Latest poll September 28, 2013
Posted by WorldbyStorm in Irish Politics.trackback
Farewell the Seanad? Well according to the latest Sunday Independent/Millward Brown poll:
When the undecided voters are excluded the margin is 64% in favour of abolition and 36% against the Government proposal that will be put to the people on 4 October.
And the party figures?
Fine Gael 28% (+1)
Labour 10% (+2)
Fianna Fáil 27% (+2)
Sinn Féin 19% (-2)
Independents/Others 15% (-5)
Hard to understand Ind/Other drop, particularly when set against other polls. Will have to look at the details.
The Cedar Lounge Revolution 
I would love to see a poll measuring support for 1 or 2 chambers.
LikeLike
The opaque nature of Millward Brown’s polling means we can’t see what their problem is but at a guess I’d say they like B&A have a ABC1 heavy sample that they are either weighting insufficiently or not at all. Also the large number of undecided responders reduces the reliability of the headlines figures as they’re taking a lot of people out of the sample by excluding them; I haven’t seen the number of undecideds for this poll but based on their previous efforts I’d say the margin of error is about 3.8%.
LikeLike
In an ironic kind of way I think that the undecided, who are always excluded, will decide this referendum.
I don`t have any facts to back this up, I can only say that I have not excluded any voter, even those who say they will vote one way and then vote the other. Or even those who say they will vote and then don`t.
Would my reasoning be counted as a valid poll?
LikeLike
It’s impossible to know fully what their going to do on polling day them which is why the pollsters exclude them as their looking for the “election result” as opposed to actual public opinion. Having said that you can ask the undecideds what are called squeeze questions to try and flush out their potential choice but those are maybe not entirely reliable.
LikeLike
To be really honest, I think that when people say ” I am undecided ” they really mean ” why should I tell you my preference “, it`s really nobodies business but mine, go ask some other fool who will answer you in a dishonest way in order to mislead your results.
I have only ever once been polled here in Limerick, I declined to answer any of their questions, so they told me they would put me as undecided. How can they get any accuracy with that method?
Maybe it`s because I am getting old but I have always felt that the only poll that really mattered was the one at the polling booth. These other polls are a device used to influence those that declare that they are undecided.
Whether it works or not has never been the subject of a poll or has it?
LikeLike
I understand that this is a tracking poll and not a new sample of people every time. That would mean firstly that any error in selection of the main sample group would be replicated time after time and secondly that there would be smaller changes in the results as people tend not to change their minds very much. Dramatic changes in opinion in my recollection only tend to be picked up with the more expensive originally selected groups each time.
LikeLike
Using tracking over such a large time period and small sample surely would be ludicrously likely to produce meaningless results as it wouldn’t represent any coherent “snap shot” of public opinion. Where did it say it was a tracking poll?
LikeLike
I assume it’s only a ‘tracking poll’ in the sense that they’re repeating the poll at regular intervals and with consistent methodology, in order to track changes in opinion, rather than it being the same sample. Certainly that’s what RedC aim to do in focusing on the trends rather than just individual poll results, which are inherently variable. So even if you can identify biases in their polling (which regularity actually might make it easier to do) it should at least be consistent with itself.
LikeLike
I think it is worth pointing out that richotto is slightly wrong in that tracking polls only use a portion of the previous polls data. So a poll that includes Monday and Tuesday would be followed by a poll that includes Tuesday and Wednesday. But if you’re doing that over two to four weeks it’s got to be regarded as barking mad as public opinion might have changed substantially over that time period; same applies to solo polls conduced over lengthy time periods.
LikeLike
Seems to me that it would be more expensive and troublesome to try and pin down the same people you polled last time. Also, companies tend to save money by tacking the political attitudes survey onto a more mundane survey on car insurance, banking, grocery shopping etc. In any event,you are still going to have to go to the trouble of contacting 1000 people in your stratified, multi-stage, cluster sample.
LikeLike
It shouldn’t be that difficult actually as you’d already have their address if done face to face or their number if done over the phone; the real trouble would be getting them to agree to do it again.
LikeLike
Yeah, I think you’re right re it being barking mad Liberius.
LikeLike
One further point, the tracking polls were the main offenders in failing to predict the outcome of the last US Presidential Election. The two main national ones were well outside the stated margin of error. Polling companies may be trying to cut corners with this method being employed more.
LikeLike
[…] at odds with Millward Brown’s effort last week but there will be plenty of dark muttering around Leinster House this afternoon. Micheál Martin […]
LikeLike
[…] these figures with the earlier Sunday Independent/Millward Brown poll and the averaged-out “banding” of voter preferences becomes obvious: FG and FF on plus 20%, SF […]
LikeLike