More lessons in the Belfast Telegraph poll…

7 views

The detailed results of the Belfast Telegraph poll into current Assembly voting intentions were published at the weekend www.lucidtalk.co.uk We can be sure that the parties will be poring over them. They will translate the changes in percentage support into votes, and then project that into individual Assembly constituencies, and also look for lessons for their next Westminster and European campaigns.

Here’s what they will find:

• The UUP appears to have lost over 14,700 voters since the last Assembly Election. A drop of 16.8%.
• The SDLP drop is almost as great – at almost 14,600. A loss of 15.5%
• The Alliance gain of over 21,200 has therefore come at the expense of both of these parties. Up 41.8%.
• Sinn Fein and DUP have consolidated and improved their positions, with increases of about 7,100 (up 3.9%) and 4,700 (up 2.4%) respectively.
• A small element of these gains, about 7,000 votes, is at the expense of independents and minor parties.
• The SDLP hold on the Westminster seats of Foyle and South Down is under threat.
• The DUP could take the UUP seat in Europe by running a second candidate next year, and could face political risks in the future if it fails to do so.

Implications for the Assembly

The potential political impact of the vote changes is best expressed by considering how different the Assembly would look if the election were run on the same boundaries as last time. The effects of boundary changes, reducing the number of constituencies by two, and possibly the number of MLA’s per constituency to 5 are separate issues.

DUP: either no change or, just possibly, 1 seat gained
UUP: losses of between 2 and 5 seats (most likely 4)
Alliance: gain of between 4 and 6 seats (most likely 5)
SDLP: net loss of 2 or 3 seats (most likely 2)
Sinn Fein: net gain of 2 seats

This would leave Alliance with between 12 and 14 seats (most likely 13), possibly overtaking both the SDLP on 11 or 12 (most likely 12), with the UUP at between 11 and 14 (most likely 12).

These figures have been arrived at by looking at each constituency individually. In each constituency I have applied the total percentage change to each party’s support shown by the poll at the overall Northern Ireland level. Where I refer to a quota, this is always first preference votes, unless stated otherwise.

In 10 constituencies no seats would change hands. The battlegrounds are:

Belfast East: With Alliance up to 2.61 quotas, and UUP reduced to 0.56, transfers will probably give Alliance a third seat (up 1) at the expense of the UUP, although it is theoretically possible that the UUP could hold on.

Belfast South: With 1.96 quotas Alliance takes a second seat (up 1) from the SDLP.

East Antrim: Last time Alliance won one seat with 1.02 of a quota. However, because of the way the votes split between the parties, 70 more votes would have given Alliance a second seat. With the poll pointing towards an increase of 1,860 1st preference votes, Alliance would gain the second seat at the expense of Sinn Fein.

East Londonderry: This is probably no change. But it is interesting. Last time the Independent McClarty had 0.61 of a quota. With Alliance rising to 0.54 of a quota and the UUP dropping to 0.49 it is likely that the transfers keep the seat with McClarty, but it depends on how the UUP transfers fall.

Foyle: Sinn Fein gains one from SDLP. Last time the SDLP had 499 more first preference votes than Sinn Fein, but this extrapolation of the poll gives Sinn Fein a 1,700 vote lead.

North Down: The calculations pointing to 1.85 of a quota for Alliance, they would gain another seat. The poll suggests that support for the Greens has grown strongly, although it is impossible to calculate a precise percentage since they did not stand in all constituencies last time. So the Alliance gain would probably come at the expense of the UUP.

Strangford: Last time the UUP took two seats starting from 1.43 quotas. This time they would drop to 1.19, with Alliance rising, coincidentally, to 1.43. So the UUP will drop 1 of their two seats. The SDLP would start on 0.52 of a quota. On the evidence of the last election they should pick up about two thirds of the Sinn Fein transfers, taking them to about 0.7. I’m calling them to gain the seat on the last count without reaching the quota. If Alliance could balance their two candidates well they might just keep them both above the SDLP long enough to take the seat on DUP and UUP transfers, of which there are plenty available since the DUP have half a quota to spare on top of the UUP’s excess of 0.19.

Upper Bann: The poll points to two changes here. Last time Sinn Fein only managed to win only 1 seat from 1.9 quotas. This time with the figures pointing to 1.98 quotas they should take a second seat, at the expense of the SDLP, who would start on 0.70 of a quota. With the DUP on 1.95 quotas, the UUP on 1.43 and TUV on 0.17, there should be sufficient transfers around to keep Alliance, starting on 0.65, ahead of the SDLP to take the seat which the UUP seem destined to lose.

West Tyrone: The most likely outcome would be no change, although there is the possibility that the DUP takes the UUP seat.

Implications for Westminster

The poll asked about voting intentions in an Assembly Election, not Westminster. We also know that some people vote differently in elections for Westminster, local council, Assembly or Europe, even where two of these elections are held on the same day. But while we cannot extrapolate directly from these responses it is significant to note the Foyle figures referred to above, which point to a higher level of vulnerability for the SDLP in that Westminster constituency. In South Down the tide would also have turned against the SDLP on these calculations, with a 2,040 vote advantage to the SDLP in the last Assembly elections turning into a short lead of 303 for Sinn Fein. With the South Belfast constituency expected to disappear, the SDLP faces a much tougher task to hold onto the two Westminster seats which would remain to them. They may be able to “borrow” enough tactical votes from other parties’ supporters to hold on in South Down, but it is hard to see where to look for enough willing lenders in Foyle.

Implications for Europe

Again it is worth stressing that the poll question is not about a European Election. But the DUP will not have lost sight of the fact that this poll shows the DUP Assembly vote (30.7%) now almost three times that of the UUP (11.0%). They could take the UUP seat simply by putting up a second candidate, and managing the vote division between the candidates reasonably well.

That in itself would be strong incentive for powerful voices within the party to argue for running a second candidate next year, and plays to the DUP strategy of presenting itself as the bulwark against Sinn Fein. Can it resist the optics of defeating Sinn Fein 2 to 1.

But there is a longer term strategic issue. Alliance has always polled below their strength in European elections. Quite simply they have always lacked credibility in a poll where the quota is 25% of the vote. And this has benefited the UUP and SDLP, and will probably continue to do so in 2013. But a 2016 Assembly Election in which Alliance win more seats could solve their credibility problem. Then the 2017 European Elections become more open.

Applying the recent Belfast Telegraph poll shares to the 484,572 votes cast in the last Euro Election gives:

On this basis the UUP can no longer be relied upon to be strong enough to take the second “unionist” seat. Indeed a second “unionist” seat is far from inevitable. Better for the DUP to establish the precedent for running a second candidate in 2013 when it is safe to do so, than wait until 2017 when running 1 candidate risks losing a “unionist” seat, and running 2 candidates against an incumbent UUP risks taking the blame if that second seat is lost anyway.

, ,

  • Drumlins Rock

    mjh, who are you? your profile is blank and you have never commented on Slugger before, like to know the angle of the person posting first before commenting.

  • http://fitzjameshorselooksattheworld.wordpress.com/ fitzjameshorse1745

    Fantasy Politics Im afraid.

    There are no votes “gained” or “lost”.
    Simply put ….if the Belfast Telegraph conducted a poll asking “Did you brush your teeth this morning?”……99.99% of respondents would say “yes”.
    But most dentists would say this is highly unlikely.

    The worst possible way of measuring dental hygene is to conduct an opinion poll. And the worst possible way of measuring voting intentions is an opinion poll.

  • Lionel Hutz

    The Alliance partyhave gained 41%. Lmao. Why? What could possibly have caused this. Surely you would look at this and conclude that it’s a mistake…

  • Drumlins Rock
  • keano10

    Essentially it’s a major slap in the teeth for all those (including most of those on this site) who are never-done criticising the current main incumbents of local power, Sinn Fein and The DUP. Often without much tangible evidence to back up their arguments.

    If this poll were to be replicated, those parties would be attracting a combined total of 58.7% of the voting electorate. Thats an enormous mandate and the fact it appears to be ever-increasing, shows beyond doubt that people here are happy with how those parties are governing.

    The leadership from both MMG and Robinson is clearly working and I cant see any joy for the likes of the UUP and SDLP in making sniping personality attacks on either of them. In fact, it’s hard to guage where the SDLP or UUP can make any impact whatsoever. Both of those parties have witnessed a number of their most high profile candidates efefctively drop off the radar lately and both are in pretty deep trouble with no obvious visible escape route.

    Alliance may well make further dents in the SDLP vote and the SDLP is rapidly becoming a lost cause in many constituencies. McDonnell’s abrasive and tetchy persona will hardly help to attract those floating voters who prefer to vote for personalities as opposed to the finer political intracacies.

    As for the UUP, it’s gonna take all of Nesbitt’s media experience to stop the rot. He needs to get a few no-nonsense policies placed on the table pretty pronto, going by this poll.

  • Drumlins Rock

    FJH, I love the next question in the poll, talk about a leading question this one takes the biscuit “The NI Assembly has a higher number of members at 108, than other UK regional assemblies, would you like to see the number of MLAs A-stay the same, B-reduced 90-108, C-below 90, D-Abolish Stormont. ”

    Also it is a lie as the Scottish Parliment has 129 members, and the suggestions don’t make sense otherwise, the real options are reduce to 96 (likely) or 80, possible long term.

  • http://www.banuanlae.org/ Ulick

    “The Alliance partyhave gained 41%. Lmao. Why? What could possibly have caused this. Surely you would look at this and conclude that it’s a mistake.”

    1. It was a telephone poll. Unreliable especially here where people may not give honest answers to a stranger over the phone. Excludes those more likely to have a mobile as their primary phone (the young and the unemployed).
    2. Apparently the ‘random’ sample was ‘weighted’ to be representative of the NI population. No detail of the weighting process is given. Let’s hope they didn’t use the outdated 2001 census data.
    3. It was conducted by Alliance man Gerry Lynch.

  • Drumlins Rock

    I note the poll has Prods+RCs at 60% whereas the last census has the total at 80%, I know secularisation has grown, but not by that much I imagine, considering that sector contributes a considerable proportion of the Alliance vote there is certainly a degree of distortion there.

  • quality

    Just so I understand…?

    So if the poll shows no support for a united Ireland, its valid. If it shows the growth of the Alliance Party, Sinn Féin and the DUP, its biased. And visa versa.

  • Drumlins Rock

    The margin of error would increase considerably, my advice Quality is to look at the raw data, http://www.lucidtalk.co.uk/Portals/d5eacc2c-c1ff-48d6-8f89-c77800a057c1/120530BelfTelResTbles-14Qs.pdf

    See what you think.

  • http://nwhyte.livejournal.com Nicholas Whyte

    Next European elections are 2014 and 2019, not 2013 and 2017.

    My gut feeling is that polls have always overstated Alliance support – I remember the one in 1996 that reported support for the party at 13%, taken the weekend after an election result of 6.5%. I imagine that there is less of this happening now, and the methodology of this poll seems to be an improvement, but I’ll still take the Alliance result with a pinch of salt.

    By the same toke, SF support has been historically understated in polls. So I take the reported increase in their support here very seriously indeed.

  • http://nalil.blogspot.com Nevin

    “There are no votes “gained” or “lost”.”

    Agreed, fjh. There were 535901 votes unclaimed – not that far short of half the electorate – and we don’t tell folks who we voted for in elections.

    “mjh, who are you?”

    It’s a bit ironic, Drumlins Rock, that we have an anonymous person commenting on a poll by lucidtalk – and being challenged by a second anonymous person!

  • andnowwhat

    I’d a guy to a poll with me last month, something to do with services used and pretty inoccuous. There were a couple of questions about Irish and Ulster-scots culture. I had no idea who he was and so I lied about my enjoyment of Irish culture and the fact that I don’t believe there is any such thing as the Ulster-scots language.

    Were it about politics, I’d have lied through my teeth.

  • Drumlins Rock

    Nev, if anyone was really interested in who either you and I were. less than 5 minutes on Google would bring up our names using the info we provide in our posts and comments, I prefer to think of it as a pen-name rather than anonymous. I have no idea if mjh is worknig for BT, Lucidtalk, the Alliance Party, MI5, MFI, or Mossad.

  • http://fitzjameshorselooksattheworld.wordpress.com/ fitzjameshorse1745

    For those of us who remember Hughie Green on Opportunity Knocks…..the lesson is that the clapometer is just for fun. Its your vote that counts.
    I dont think that posting without a profile is a big crime. I did for months until I was outed as being Brian Feeney. The first person to “out” me …..well it would be unfair to name him but he had Alliance Party affiliations.
    And opinion pollsters…..no less than teachers, doctors, journalists, taxi drivers..have political affiliations…..which dont affect their professionalism.
    But as always ….its nice to know who is claiming what.

  • http://nwhyte.livejournal.com Nicholas Whyte

    FJH / DR,

    I have always posted here under my real name. But I don’t have a problem with people posting here under pseudonyms, unless they hide behind them to make personal attacks on people who appear here under their real names, or indeed to attack others for using pseudonyms!

  • Drumlins Rock

    Nic, I think most comentators on here behave pretty well, using a pen name or their own name makes no odds, I have a tendancy to play devils advocate too much hence a pseudonym!

    Whats your take on the number crunching above? you are the top expert on this in my view!

  • http://fitzjameshorselooksattheworld.wordpress.com/ fitzjameshorse1745

    Indeed and not something I would ever do.
    The moderation here would never allow that.
    I indeed have no trouble with people using pseudonyms or indeed real names.
    The real point is knowing the connexions of people and organisations.I always think thats relevant. Why do people say what they say?
    Drumlins Rock makes no secret of his affiliation to the UUP.
    I make no secret of my affiliation to the SDLP.
    Clearly a matter for others whether or not they declare an interest. In most threads its irrelevant. Unfortunately sometimes people have to be reminded of their “interest”.

  • http://sluggerotoole.com Belfast Gonzo

    keano said:

    If this poll were to be replicated, those parties [SF and DUP] would be attracting a combined total of 58.7% of the voting electorate. Thats an enormous mandate and the fact it appears to be ever-increasing, shows beyond doubt that people here are happy with how those parties are governing.

    Wrong.

    Only 8.4% rated the Assembly’s performance positively compared to Direct Rule. That is utterly, utterly pathetic.

    People clearly think the governing parties are pretty useless, but put up with it because the alternative is too awful to contemplate.

    This survey is many things, but a ringing endorsement of SF/DUP’s ability to govern it ain’t.

  • http://nwhyte.livejournal.com Nicholas Whyte

    DR,

    My take on the numbers is that for every single party the reported change in support from the last election is within the 3.1% margin of error on the poll – except Alliance and the Greens, but see my earlier comment re polls overstating Alliance support (and the same applies to the Greens, plus the fact that they were not represented in all constituencies in the election).

    I find the increase in support for Sinn Fein the most interesting figure, again for the reasons given earlier. Other things being equal (which they may not be) that would hit the SDLP hardest.

  • Drumlins Rock

    Nic, I was just noticing that when it was transfered to the Euro vote, SF have been just on the quota, this has them with a good bit to spare, there was a long shot the SDLP could come back there, but this seems extremely unlikely gonig by that poll, a good candidate could change things, not sure if Martina will have any impact.

  • michael-mcivor

    The polls have Sinn Fein support going up across the 32 counties-

    So in goverment or in opposition it is Sinn Fein who are seen to be doing the work-

  • mjh

    Drumlin Rocks

    mjh, who are you? your profile is blank and you have never commented on Slugger before, like to know the angle of the person posting first before commenting.

    I have no idea if mjh is worknig for BT, Lucidtalk, the Alliance Party, MI5, MFI, or Mossad.

    Not working for anyone. I’m retired. I have not lived in Northern Ireland for over 30 years. I have not contributed before because I have never before felt that I had anything particularly interesting to say. I follow Slugger because I often find interesting what you and others have to say.

    When I lived in Northern Ireland I was an active Alliance supporter, but I have not spoken to anyone in Alliance since I bumped into the late Basil Glass at Belfast Airport at least 20 years ago. I hope that is enough information for you and others to feel that I am not disrespecting you with my anonymity. But now that you know my political view please treat the analysis on it’s merits. I was conscious when doing it that I needed to avoid my heart ruling my head.

    I enjoy digging around in data. The only reason that I have commented on this occasion is that it seemed to me that there were things in this poll that had not been surfaced and so I put it forward.

    A number of commentators have dismissed this poll on the grounds that they do not take opinion polls seriously. That is fair enough. But do not think that political parties share that view. Politicians are very interested in polls, otherwise parties would not spend serious money on commissioning their own private polls. And senior members of all parties will be using this poll to inform their thinking. At least we can all know what they know, or think they know if you prefer.

    Ulick

    It was a telephone poll. Unreliable especially here where people may not give honest answers to a stranger over the phone.

    Some major polling organisations use telephone polls, and even the internet. Many polls are no longer conducted face to face. Anyone who is interested in the issues of sampling and weighting can find more information on Anthony Well’s site http://ukpollingreport.co.uk/ in the ‘Articles and FAQ’s’ section.

    Actually one of the most noticeable and pleasing points about this poll is that those questioned had no problem expressing their support for the DUP and SF, as noted by Keano. Many past opinion polls certainly did overestimate Alliance support, but they also underestimated DUP and Sinn Fein support often by huge margins. People were unwilling to admit to a stranger that they supported those parties. It is the same issue as the “shy Tories” who wouldn’t admit their opinion for most of the last 20 years. Who knows, maybe it is partly because this poll was conducted by telephone that people have not felt shy to give their true opinion. Maybe it is partly the work and efforts of those parties. Maybe it is a reflection of a more secure and comfortable society.

    Thanks for the dates correction, Nicholas Whyte.

  • Drumlins Rock

    mjh, thanks for the info, it is appreciated, as is the effort to put the post together, you seem to have just enough bias to make it credible! However I think as NW states the changes are within the margin of error and some questions arising it is in reality a waste of effort!

    I think we would need a series of similar polls from the same group to get a truer picture.

  • http://www.wordpress.ianjamesparsley.com IJP

    On the party preference, I would have to say that the raw data are too limited. Just over 1,000 people, over half going for Other/Wouldn’t vote, isn’t really a sufficient number on which to base very much.

    The figures given for each party exclude the “Other/Wouldn’t vote” collectively – “Other” would usually account for 3-5%, including some fairly significant Independents. On those grounds alone, you have to reduce each party’s share (albeit by a roughly equal amount for each).

    For Alliance, also, the figures from the outer postcode districts seem far too high, coming in often at around 4% of those voting. There isn’t a box in the Border Area, Tyrone, Fermanagh or Derry City where the party is on that, and it is rarely anywhere near it. That said, it may indicate a willingness to vote Alliance once it produces well known local candidates and gets to the stage where it is seriously competitive (e.g. Patrick Clarke’s win in Newcastle at Council level last year).

    I would also begin to wonder genuinely if there isn’t a “Shy UUP” factor coming in. It is such a chaotic party, you just wonder how many people would wish to admit to voting for it. That said, I found no evidence of this conducting my own exit poll last year.

  • Charlie Sheens PR guru

    Usual health warnings applying,

    wouldn’t this point to a possible SDLP ( or alliance gain at euro level?)

    if the SDLP is hanging around longer than alliance and gets approximately the same level of transfers from alliance as the UUP get, wouldn’t an UUP elimination lead to lots of alliance transfers from UUP staying in the center ground rather than to a 2nd DUP runner? Same goes for greens who are usually SDLP transfer friendly.

    Nicholas, have you an inkling for the meandering nature of alliance transfers?

  • FuturePhysicist

    I have to agree with Ian and Nicolas here.

    Ian on the point that the sample size is ridiculously small to draw conclusions and Nicolas on the point of overstating Alliance support.

    If we took this with a pinch of salt we’d be having hypernatraemia.

  • RyanAdams

    This poll is useless for STV, especially where the last seats are often called by transfers.

    For example last time, Alliance transfers saved the UUP’s arse last time in South Belfast. Obviously if Alliance had two candidates they would be utilising those transfers themselves – therefore on top of the ‘UUP decline and DUP growth aforementioned, DUP also gain UUP seat there?

  • http://www.lucidtalk.co.uk LucidTalk

    On behalf of LucidTalk we would like to make the following points in response to some of the points already stated:
    – Gerry Lynch is a Consultant to LucidTalk and provides advice, guidance, commentary and interpretation on the poll results. He was not involved in the conducting of the actual poll. Gerry is no longer a member of the Alliance party, as No officers, employees, or associates of LucidTalk (e.g. polling staff etc.) are members of any political party, as this is prohibited.
    – There was a small sample of people polled via their Mobile phone – their area of residence was determined by a follow-on poll question.
    – The sample size was more than adequate for the size of the Northern Ireland electorate – see National UK Polls e.g. Comres, YouGov etc.., and you will see similar sample sizes used to project/predict political opinion in GB, which has of course, a much larger electorate! NB You don’t increase results accuracy that much, by continually increasing sample size – see mathematical sampling theory for a fuller description of this concept.
    – Regarding Ryan Adams points re. STV – It is not true that the poll is ‘useless for STV’. Asking someone what their first preference vote would be at an Assembly election is key. Look at last Assembly election results, and also the 2007 & 2003 results (see Nicolas Whyte’s ARK site), and you will see in the vast majority of constituencies that the first six candidates on the first count (i.e. first preferences) ended up being the 6 elected! In the few exceptions to this, it was the person who came 7th on the first count who was elected instead of the 6th or 5th ranking person, i.e. very, very rarely did anyone who came 8th, 9th, or beyond, on a first count under STV ended up getting an MLA seat! As you can see, first preferences ARE KEY and this is what our poll researched!! A rule that backs this up is that unless you get in the top six on the first count, and/or get half the quota on the first count, in a Northern Ireland assembly election, the chances of getting an MLA seat are less than 3% – again review Nicolas Whyte’s site and see this ‘rule’ in operation for yourselves!
    – It is planned to hold a similar Northern Ireland wide poll every six months, dependent on commercial support and funding (as you would expect). This will provide the most useful part of regular polling in that this should show emerging trends, and improvement/dis-improvement in e.g. political party support. It is these trends and changes in opinion, that polls can show on an ongoing basis.

    LucidTalk plan to run ‘live’ roadshow events in September which will cover and present the poll results in more detail, and will also announce plans for the planned next poll in November 2012.

    For more details – Follow LucidTalk on http://www.lucidtalk.co.uk, and also on Facebook and Twitter.

  • Drumlins Rock

    Charlie, try Nicolas excellent website, http://www.ark.ac.uk/elections/fe09.htm give you all the last euro elections details, the relevant quote would be,

    “The vote from the centre parties split fairly evenly, 37% to the remaining Nationalist candidate, 42% to the three Unionists (and 17% to nobody).”

    The Chances are this time the TUV vote will be below the AP one, it went 2;1 to UUP on transfers the last time, which going by the percentages here would push the UUP above the AP, who would then be eliminated.

  • Charlie Sheens PR guru

    DR,

    Thanks, but I’ve read that already. While it’s a useful illustration I was more interested in the post UUP elimination ( even though I don’t think that is likely at all). Likewise if the shoe was on the other foot, would votes eliminated SDLP votes originating from alliance stay close to the center ground. I didn’t see too many examples of this as SDLP tends to get the runners up spot a lot.

    Interesting to see Anna Lo’s surplus ( with all parties still in the race) went 40% SDLP, 21% UUP, 20% Green ( which ended up with NcDevitt) 8% DUP, 5% SF, 5% lefty types.

    Since some ppl here call alliance a small u unionist party the breakdown is curious.

  • Drumlins Rock

    Charlie, playing the what if game of DUP & SF elected with votes redistributed and Greens & TUV eliminated, leaving SDLP & AP on the same vote and UUP one behind and therefore eliminated next, I think up to half would not transfer, of the rest it would be 4 to 1 at least to Alliance. Virtually the same vice versa.

  • Reader

    LucidTalk: - The sample size was more than adequate for the size of the Northern Ireland electorate – see National UK Polls e.g. Comres, YouGov etc.., and you will see similar sample sizes
    In general, you are right, as a sample of 1000 gives a margin of error of about 3%, and it’s difficult to improve on this without a load of extra expense. And the population size is irrelevant, as you know, but isn’t widely accepted.
    But the 3% margin of error overall isn’t really good enough to measure small party support against a quota within a six member constituency.

  • http://www.lucidtalk.co.uk LucidTalk

    Reader: You are correct. However LucidTalk did not take the poll results and extrapolate them into election predictions based on individual constituencies – It has been the contributors to this thread that have carried out these projections! It is of course natural for poll results to be analysed and interpreted in this way, but LucidTalk don’t publically do these sorts of analyses, because, as you say, the error factor increases as you do these sorts of ‘drill-down’ calculations and projections.

    As we say above, hopefully with our plan to carry out these sorts of poll projects every six months, this will allow us to track trends in increasing/decreasing support for political parties and various issues.

  • Drumlins Rock

    Lucid, while I may have some criticisms! I do welcome the fact we now have polling, I guess we won’t find out its true accuracy till at least 2015, but in the meantime as you said it is the trends that will be interesting.

    One area I mentioned earlier, regarding religious balance your “others” category recorded 40%, wheras the last census had it at only half that, surely it hasn’t changed that much in 10 years? And regard the number of MLAs question, tut tut, that was just sloppy and a leading question, I’m shocked Bill would let that pass!

  • http://www.wordpress.ianjamesparsley.com IJP

    Lucid

    First off, congratulations on establishing the company and arranging the poll. Also congratulations on getting Gerry on board.

    Just one thing, I’m not sure I can allow your argument that the sample size was sufficient. As Reader implies, the difficulty with a 5-7 party system (versus the 2-4 party system in England) is that even a slight margin can be magnified more significantly. My point was that you have given data for postcode areas and, as I suggest, it is too small to be significant (and, as with the case of Alliance in border areas, it simply doesn’t ring true – and believe me I’d love to think they were right!). That is not a criticism of the polling, but rather a big caution against taking the results too literally and particularly against taking the breakdown too literally.

    Since you’re here, an obvious question: what next? My own view is the best way to get a clear read out of the situation across NI would be to run such polls consistently, at least monthly. Is that an option? If it isn’t, what can Sluggerettes and others do to assist make it so?

  • Drumlins Rock

    IJP, monthly might be a bit much! lol

    Was just wondering, if there was minor flaws in the methodology and they are corrected, does that distort the picture as it develops over time?

    PS, can I email my phone number so that I’m included in the next survey?

  • http://nwhyte.livejournal.com Nicholas Whyte

    Replying to the specific question about the Euro elections:

    Alliance’s vote will have to increase a lot more than this poll suggests for the party to be in the running for a Euro seat. Taking MJH’s figures, the UUP (or in any case the second Unionist candidate) would benefit from the DUP surplus and go from 0.44 quotas on first preferences to something closer to 0.69 with transfers. Likewise the SDLP would go from 0.51 to 0.63 with SF surplus votes. That does leave Alliance transfers effectively deciding the outcome; but as has been discussed they tend to split too evenly to make up the difference for the SDLP.

    The other scenario under MJH’s figures is that the DUP balance two candidates ahead of the UUP, who might be eliminated with Alliance still in the race. In that case I am confident that the second DUP candidate would get in on transfers from the eliminated UUP candidate. Charlie Sheen’s PR Guru’s suggestion that UUP votes might “stay in the centre ground” is flawed, because it’s not at all clear that the UUP’s votes are based in the centre ground. Back in the early days of the Good Friday Agreement, when the DUP were clearly wreckers, you would find UUP transfers sometimes going to all kinds of other people against them. Now, it is the DUP who have delivered a second term of power-sharing government without interruptions and the UUP who are the ones yelling about homosexuality. Which is the more moderate Unionist party today?

  • Bangordub

    Now, it is the DUP who have delivered a second term of power-sharing government without interruptions and the UUP who are the ones yelling about homosexuality. Which is the more moderate Unionist party today?
    Nicholas, Not your usual area of expertise but brilliantly put and a subject for a whole new post

  • Bangordub

    “Now, it is the DUP who have delivered a second term of power-sharing government without interruptions and the UUP who are the ones yelling about homosexuality. Which is the more moderate Unionist party today?”
    Nicholas, Not your usual area of expertise but brilliantly put and a subject for a whole new post
    Sorry, forgot the comma’s

  • http://www.lucidtalk.co.uk LucidTalk

    Some comments re. recent points:

    The sample size is/was more than required. It would make very little difference to the accuracy of the poll to have a larger sample size. NB You’re not meant to extrapolate (or shouldn’t), specific predictions from sub-sets of data in the poll e.g. the postcode/area data. IJP – you can’t project Alliance performance in the West of Northern Ireland, just from sub-sets of data.

    Re. the No. Of MLA’s question, it is not a leading question, as its quite proper within polling to set the context and ‘set-the-scene’ in terms of No’s of members in other UK assemblies etc.

    Re. the point regarding frequency of polls, as I said above, we are planning to run this type of poll twice per year. In terms of frequency, there is fact about polling and frequency of polls – Money/Funding! Professional Opinion Polling is expensive, and we rely on a commercial model to carry out polls such as the recent one in the Belfast Telegraph.

  • Charlie Sheens PR guru

    Nicholas,

    Again I think I’ve not been understood correctly, so I’ll try once more.

    You correctly identify that most alliance votes split to evenly to make a difference on their own. The point is THIS:

    If these numbers are, for the time being, taken at face value,

    UUP -11% ,Alliance 10.9% , SDLP 12.5%

    Then in a situation were the DUP has two candidates on 15.3% each, the alliance will be eliminated and lets say it goes : SDLP 4% UUP 4%, DUP 0.9%, dead 2%

    Gives 16.5% SDLP
    15.0% UUP
    15.8% DUP1
    15.8% DUP2

    Following SF surplus: 2.5% SDLP, 0.5% dead gives:

    18.0% SDLP
    15.0% UUP
    15.8% DUP1&2

    Next to go is UUP:

    Unlike Nicholas’ Whyte’s interpretation of my analysis as studying UUP transfers, I’m MORE interested in the 4% originating from the alliance party.

    I would suppose that ~2% of that 4% would go to the SDLP. But I’d like to debate it more. If it’s given that the UUP electorate themselves would 2nd (or 3rd preference after alliance) the SDLP a further 1% of their initial 11%, we can add another 3% to the SDLP pile and 10% of the remaining 12% to the DUP giving something like:

    SDLP: 21%
    DUP1: 20.8%
    DUP2: 20.8%

    Again, it’s taking these numbers at face value, did not include TUV and GREEN (or PBP) and we are still someway out from an election. So all I want to get at is:

    IF the eliminations happen something like this, what happens to the alliance transfers as they can make a big difference.

  • Drumlins Rock

    Lucid, pull the other one, that was a leading question, and an inccuarate one at that, Sir Humprey would have been proud of it.

  • mjh

    Charlie Sheens PR Guru

    If I may comment on your last posted question to Nicholas Whyte.

    I had written a 3 page answer complete with figures and rationale. But since even I was getting bored reading it forgive me for cutting to the conclusions. If you want all the gory details you can have them.

    You describe a European Election in which the SF candidate is elected, and the Alliance eliminated. Some of the Alliance transfers go to the UUP, which you have called at 4% of the total votes cast. When the UUP candidate is eliminated you want to know how those UUP votes which had originated with Alliance will transfer between the SDLP and the 2 remaining DUP candidates.

    First the health warnings. Once votes have transferred from Party X to Party Y there is no published information on what happens to them when Party Y’s votes are transferred. Secondly the evidence which could point to an answer to your question is only visible in 7 of the 18 constituencies, and even then is sometimes based on small numbers.

    Still, making the best of what we have it looks to me like at least a third of them will not transfer at all, and this figure could be as high as two thirds. Best guess about half =2%.

    Of the remaining 2% at the very least a third goes to the DUP candidates, but (and this last part is personal judgement) this is likely to be much higher.

    And as you say your scenario has not allowed for TUV and Green transfers. So even on your assumptions I reckon the SDLP would probably have to have more first preference votes to be in with a shout.

  • http://nwhyte.livejournal.com Nicholas Whyte

    CSPRG,

    Not wanting to go on about it, but I agree with MJH. Your scenario has Unionists on a combined total of only 46.6% *after* SF surplus transferred *and* Alliance elimination. That’s unrealistic enough to make the question fairly pointless, I’m afraid.

  • PaddyReilly

    Your scenario has Unionists on a combined total of only 46.6% *after* SF surplus transferred *and* Alliance elimination.

    Probably not far wrong because, if you examine mjh’s data, you will find that he is only dealing with (1.23 + 0.44 + 0.44 + 0.50 + 1.12 =) 3.73 quotas, whereas European Elections traditionally involve a return of 4 quotas, bar a couple of votes.

    So let us recapitulate: we have an opinion poll, only slightly more reliable than a dream, which has been partially quoted and then used to build a whole aerial castle of speculations.

    Lionel Hutz asks how it is possible that the Alliance Party can gain 41%. The answer is that in the context of multiple-choice voting, this gain is only apparent. People merely place Alliance in first position and then assign their traditional choice to second position. Alliance is eliminated and the traditional choice is elected.

  • http://www.lucidtalk.co.uk LucidTalk

    Drumlins Rock: Without getting into an argument about wording of questions – LucidTalk will be running roadshow events in September as part of the build-up to the next scheduled poll in November. This will be an opportunity for people like you – DR, to have a direct input into the wording of poll questions, as well as other aspects of the poll, including polling methodology.

    PaddyReilly: The poll is, and was not, ‘a dream’ as you put it. However, as I said before, the key aspect and benefit of polling is to carry out polling on a regular basis, and spot the trends – a one-off poll is not as useful as comparing poll results over a 1 to 2 to 3 year period, and analysing the trends. This ‘first’ poll is now a baseline and the important and useful data will be to see how future polls (NB carried out using the same methodologies) compare with this baseline, and with the other polls moving forward – this will show key increases and/or decreases in political party support over time. E.G. As far as the Alliance party is concerned, it will be far more useful to see if the Alliance party can hold (or even increase) the poll rating they achieved in this recent poll, over the next 6 to 12 months – if they do this, then we then know for certain that this increase in support is genuine etc.!

    As I said before this all this depends on our commercial model being successful so that we raise the necessary funding!!