Sunday Business Post-Red C poll 10 April; Kenny-a believe it?

Adrian Kavanagh 14th April

A new Dail and a new series of opinion polls leading up to a future general election in 2015 or 2016! As the constituency level poll analysis held up remarkably well in light of the eventual election results (and the last analysis involving the results of the Millward-Brown poll of 23 February was remarkably close to the actual result in terms of seats allocations), it is interesting to see what seat allocations would be based on the first of the post-election polls, the Red C poll published in last Sunday’s (10 April) Sunday Business Post.  This estimated party support as follows: Fine Gael 39%, Labour 18%, Fianna Fail 16%, Sinn Fein 11%, Independents and Others 16% (including 2% for the Green Party). What is striking about these figures is the degree to which they largely mirror the February 25th election results, but when seats are allocated based on the constituency-level poll analysis model the difference is somewhat more pronounced – this estimates seat levels as follows:  Fine Gael 80, Labour 29, Fianna Fail 22, Sinn Fein 13, Independents and Others 22 (including 6 seats for United Left Alliance candidates). When the effect of vote-splitting between multiple candidates is accounted for, particularly amongst the independent and Fianna Fil ranks, the numbers would probably read as Fine Gael 80, Labour 33, Fianna Fail 20, Sinn Fein 15, Independents and Others 18 (including 5 seats for United Left Alliance candidates).  

The main difference in this poll analysis is that the constituency estimates are now based on applying the percentage change between the poll results and the February 2011 results, whereas previous analyses were relating poll figures to the May 2007 results.

Based on these Red C poll figures the model estimates party support levels by constituency as follows:

Based on these constituency support estimates, the destination of seats by constituency would be guesstimated as follows (looking solely at first preferences and not considering likely transfer patterns – allocating seats based on the D’Hondt method):

This is a very rough model based on a “uniform swing” assumption – assuming that the national swing between the general election and the Red C poll would be replicated exactly in each constituency. In the pre-2011 General Election analyses the model did tend (or seem!) to over-inflate constituency support estimates in constituencies where parties/groupings were already starting from a very high base (e.g. Fine Gael in Mayo, Labour in Kildare South, Others in Dublin Central, Galway West, Tipperary North and Wicklow).  But, in terms of the overall national estimation of seat estimates this was compensated for by the under-estimation of support levels for those parties in other constituencies (e.g. Fine Gael in Dublin Central and Wicklow, Others in Dublin South and Wexford). The award of an extra seat to a party, or grouping, in an “over-estimated” constituency may thus be compensated by the “non-award” of a seat in an “under-estimated constituency”.

Seat estimates for the Others grouping are over-estimated by the model, which simply allocates seat estimates based on the estimated constituency first preferenence vote based on D’Hondt principles. In certain constituencies, seats allocated to this grouping are based on a 2011 General Election figure for this grouping involving multiple independent and smaller party candidates (including the United Left Alliance). In the case of Laois-Offaly, the allocated seat here is based on a 20.4% share of the vote won by eleven candidates in the February election, with only three of these winning more than 1% of the vote (Leahy 6.6%, Foley 6.0%,  E Fitzpatrick 3.4%) and in reality this seat would more likely fall back to Sinn Fein or else go to a third Fine Gael candidate.  Other constituencies that would fall into this category include Clare, Cork North Central, Dublin South East, and possibly also the two Galway constituencies, meaning that the overall number of seats for Independents and Others can probably be estimated, based on these poll figures, as lieing between 16 and 18 seats.  In a similar vein, the model estimates of narrow Fianna Fail seat gains in some constituencies can probably be discounted where these narrow wins are based on a Fianna Fail general election vote share won by two, or more, party candidates (and also if unfavourable vote transfer patterns are factored in) - constituencies that would fall into this category would possibly include Cork East, Cork South West, Donegal South West and Sligo-North Leitrim, so a more realistic estimate of their seat tally would probably fall between 18 and 20 seats.

The biggest change between the election results and the model’s seat allocations involves Labour, which is seen to win 8 fewer seats, despite experiencing a loss of support of just under 1.5%.  In part this is a function of the model’s tendency to over-estimate independent and Fianna Fail seat allocations, as discussed above, and some of the seats viewed as being lost by Labour to these other political groupings would in reality remain with the party based on the the poll levels of support. This also highlights the transfer-’ttraction of Labour and highlights the degree to which the party can enjoy a seat bounce based on their ability to draw transfers from a wide range of different political parties and groupings. But the model also highlights just how marginal Irish electoral contests are, especially given that seats were won or lost in February 2011 by margins of less than 1% of the vote in over one third of all the constituencies, and shows that even a relatively small loss of support in percentage terms could see amount to relatively significant seat losses especially if luck is not on your side and major competitors for your sections of the vote base (Fine Gael and Sinn Fein) are gaining in support.