Actually they don't show that. Prior to elections all local polls are distorted by what is called the incumbency factor. People know the names of existing TDs so they show up with more support than occur in the actual election, when the non-incumbents get mass exposure too.Originally Posted by The Langerboy
So reaching conclusions based on pre-election local polls, or even early polls, is pointless. The last time most early polls proved worthless in terms of reliability.
Secondly, most local polls make claims as to how last seats will work out. In reality it is impossible to make those predictions. Local polls are small to start off with. Their transfer predictions are based on such small subsets of polls that the margin of error can be 10% of higher. In other words, any poll that predicts transfers should be binned. As pollsters know well, you can no more predict transfers based on small samples than you can predict results based on reading tea-leaves. The subsets of samples are way too small to have any reliability whatsoever.