Exit polls

Excellent overview of the virtues and flaws of exit polls here by Michael Marsh on RTÉ. He notes three key aspects:

There was much talk in Britain about “shy” Tory voters who biased polls because of their unwillingness to be interviewed. It was a good story, but thorough investigations into polling failures found little evidence to support it and good evidence for other explanations of error.
It is harder to see why the big parties are generally underestimated. That is not consistently the case in pre-election polling.

And:

Exit polls could be made more precise if any organisation, governmental or private, was willing to provide the necessary money to increase sample size, locations, interviews and time windows.
It is noticeable that the BBC exit polls have greatly improved since they substantially increased their spending.

But this really resonates with me:

We must have some sympathy for those voters who were willing to stop and answer questions. It was hard enough to search through more than half a meter of ballot paper without having to remember afterwards what you did, who you did it to, and in which election.

I was struck by this latter phenomenon leaving the polling place on Friday evening last. Just going over my vote it struck me I knew the first three I’d gone for in both local and European elections but after that while I was pretty sure I had the next three and I think that was pretty much that for me on the locals I couldn’t quite recall which order were 4/5. Then again I know precisely who got my 1st preference.

I really don’t think exit polls should be regarded as much more than filler – indicative of but not precise as regards outcomes. Then again, as I noted elsewhere we also have tallies to contend with even later in the election process.