Local Heroes
While the IPR spreadsheet extrapolates from national polls, there’s always been a (justifiable) argument made that things will be different in the individual constituencies. I’d argue that these differences tend to cancel out on a national basis, although of course, in the case of small parties in particular, you can be lucky or unlucky in that.
There’s been a number of constituency polls using mock ballot papers over the last few days from reputable companies (as opposed to other polls where IT students did a project in some areas for the local rag), and rather than pore over each of these, given they generally have individual margins of error of 4.5%, I thought it would be interesting to look at a composite of these (which would have a smaller margin of error) and compare this with the most recent IPR projections in those constituencies.
The polls in question provided the following figures;
FF (%) | FG (%) | LP (%) | GP (%) | SF (%) | OTH (%) | |
Cork NC (average) | 14 | 23 | 30.5 | 1 | 14 | 17.5 |
Cork SC | 27 | 36 | 19 | 3 | 8 | 7 |
Donegal NE | 15 | 29 | 15 | 0 | 28.5 | 12.5 |
Dun Laoghaire | 18 | 32 | 32 | 4 | 0 | 14 |
Galway W | 19 | 31 | 12 | 3 | 6 | 29 |
Kerry Sth | 15 | 31 | 15 | 0 | 1 | 38 |
Tipperary Sth | 10 | 31 | 14 | 0 | 5 | 40 |
Average | 16.86 | 30.43 | 19.64 | 1.57 | 8.93 | 22.57 |
So how do these stack against the most recent IPR projections? These were as follows;
FF | FG | LP | GP | SF | OTH | ||||||||
Cork NC | 9.80% | 40.40% | 22.00% | 2.00% | 11.30% | 14.50% | |||||||
Cork SC | 17.30% | 41.40% | 17.20% | 5.00% | 7.30% | 11.80% | |||||||
Donegal NE | 21.20% | 33.10% | 12.80% | 1.10% | 22.80% | 9.00% | |||||||
Dun Laoghaire | 11.70% | 28.80% | 31.70% | 4.60% | 3.60% | 19.60% | |||||||
Galway W | 12.20% | 35.90% | 15.20% | 4.30% | 5.80% | 26.60% | |||||||
Kerry S | 15.50% | 31.10% | 16.90% | 0.50% | 5.90% | 30.10% | |||||||
Tipperary S | 13.50% | 29.50% | 14.40% | 0.50% | 5.20% | 36.90% | |||||||
Average | 14.46% | 34.31% | 18.60% | 2.57% | 8.84% | 21.21% |
Well, interestingly, pretty close overall. In these polls, FF do 2.4% better, FG 3.9% worse, LP 1% better, GP 1% worse, SF 0.1% better and OTH 1.4% better. Very odd, given the protestations from some that the spreadsheet is biased against FG….
The main cause of FG’s underperformance in the polls relative to the projection is the retirement of Bernard Allen in Cork NC, and my underestimation of the individual effect this would have on their fortunes there. Interestingly though, what deviation there is in the other constituencies is also largely downwards. Even taking CNC out of the equation, the average FG vote in these polls is 31.7%, compared to 33.3% in the spreadsheet projections for those constituencies, which may temper the expectations among some that FG will significantly over-achieve compared to their national polling. They may of course do better on the day, but if they do, it will be the result of a national swing to them, rather than 43 cases of ‘local factors’ all going their way.
FF will draw some comfort that they are polling slightly higher in the constituency polls than in the projections, but given 2 of the 7 polls are in Cork City, this is likely to be at least in part a result of the Martin factor (a local langer leap) being underestimated by me. Interestingly, only two of the seven constituencies sees them make a quota despite Cork City being there twice, although they could be expected to get 4 seats out of these constituencies on these figures (CSC certain, Galway W pretty certain, Dun Laoghaire very probable, and touch & go in CNC and DNE), which is 1 seat more than I call it in the spreadsheet, so that couple of percentage points could make a big difference to them on the day. The big issue for them will be the 20-30% reported abstentions in these polls – whether these are “shy FFers”, or anti-FF voters who have yet to decide the flavour of their opposition, is the factor that is likely to seal their fate.
Labour should be happy enough with these polls overall. They’ve only 3 TDs running in these 7 constituencies, and their vote is 1% higher than the most recent projections. If I had to call the seats on these 7 polls, I’d give LP 7 seats , although 9 would be possible. This compares slightly favourably to the projections, where I have them getting 6 seats. Certainly, if offered these FPV shares at this stage of the campaign, they’d probably take them (if over the objections of Paula Desmond, who will be disappointed with the CSC figures).
After the largest 3 parties, it becomes harder to call given the margins of error, but it is noticeable that the SF vote is 0.1% higher in the polls than the projections, indicating that they should be able to translate their national fortunes locally, provided they get their voters out. The Greens however will be concerned that they under-perform in all 7 constituency polls, particularly as they are running an outgoing Oireachtas member in half of them, and so these are constituencies where their candidates are well known.
OTH do slightly better, although the nature of these candidates makes it hard to extrapolate nationally. It does suggest though that lack of high profile candidates will not impact on their appeal at the ballot box.
That’s all for tonight, I’ll have something for you tomorrow after the RedC poll breaks.
D
30% average for FG in the polls is very interesting. If the party is pushing 40% nationwide then there must be some serious surprises in store in certain constituencies where the FG swing will be massive!!
David Higgins
February 18, 2011 at 11:20 pm
Certainly it would appear that they must be pulling in the votes somewhere, and if they’re not showing up in these constituency polls, you’d expect them to be somewhere. It doesn’t appear to be Dublin, given the poll their recently, so the question is where? I think there’s a few more constituency polls due over the next few days (heard there’s one in Mayo) which might answer it….
Dotski
February 18, 2011 at 11:28 pm
Thanks Dotski, excellent analysis. I’m learning.
Toland
February 18, 2011 at 11:21 pm
This is a very interesting analysis. That FF is doing better on the averaged local polls while FG is doing worse, suggests that when transformed into seats FF should get more seats than originally expected and FG getting fewer.
Don Din
February 18, 2011 at 11:24 pm
Possibly, although the Cork factor might be disproportionate, and there may be some area that FG are doing much better (e.g. Connaught, FG are above average in Galway West, and the more rural parts of Connaught may be swinging to FG even more)
Dotski
February 18, 2011 at 11:32 pm
I wonder if we’re seeing the start of a reaction to the possibility of Fine Gael winning an overall majority. Elsewhere there’s often been a disinclination to give the leading party too much power – never mind that to many Irish voters that may sound like repeating the big mistake of the past.
The other query I’d have would be what happens to all those ‘Others’ votes, especially where there’s no obvious TD to be elected. Are many of them so alienated from all the Parties that they’ll just stop transferring? Or will they benefit the main Parties in unexpected ways? (Yes I know every constituency will be different, but still, it’s a lot of votes).
Roger Mexico
February 19, 2011 at 12:36 am
Very poss, but I think there’s 2 reactions to the possibility of a FG only govt. The negative as you say, where some ppl will be concerned, but there’ll also be others who want to “join the winning team”. I’ve no idea where the net effect will lie, but suspect it might leave them about level.
On OTHs, I suspect that in real life they won’t use as many preferences as they did in these polls – knowing someone is hearing how you vote means your more likely to use the preferences I think, as you might “feel you should”. I suspect that they will in many cases stop at pref number 4 or so, seeing a greater seat bonus for the other parties (in Kerry South this may benefit LP, for exmple, in DSE it could benefit FG.
Dotski
February 19, 2011 at 11:16 am
I suspect that they will in many cases stop at pref number 4 or so, seeing a greater seat bonus for the other parties (in Kerry South this may benefit LP, for exmple, in DSE it could benefit FG.
Would that be easy enough to check? I’d hav thought looking at the non-transferable votes in later counts give a reliable figure for those who stop (at least, for those who didn’t vote for one of the leading candidates).
[Or, for the hi-tech, some parties could send in tally persons with cameras on lapels to film a sample of ballot papers as they are opened or at some other stage in the count to see where people stop.]
Tomboktu
February 19, 2011 at 12:48 pm
It’s be easy in elections with only a few candidates but when there’s 12+ it’d be impossible – the fact that a vote didn’t transfer when someone has already been eliminated means that you’ve no idea whether they transferred to them or not. After a couple of counts wouldd be impossible (unless transfer rate was very high, so you could assume then it was no lower).
From my general experience of tallying (many years ago) I’d have thought that about a third of papers vote for 3 or less candidates, about a third for between 4 and 9-10, and the remainder go practically all the way, but I’ve no data to support that – would anyone else here who has tallied comment? (My experiences were in Dublin)
Dotski
February 19, 2011 at 3:32 pm
Hi,
I have never tallied but I have examined the electronic voting results from 2002 (Meath, Dublin North and Dublin West) and from these it seems that most people vote 1, 2 and 3 but it falls off very rapidly after that.
You can see some graphs at http://guthanphobail.net/scrudu02.htm .
It lets you see graphs of the 2nd and subsequent preferences of those who gave their first preference to any candidate or party. You can also, for example, look at those who used all their preferences giving the last preference to Labour and see what their first preferences were, etc. Labour and the Greens may be marginally more likely to use more preferences than FF or SF with FG somewhere in between. It would be interesting to see if there is a different pattern when a party runs 4 candidates as FG are doing in Laois-Offaly and Galway West.
Whether people behave the same way with electronic (or postal) voting as they do with a normal ballot paper is another question…
CQ
February 19, 2011 at 6:35 pm
On the subject of a possible FG majority, when it looked like Haughey or Ahern were on the verge of achieving that there seemed to be a reaction against that prospect. Things could be different now, for all his faults Enda doesn’t have the “colourful” background that Charlie and Bertie did so perhaps the prospect of an FG majority doesn’t frighten people in the way that an FF majority did?
Personally I don’t think FG will get a majority, they could well be in the mid to high 70’s allowing them to have a minority government if they wanted. Tonight’s polls will be very interesting!
The Oncoming Storm
February 19, 2011 at 1:41 pm
Indeed. It will be interesting if they both come out around the same time, and if they tell the same story, as they’ll set the tone for the last week….
Dotski
February 19, 2011 at 3:33 pm
I have been shown suggested projections for Dublin West, based on canvas returns and projections from the local authority elections, which showed the following figures;
Leo V 27% (FG)
Joan B 21% (Lab)
Joe H 16% (Trot)
Brian L 13% (FF)
Patrick N 10% (2nd Lab)
Kieran D 6% (2nd FG)
Paul D 4% (SF)
Others 3% (Green, 2nd FF, Ind)
I wonder Dot, do they vary much from your projections for the constituency? The SF vote is low because of the Joe Higgins factor and a new candidate. The LP are trying to split their vote, giving some of their
strongest areas to the Patrick Nulty. This was not one of their expected hits, so they must be reasonably happy.
Old liver & chips might yet hold on if the transfer rates inside FG are poor. I can’t see a strong transfer rate from FG to Lab, if Nulty stays ahead of Dennison.
The key area is River Valley and surrounds, the bit of Swords added on. There are solid Labour & SP votes there and no Lenihan family tradition. FF only received 16% of the vote in Swords LEA back in the heady days of 2009.
Vote management by Labour or FG could produce results, but if Dr. Leo pigs out in first prefs they have no hope. I could see some leakage of votes back from the residents of The Georgian Village & Deerpark back to the local boy Brian, letting him off the hook.
Niall
February 19, 2011 at 7:27 pm
Hi dotski,
There have been four polls done in Kerry South in the past couple of weeks, all with a sample size of less than 500. How reliable would it be to add these four polls together? Obviously a poll with sample size of close to 2000 in a three-seater constituency would be a reliable poll with small st dev. But since the four polls were carried out by three different polling organisations with different methods of collecting data, would the figures generated from this exercise be more credible than one 500 smaple size poll. Of course I would have to assume that no voter was polled in more than one poll, otherwise thier vote would count as double in my ‘poll’. I think the best way of calculating FPV% is as follows: If a candidate got 50 out of 400 in poll ‘A’, 70 out of 500 in poll ‘B’, 40 out of 400 in poll ‘C’ and 80 out of 500 in poll ‘D’, then I would give them 240 out of 1800 or 13.33% of FPVs and so on. I may also try to track down the transfer details from the polls and do out a count by count calculation. I understand the limitations of relying on transfer data in polls to predict what will happen but it may still be interesting.
The four polls are:
Kerry’s Eye / IT Tralee – 3rd Feb.
Sindo / Quantum – 13th Feb.
The Kerryman / IT Tralee – 16th Feb.
Irish Examiner – RedC – 17th Feb.
Any thoughts or advice before I crunch the numbers would be appreciated.
Regards,
Padraig
Padraig
February 20, 2011 at 11:55 pm
Hi Padraig,
Rushing off to a conference, but I’d be somewhat wary of the QR poll as that may not even exist, but otherwise it could be an interesting exercise – let us know the result?
Ta,
D
Dotski
February 21, 2011 at 9:11 am