Cooper was talking at Impact 2015 yesterday ( 17 March), in a session presided over by Deborah Mattinson, co-founder of consultancy BritainThinks and a former pollster to Gordon Brown, with Mattinson pronouncing the upcoming General Election as the “hardest to read in a long time”.

Mattinson found agreement among the panellists who joined her onstage – Populus founding director Cooper; Mike Smithson, a former Liberal Democrat politician and founder of PoliticalBetting.com; and James Morris, a former speechwriter to Labour leader Ed Miliband and director at Greenberg Quinlan Rosner Research.

“The hard thing is not predicting poll shares but working out what they mean in terms of the seats that parties will win,” said Smithson. “It’s those 650 seat battles which are important. I think we’re seeing a fight which is going to be focused on the marginal seats.”

Meanwhile, Morris extolled the virtues of political research in the UK saying that of all the countries he has worked in “Britain’s opinion and political research is by far and away the most open, transparent and voluminous”, allowing “pundits, voters and academics to understand what is going on, that is not available elsewhere”.

While admitting the General Election result is hard to call, Smithson said that Scotland is probably where the election will be won or lost. “Labour have picked up in England and are in a strong position and what has happened in Scotland will probably determine the result,” he said.

Cooper agreed that Scotland would be “quite critical” to the outcome, with Labour irrecoverable in the region, highlighting one fringe party that will hit the Tories hard in the election.

“Our analysis suggests that if Ukip are above 12% [vote share], it will start to cost the Conservatives a lot of seats,” he said.

“A lot of people are deciding that they don’t want to vote for anyone at all. There’s an extraordinary phenomenon – [polls saying] that 25-30% of people will vote for the fringe parties. It’s unlikely they will do that, but if the turnout at the polls is only two-thirds of the electorate, which two-thirds that is makes a huge difference.

“We don’t have a party with broad enough an appeal to win convincingly. It’s more difficult this time because of the rise of the marginal parties… But if I had to bet – it’d be on David Cameron remaining Prime Minister in some messy government, then more polling on the EU referendum.”

Lib Dem alumni Smithson disagreed: “I expect a Labour government and a Labour win in share and in seats,” he said, while Morris also hedged his bets on a Labour government being in power post 7 May.

Scottish referendum

Meanwhile, the panel were asked by a member of the audience about how pollsters forecast the outcome of 2013’s Scottish Independence referendum so wrong, a point that the panel were quick to quash.

Cooper argued that it was one pollster out of 100s whose prediction sided with a ‘yes’ vote, a forecast seized upon by the media, and Morris underlined his point. “It wasn’t the polling industry that got it wrong. It was another industry,” he said, with a clear nod to the media industry.