Projections from poll findings into seats

By Michael Gallagher

In essence there are two ways of trying to convert opinion poll findings regarding voting intentions into seat totals for the political parties. One, employed in the earlier post by Adrian Kavanagh (5 Feb, two posts below this one), is to assume uniform swing across the country; apply this to the party vote shares across each of the 43 constituencies to arrive at new predicted vote shares in each if them; make inferences about what this vote distribution within each constituency would mean for the allocation of seats there; then add up the constituency totals to produce an overall national result.

The other, outlined here in more detail on 9 December last, is not to try to predict the result in every constituency but, rather, to try to convert national vote shares into national seat totals. This is not an exact science; the relationship is affected by the small constituency size employed in Ireland (on average fewer than 4 TDs per constituency, an exceptionally small number for a proportional representation system), which creates the potential for disproportionality and in particular makes life more difficult for smaller parties and independents; by the impact of transfers; by candidate selection strategies; and by the distribution of votes among a party’s candidates.

As a result of this, two parties, or one party at different elections, might win the same number of first preference votes yet different numbers of seats. At recent elections FF, thanks mainly to losing its previous ‘transfer-repellent’ status, has received a large benefit: thus in 2002 it received 49 per cent of the seats with 41 per cent of the first preference votes, and in 2007 its vote share was 42 per cent and its seat share 47 per cent. FG received a record ‘bonus’ in 1997, with 33 per cent of the seats from 28 per cent of the first preference votes, and a record under-representation in 2002, when its vote share was 22 per cent and its seat share 19 per cent.

Summarising from the 9 December post, the national-level seat projection method assumes that FF will receive more or less its proportionate share of seats, benefiting from being (still) one of the largest four parties but suffering from the likelihood that other parties’ supporters will use their lower preferences against it; FG will benefit from its size but will not attract many transfers except from Labour voters; Labour will attract transfers from across the board, but it remains to be seen how the party copes with the tensions and difficulties that can arise when a party expands its number of candidates; Sinn Féin will be more transfer-attractive than ever before but many of its candidates are relatively unknown.

This gives a set of steps that lead to national-level projections, and employing these the Red C poll figures published in the Sunday Business Post on 6 February suggest national seat totals of around FF 29, FG 62, Lab 40, SF 22, Others 13. Comparing these with those from Adrian Kavanagh’s 5 February post, this approach suggests fewer seats for FG and more for SF, with FF and Labour around the same by each method. For more information about projections from other polls during the campaign and general discussion of the election, see this site.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s