be representative, but it isn't always. In Poland, voters of PiS and the Democratic Left Alliance (SLD) tend to be more reluctant to talk to pollsters or to admit they like parties often denigrated by others, especially the louder ones. In a country in which totalitarianism is not yet such an old memory, assurances of confidentiality are often little assurance at all.

If the poll is a telephone one, as the Homo Homini one is, big biases could show up. Many rural Poles don't have telephones. Poor ones as well. This would tend to skew the results to the PO. But young people are giving up landlines in favour of mobiles, potentially pushing the chance of error the other way.

Poles also love to back the winner. Many will say they support the PO every single time asked but vote totally different in the actual election. The PO used to suffer massively from this problem. I suppose the flipside of its turn toward populism means the PO could in fact be less impacted by this effect. But then wouldn't it be sad if the PO dropped its previous beliefs in order to build support levels only to seem them decline in the election itself. It would have sold itself out for nothing. And that would be a problem.

1 comment:

  1. Hello, this is the Poland X ombudsman here. Just wanted to write a quick note to say question marks in headlines are a big pet peeve of mine. Don't do it!