Why did some polls fail to spot FF resurgence?

Anomaly a combination of ‘overestimation, underestimation and a late swing’

“The Ipsos MRBI poll for The Irish Times underestimated support for Micheál Martin’s party by 1.5 per cent.” Photograph: Leon Neal/AFP/Getty Images

The performance of Irish polling companies in the general election – while far from perfect – could not begin to be compared to the dismal failure of British opinion polls to identify a clear Conservative victory in last year’s Westminster elections.

But it will hardly be remembered as vintage.

Not one poll came remotely close to predicting that Fine Gael and Fianna Fáil would end up between one and two percentage points of each other, although the Ipsos MRBI exit poll for The Irish Times on the night of the election came closest.

The final surveys conducted by four major polling companies in the week of the election put the gap between the two largest parties at between four and 12 points. The closest was Millward Brown, which put it at four points (27 points to 23) followed by Ipsos MRBI, which had a five-point gap (28 to 23).


Red C had Fine Gael at 30 and Fianna Fáil at 20 in its last poll published on the Tuesday before Friday’s election – a 10- point gap.

Behaviour & Attitudes had a 12-point gap in its last poll before its exit poll: 30 points to Fine Gael and 22 points to Fianna Fáil. *

On Sinn Féin, only Millward Brown was significantly off in its last poll, ascribing 19 per cent support to the party, 5.5 points above what it received in the poll.

The other polling companies recorded figures between 15 and 16, which were within the 3 per cent margin of error.

Generally the polling companies’ results were close to Labour’s final figures (suggesting the party got no swing in the final days). They were also broadly accurate on Independents and others, although Millward Brown underestimated the support levels of non-aligned Independents to an extent outside the margin of error.

The two exit polls, with much larger samples, were much closer to the final outcome. However, the Behaviour & Attitudes Poll for RTÉ underestimated Fianna Fáil support by more than 3 per cent and overestimated Sinn Féin support by more than 2 per cent. The Ipsos MRBI poll for The Irish Times underestimated support for Micheál Martin's party by 1.5 per cent.

So why did the opinion polls not identify that gap? Was there a dramatic late swing to Fianna Fáil and away from Fine Gael and Sinn Féin? Or did the polls simply overestimate the support levels of Fine Gael and Sinn Féin and underestimate that of Fianna Fáil?

Poll analyst Sean Donnelly said the figures for the four big parties were virtually the same as they had been at the local elections, suggesting little long- term variation.

“These figures are practically the same. The wisdom has been to ignore local elections. Well, you can forget about that one,” he said.

He said there had been some very optimistic seat projections on the back of the poll results, with some commentators predicting Fine Gael would have a seat tally in the mid-60s. In the end, it scraped over the 50 mark thanks to some astute vote management in key constituencies, which gave it a seat bonus. That kept a small gap between it and Fianna Fáil despite them being very close in percentage terms.

Prof Gary Murphy of the DCU School of Law and Government said there was clearly an anomaly between the poll results and the outcome.

On whether there was underestimation, overestimation and a late swing, he said there was probably a mixture of all three.

“It is very hard to posit why that is the case. I was one of those people who thought Fine Gael would get over 30 per cent. The campaign clearly mattered more this time than the elections in recent memory.

“You would honestly ask the question why Fianna Fáil, which got within one point of Fine Gael, why that was not reflected in the polls,” he said.

Prof Murphy said he thought the polls succeeded in reflecting the fall in Sinn Féin support during the campaign, partly attributable to poor responses to specific questions on economic policy by Gerry Adams in a series of debates and interviews.

*This article was amended on March 2nd 2016 to correct an error