And our survey says...

The first thing you would have heard from me on election night 2015 was a sharp intake of breath. If I sounded surprised as I talked listeners through the exit poll, that’s because I was surprised. The biggest surprise was that I didn’t fall off my chair.

Almost nobody predicted the Conservatives would win an outright majority last May — including many Conservatives. Now a review for the polling industry has revealed the “systematic over-representation” of Labour voters in surveys, coupled with an under-representation of likely Conservatives, led to the false presumption that the election was neck-and-neck, and a messy coalition inevitable.

These two graphs explain exactly how the Conservatives won:

The older voters pollsters failed to find were not only far more likely to support the Conservatives, they were almost twice as likely to vote as those under 25.

But modern polling firms find it far harder to reach older voters. In the past, you’d stand on a street corner with a clipboard and find enough elderly ladies to match the population. Now, you sit in a call centre or wait for people to respond to e-mails, and try to balance up your sample later.

So we missed countless older people who were firmly in the Tory camp, some of whom would have told us had we only asked them.

The younger voters who pollsters did reach were far more likely to back Labour — but their enthusiasm took no account of the 57% of under-25s who simply wouldn’t bother to vote.

Add in a Lib Dem collapse, gifting the Tories a swathe of seats across the south, and you can see how the Conservative majority was built.

A little bit me... a little bit you too...

All this, of course, is with the benefit of hindsight. And it’s easy to demonise the pollsters for failing to accurately poll public opinion.

Let’s not forget, however, the role of both of the press and politicians. On election night, the Tories and SNP insisted the forecast was beyond their most optimistic private polls, while Labour and the Lib Dems said it was beyond their worst-case scenarios. The truth is, the parties allowed themselves to be led by the polls, as did the journalists covering the campaign.

A year before polling day, Labour abandoned any effort to reach beyond the 35% of voters they felt would be enough to finish up on top.

The Conservatives put a series of hardline measures in their manifesto, partly to appease supporters flirting with UKIP, but partly because they expected to jettison some policies as the price of the inevitable second coalition. Now they’re committed to massive welfare cuts, and almost impossible EU reforms.

Meanwhile, from the very start of the campaign, those of us covering the election framed it as a pointless stalemate — no more than a warm-up for the real business, the coalition negotiations that would clearly follow polling day.

The campaign was fought through the prism of coalition — endless debate about who would do deals with who, what they’d give away for a taste of power.

As a profession, we spent too long in our studios and newsrooms, and not enough talking to actual voters. When we visited a constituency, it was to ask the candidates whether Labour should do a deal with the SNP.

Had we spent more time on the ground, we would have picked up just how few people thought of Ed Miliband as a credible Prime Minister, how little confidence there was in Labour’s economic credentials, just how much trouble big-name Liberal Democrats like Vince Cable and Simon Hughes were in.

The pollsters have a lot to learn from the 2015 campaign. So do those at the top of politics, and those who report on them.