Analysis: Polls are a flawed but important tool

I WON’T repeat here the word that left my mouth the moment after David Dimbleby announced the outcome of the General Election exit poll on the night of May 7 last year.

But my slightly more considered response after a couple more seconds of contemplation was “that can’t be right”.

With a deadline to get our first print edition away fast approaching, the discussion in The Yorkshire Post newsroom centred on how many caveats we should place in our front page story. How could we write that David Cameron appeared to be heading back to Downing Street while also leaving wriggle room in case by the time people read it in the morning, the poll had proved to be horribly wrong?

Shaped by weeks of poll results, election night had started with expectations of weeks of coalition discussions lying ahead and by the following morning Mr Cameron was leading a majority Conservative government and Shadow Chancellor Ed Balls was looking for a new job after losing his Morley and Outwood seat.

It was not a comfortable moment for the polling organisations and seven months on the findings of the inquiry into what went wrong have been published identifying problems with sampling and companies “herding” to ensure they published similar results.

In the aftermath of the election there was talk of a loss of faith in opinion polling. And yet last weekend the Mail on Sunday’s front page story was based on a poll suggesting voters wanting to leave the European Union are now in the majority.

Because despite the limitations ruthlessly exposed last May, polls remain an important tool in discovering public sentiment, not just for journalists but also for political parties and Government bodies as they shape public policy and spend taxpayers’ money.

And despite my experience last May, I’ll still be looking to them to offer a guide to how the arguments of the remain and leave campaigns are being received when the EU referendum campaign gets underway.

I hope they get it right this time.