Details of today’s Irish Times/Ipsos Mori poll can be seen here.
Quite a shift since the last Red C poll, discussed in a previous posting. Even allowing for the usual issues of margins of error and different polling methodologies, this is a dramatic result for Labour and not the best of news for Fine Gael who have slipped into second place and whose leader’s poll ratings plummeted almost as dramatically as the Taoiseach’s. Already the mutterings are beginnings.
Ironically, while the poll news is just as bad (actually worse) for Fianna Fail, the impact on Fine Gael is providing some welcome distraction for the embattled Taoiseach, and perhaps also a hint of a chance to cut and run for a snap election sometime soon?
Irish Politics Forum 
I can’t see either FF or the Greens wanting to cut and run. What’s in it for them? They have particularly bad prospects at those elections (even if FG has failed to convince the electorate). In fact the poor opinion polls may be used by Cowen to ensure that there is no more leaking of support within the party. The threat of an election might help keep the government together. The only problem with this is that a small number of FF TDs might see their prospects at the next election improved by sniping at the party leadership. If there were a serious challenger for Cowen’s job within the party, now might be the time to move (though would anyone really want the job right now?)
Interesting to read the ‘note on methodology’ attached to this poll. At it makes clear this poll is ‘calibrated’ to take account of historical voting trends — the problem now being that these poll results are hugely out of kilter with those trends. So, the model is misspecified. As the polling agency indicates, a new calibration is in preparation. This will mean that, whatever about the actual poll trends in their next survey, Labour can expect to see its vote drop simply because of this recalibration.
I’d love to see the uncalibrated data underlying this finding. How calibration is achieved is far from clear inthe note on the methodology section in the IT
(http://www.irishtimes.com/newspaper/ireland/2010/0611/1224272271214.html?via=rel).
I assume the calibration model works by weighting the data somehow.
Does the static model downweight deviations away from the historical mean levels of party support?
Would a ‘dynamic’ model only account for more recent polling information or progressively discount older polls/election results?
At the moment is the calibration system downweighting statements of support for FF and overweighting statements of support for Labour?
If the calibration system is proprietary, I guess we’ll struggle to understand these questions.
Given this, when the new calibration scheme is up and running, it may become very difficult to distnguish changes in the preferences of the population (i.e. a dip in Labour support) from changes arising from the new calibration scheme.
Actually – thinking about it – if the static calibration scheme works more or less as described above, could it actually be underestimating Lab support (which is now way above its historical average) and overestimating FF support (which is now way below the historical average)?
If so, could the re-calibration actually see a further INCREASE in the Labour vote estimate and a further drop in the FF estimate?
The TNS MRBI poll findings have long been in doubt due to the downward adjustment made to the Fianna Fail vote on the basis of its “big brand” effect. For example at the last local elections the final TNS mrbi pre-local elections poll gave Fianna Fáil only 18 per cent of the votes, in the event it secured 25 per cent. It is also consistently has Labour higher than near contemporaneous Red C polls. Nonetheless, despite these caveats, the trend is clear. Fianna Fail is dropping support and the Labour Party is picking it up. It is quite possible, and indeed likely, that many Fianna Fail voters are heading towards Fine Gael who are simultaneously losing voters to Labour. This is strikingly similar to the pattern in advance of the Spring Tide in 1992. At that time, Reynolds had replaced Haughey and continued the long term decline in the former’s leadership ratings. Fine Gael also had a new leader in John Bruton who had replaced Alan Dukes only to resume the inexorable decline in his leadership ratings. Spring by contrast saw a general upward trend in his leadership satisfaction ranking over the three years leading to the election. Focus groups at the time, as David reported in How Ireland Voted 1992, found that voters were more concerned about leadership than they were about policy. Spring was perceived as a strong leader who could keep his party under control. Dukes on the other hand faced down neither the left nor the right of his party while Bruton was associated primarily with the right. This probably led to the notable shift of the liberal vote to Labour at that election, perhaps the most obvious pattern in voting that year. At the same time all the parties were focusing on policy and particularly unemployment and job creation. Change was the dominant theme of the campaigns (sound familiar?). Nonetheless, research later found that much of the talk of change from Fine Gael was perceived to be empty rhetoric. Fine Gael could do worse than going back and looking at that campaign in detail.
See here for further analysis of the opinion poll and how it would translate in terms of seat shares, by Adrian Kavanagh (NUI Maynooth).
Given that the poll was carried out in the context of the publication of reports blaming (at least in part) leading FF personnel for the post-Celtic Tiger economic collapse, the poor showing for Fianna Fail cannot have been unexpected. If anything, it’s a worse result for Fine Gael, who face the possibility of being eclipsed by a surging Gilmore Gael and who really should be doing better than standing just 1% above their 2007 support levels.
It’s fair to say there’s a good deal of scepticism as to whether those figures would materialise in an actual election, given (a) the inadequately justified ‘adjustment’ applied by IPSOS / MRBI, which seems especially inappropriate now that we’d expect an under-estimating of FF support rather than an over-estimating as in the past; (b) one poll could always turn out to be a rogue; and (c) as Jane S observes, given the track record we would want to see any poll finding validated by an actual election result somewhere to feel confident about it.
Nonetheless, even the unadjusted figures make Labour the strongest of the three parties (29% compared with 27% for FG and 21% for FF), and that does the raise the question of whether Vincent Browne’s hope of a left-wing government, discussed by Michael Marsh in his 9 June post on this site, could be somewhere close to realisation.
Sticking with the adjusted figures, Labour on 32% and SF on 9% already takes the putative alliance over the 40% mark. SF voters are likely to transfer to Labour more strongly than to any other party, though we wouldn’t expect reciprocation under present circumstances, and Labour will benefit to some extent from FG transfers. Add in other left-wing TDs – Joe Higgins, maybe one or two other Socialist Party TDs, possibly a PbP TD or two, Finian McGrath – and the sum total might not be far short of the requisite 83. Perhaps the Greens would be a potential part of the alliance too: with the current government having no chance of re-election, it’s unlikely that the Greens will campaign for its return but rather, during the next election campaign, they will operate as free agents open to participating in any government receptive to their policy agenda, just as they did in 2007.
Or perhaps there’s an air of unreality about this, in that SF may still not be regarded by any of the other parties as coalitionable, a coalition made up of, or at least reliant on, so many elements would look anything but stable, and as Michael Marsh observed the left remains unusually weak in this country in terms of political culture as well as party strength.
But it does highlight the question of just what government will emerge after the next election. FG seem to have convinced themselves, in the face of a lot of evidence, that a FG majority government is somehow a possibility, and as a consequence they’ve taken up policy positions with no regard for these policies’ post-election compatibility with Labour policy and in some cases have expressed views that would be pretty much unacceptable to Labour. A FG–Labour (or should that now be Labour–FG?) coalition is still very much the bookmakers’ favourite but there would be a lot of internal tension, in terms of ideological incompatibility rather than just differences over the details of day-to-day policy. In policy terms FF might be a more compatible partner for both FG and for Labour than the other main opposition party is, but almost certainly both FG and Labour will explicitly rule out any coalition with FF after the next election, and they know they would be lynched (metaphorically) by their supporters if they did help put FF back in government.
As things stand the next election looks like being one of those where the voters can have no real idea what the possible governmental options are. Maybe we will see a clearer definition of alternatives as election day approaches.
MRBI should have ditched the calibfrations last autumn when FF dropped to 17% for the 1st time and when their numbers were shown to be so out of kilter with the local and European election results.
http://www.irishelection.com/2009/05/irish-times-poll/
Fianna Fáil 21% (-1)
Fine Gael 38% (+6)
Labour Party 20% (-4)
I would agree that FG and Lab are now in a head to head contest as to who is the largest party in terms of votes with FF behind them but when it comes to the election the FF influences of incumbency and local patronage will come into focus again. It’s a 3 way contest for sure just not with Labour some way out in front as this poll suggests.