Belfast Telegraph

European election predictions: Full analysis of how DUP, Sinn Fein, UUP, SDLP and others could perform in May

The Belfast Telegraph's polling partners, LucidTalk, have analysed how Northern Ireland's European election candidates could fare on May 22

The European Parliament buildings in Strasbourg. Botond Horvath /
The European Parliament buildings in Strasbourg. Botond Horvath /

Just in case you don’t know, or haven’t yet noticed the posters, Northern Ireland will be holding European parliamentary and local government council elections on May 22.

Click here for our full European election table

This review covers the European election when Northern Ireland will be treated as one large constituency, with 1.1 million electors, and will elect three members of the European parliament under the PR-STV voting system i.e. expressing your vote in terms of preference 1,2,3...

It’s always dangerous, indeed could be suicidal in terms of reputation, for a pollster to take several different polls, other data, and various weighting assumptions, and then collate election projections. But I’m a brave fellow, or at least foolhardy, it’s also fun, and in any case it’s a great excuse for an argument. So I’m going to have a go.

First of all, it’s important to note that the Euro elections are different to Assembly and Council elections. One key difference is that first preference votes are more important. It’s obvious that first preference votes are very important in any PR election, but they are particularly important at the Euro election. As you move towards the assembly, and particularly the council elections, the vote transfers from other candidates take an increasingly important role in terms of who gets elected. This is shown by the fact that in all (yes all) NI European elections since they started 35 years ago in 1979, the first 3 candidates on the first count, were the 3 candidates that eventually got elected. They may as well just have had the first count, declared the top 3 elected, and then went home. However, in terms of our predictions we are going against this trend this time, and are predicting that this forthcoming election will be the first NI Euro election that will break this ‘rule’ – see below. Part of the reason for this is that we have a larger field of candidates, compared with previous Euro elections – 10. As such transfers, and also the level of non-transferable votes at each stage of the count, start to become more important. This could be crucial on the unionist side of the contest, as it should be noted that of the 10 candidates, only two are Nationalist/Republican, with the other eight either being pro-union or neutral. Our predictions are in % terms, as it’s impossible to predict turnout, although all the LucidTalk 2012-2013 polls recorded large despondency with politics in general, and this was again confirmed by our recent opinion panel survey (see below). This is reflected in increasingly low turnouts at elections, and as such we predict a turnout of perhaps 50-55% at the forthcoming May 22nd elections, with the turnout being boosted because of running both the European and Council elections on the same day. 

Prediction Methodology

In terms of our approach and methodology, the main input data to the projections are:

LucidTalk Northern Ireland (NI) -Wide polls from May 2012, November 2012, and September 2013, which were related to political party preference at NI Assembly elections. These individual poll results have been included in the modelling, along with the trends. The trends show how political party support has been growing, declining, or remaining the same over the 2012 to 2013 period.  

Interviews with the 220 member LucidTalk Opinion Panel (balanced to reflect NI society), carried out in the last two weeks.

Previous election results including NI Assembly, Westminster, and the previous Euro election in 2009, including vote transfer patterns.

Various weighting factors have been applied to all the polling data and research to reflect the structure of the forthcoming European election e.g.

A. Party machine strength. The NI Euro election suits big parties with large No’s of canvassers, and large No’s of council candidates running in the local government elections which take place on the same election day. When compared to the smaller political parties, this particularly suits the ‘big-two’ of the DUP and Sinn Fein, and also to a lesser extent the SDLP and UUP. 

B. Name-Profile of each candidate e.g. Jim Allister running for the TUV is different from the TUV running a different candidate. The same applies to all of the other party candidates contesting the election, who all have various levels of ‘brand awareness’ in terms of name recognition. As such, weighting factors have to be applied for well known names, and powerful party brands.

C. Party Over-performance/Under-performance: Certain political parties at European elections perform better or worse than their performance at NI Assembly elections, due to the different structure of the Euro election. As an example, the UUP usually performs better at European elections compared to Assembly elections because there are less unionist candidates running. Alliance performs less well at the Euro elections because they are localised around Belfast and are disadvantaged in terms of the one large NI-wide constituency, in which the Euro election takes place.

All of the above date, research, and weighting factors were included within the LucidTalk computer models which then forecast the possible poll results for 1st preferences in percentages for each candidate.

Cutting straight to the chase, we are predicting the same three parties returned again i.e. Anderson (Sinn Fein), Dodds (DUP), and Nicholson (UUP). However, even if this prediction turns out to be correct, Unionism may have to have a long hard look at itself after these elections are over, because as you will see from the analysis, the results at the last count stage could be very tight!

Any additional assumptions and weighting factors are included within the individual candidate analysis below.

Candidate Profiles and Analysis:

Jim Allister (TUV)

A ‘big name’ who’s top of the ballot paper, which is always an advantage. There is a big difference between the TUV running a candidate for the Euro elections, and that candidate being Jim Allister, the TUV leader. Allister polled 13.7% of the vote at the last Euro election in 2009, but then he was a sitting European parliament member. The TUV vote went down in the 2011 Assembly election to 2.5%, and it scored at a similar level in our 2012 and 2013 LucidTalk polls, which related to NI Assembly elections. However, in Assembly elections, Jim Allister is restricted to running in North Antrim, and in the forthcoming Euro election he once again has the whole of NI to target. He polled well in our recent opinion panel survey research, not particularly because people agree with his policies, but there seems to be a sneaky respect for Jim Allister and his sort of one-man opposition crusade in Stormont. Many, not all, of these people are 1st preference voters for other parties, but they may vote-transfer to the TUV leader, which could become important in later stages of the count. However, our systems don’t predict he will do as well as he did in 2009 mainly because he’s not a sitting European MP for this election. Taking the various weighting factors into account, and in particular the ‘big name’ weighting factor, plus our poll data research, our system predicted a score of 7.8% for the Euro election.

Martina Anderson (Sinn Fein)

Probably the easiest party and candidate to predict in terms of the result in the Euro election. The reasons for this prediction are several and obvious. Sinn Fein have the most powerful political machine (canvassers, party workers etc.), on the island of Ireland, probably only rivalled by Fine Gael. Political machines are important, particularly in NI-Wide Euro elections, because of the huge NI constituency, and the importance of ‘getting out the vote’ on Election Day. This favours big parties, and big political machines like Sinn Fein and the DUP. Sinn Fein has consistently polled 26-29% in our 2012-2013 NI Assembly election polling, and equally strongly in our recent Opinion Panel survey. Yes, there has been a slight downward trend over the last two years, but our models are still predicting Anderson will get elected on the first count, with a % vote around 26-27%, and a small surplus. Unlike the last Euro election, we expect that this time this surplus to be transferred as the second stage of the count process – see Table. This is because, unlike in 2009, there is a larger field of candidates, so it’s likely that there will be candidates at the bottom of the first count score sheet whose first preference votes are close, hence leading to Anderson’s, probably small surplus (at least according to our predictions), being transferred first. Our computer prediction model predicts a First count score: 26.2%

Alex Attwood (SDLP)

The SDLP, like the UUP, always over perform at the Euro election compared with their Assembly election performance. Our 2012 and 2013 LucidTalk polls have shown a healthy upward trend for the party, and indeed this was reflected at the 2013 Mid-Ulster by-election (the most recent election in NI). They also performed reasonably well in our Opinion Panel survey, held over the past 2 weeks. Although they’ve taken an electoral battering over the past 10-15 years they still have a reasonable party machine, admittedly not as powerful as the big-two Sinn Fein and the DUP, but certainly still better than the UUP’s machine. As detailed before, the party machine is important in the Euro’s, and like the other big parties they are running a healthy slate of council candidates so they will have a large No. of party workers canvassing, and knocking doors. Attwood is third on the ballot-paper, directly following Anderson (Sinn Fein) and this will be a help in getting transfers from SF, which could become crucial in later stages of the count. With the various 2012-2013 poll results, and our opinion-panel survey research, trends, and various weighting factors, our system is predicting an SDLP Euro poll result of 16.7%.

Mark Brotherston (Conservative)

The Conservatives have valiantly tried to bring what they term national politics to Northern Ireland, and they’ve had several attempts at this. One of these attempts was their ill-fated link-up with UUP with the UCUMP project. It’s extremely hard to predict performance for this election as all LucidTalk polling only covers political parties that are currently in the NI Assembly, and of course the NI Conservatives have no assembly members. We did include them for our recent Opinion panel survey, and although they scored some support, it was very small. They suffer from fact that they don’t have an electoral machine, which as stated before, is a big disadvantage in a NI Euro election which suit the big-boys like the DUP, Sinn Fein and also yes the UUP and SDLP. As such, our system predicted a euro poll result of 0.8%.

Ross Brown (Green Party)

The Green party scored a very respectable 3.3% at the last Euro election, with their candidate Stephen Agnew performing very well during the campaign. No doubt this assisted him in securing the Green Party’s one Assembly seat in North down in 2011. The party have been around in NI elections for a long-time, are well funded, but again we would ask questions about their electoral machine and the ability to cover the whole of NI. However they performed adequately in our Opinion Panel survey. As such, our system predicted a 1st preference vote score of 2.3%.

Diane Dodds (DUP)

A sitting European Parliament member, and ended up being the sole DUP candidate after all the posturing from Peter Robinson about the possibility of running two candidates. Like Sinn Fein, our 2012-2013 NI Assembly election polls, showed strong solid support for the DUP at a regular 27-30% score. Our recent Opinion Panel survey showed this still strong support, but with a small downward trend. The big question is how Jim Allister will affect the DUP performance in this election, as our poll research shows that Allister get’s 80% of his support from former DUP voters. NB This used to be 90-100%, but our research also shows that Allister is now also attracting increasing support from other parts of the Unionist family. The big factor in Dodd’s favour is the formidable DUP election machine, backed up by their sizable field of council candidates, canvassing, and knocking doors. Again, with the various 2012-2013 poll results, and our opinion-panel survey research, plus trends, and various weighting factors, our system predicts a Euro poll 1st preference result of 20.8%. 

Anna Lo (Alliance)

A well known name from a long standing and established party. How her recent comments regarding her views about a united Ireland will affect her performance at this election, remains to be seen. As mentioned in our weighting models, Alliance always score 1-2% less at European elections compared with Assembly elections because of factors mentioned before. There is also an additional reason in that Alliance are localised around Belfast, and are very weak in the rest of Northern Ireland. This doesn’t suit running in an NI-Wide election, with its huge 1.1 million electorate! Our polling and the opinion panel survey results, show that the flags issue hasn’t had much affect one way or the other. Again, with the various 2012-2013 NI Assembly poll results, and our opinion-panel survey research, trends, and various weighting factors, our system predicts a Euro poll prediction of 5.9%. 

Tina McKenzie (NI21)

NI21 scored over 5% at our last NI Assembly poll in 2012, but how much of this was due to a novelty factor of being the new kid-on-the-block, and being well covered in the media at that time? This is the only full NI-Wide poll score we can take, as NI21 didn’t exist for our 2012 polls, and this last poll result was eight months ago. They scored less than this 5% in our recent opinion panel survey, which showed that maybe the initial heady excitement of being a new party has worn off to some extent. A lot will depend in how their unknown candidate performs during the campaign, and how much media time she can get to raise that crucial awareness factor across NI, so important in elections like the Euro’s. For a start-up party, NI21 are running a sizable slate of 45-50 council candidates, who again should provide some semblance of a party machine in terms of canvassing votes. They’re going for the advertising strategy of large billboards as opposed to the lamp-post posters, and I’ve received favourable reviews of McKenzie’s billboard posters. NI21 are also trying to appeal to that vast and growing reservoir of non-voters, but the old heads in politics know that this is a difficult thing to do. If turning non-voters into voters was easy, then the answer would have been found a long, long time ago. How all of this affects their poll-score in the coming election remains to be seen. This will be NI21’s first election, and it’s incredibly difficult to predict performances of new parties – indeed when the Alliance party faced its first election in 1973 they scored over 13% of the poll, and no-one saw that coming, but it’s interesting to note that, the Alliance party have never achieved anywhere near that poll score since! However, when all this ‘little’ data and research were entered into our system, it produced a prediction of 3.2%. I would say that if this poll score is achieved, or anywhere near this, then it will be a very credible performance for a new start-up party.

Jim Nicholson (UUP)

A long standing European Parliament member, from the once mighty UUP, Nicholson is the old experienced campaigner in the field. Our 2012-2013 Assembly polling result showed the UUP had dropped again from their 2011 Assembly election performance with their poll score coming in at around 11.5%, but with a small growth trend. Our Opinion panel survey also showed a small upward trend, perhaps suggesting that the UUP had reached their electoral floor, and that their core vote was pretty solid. Put it this way, taking in to account all the banana-skin mistakes the UUP have made over the past 15 years, if the people who are currently supporting the UUP are still voting for them, then they are a pretty stubborn loyal bunch, and are unlikely to change their party allegiance now! The UUP also benefit from the weighting factor, mentioned before, in that they always perform 1-2% better at the European elections because of the smaller No. Of Unionist type candidates in the field e.g. Nicholson may pick-up a chunk of PUP votes, because the PUP isn’t running a Euro candidate. However Nicholson is not on a particularly good position on the ballot paper being one form the bottom, i.e. you have to go and ‘look for’ Nicholson’s name. Some people laugh at these sorts of silly points, but the research shows that they count at elections, at least to some extent. This is reversed with Allister, who has an advantage in being top of the ballot paper. However again, with the various 2012-2013 poll results, and our opinion-panel survey research, trends, and various weighting factors, our system predicts a Euro poll result of 14.2%. However, this is still down from the 17.1% poll score for the UUP in the 2009 Euro election. 

Henry Reilly (UKIP)

Apart from the Conservatives the only candidate who can really say he’s standing for a UK national party. However, outside South Down, Reilly’s home constituency, and Strangford which has NI’s only UKIP MLA in David McNarry, Reilly will suffer from the name awareness problem. However, he will be able to play on his national leader’s prominence. I’ve tried this –When you ask people about Henry Reilly, they say who? When you mention Nigel Farage, they say Oh yes, I know him! This is no criticism of Reilly, and indeed if he uses the undoubted popularity of his national party leader, then he could pick-up votes. UKIP, were included in the LucidTalk 2013 Assembly polls as they had, by then, a local MLA in David McNarry, but were not in our 2012 polling. As such, like NI21, the 2013 poll result is the only full NI-Wide poll results data we have. They did perform average-well in our recent Opinion Panel survey, mostly seemingly due to Farage’s large national UK profile. As such, this weighting, and one poll result, have been included in our prediction system, which has forecast a UKIP Euro poll score of 2.1%.    

Projection and election forecast


The 1st preference votes detailed in the table are estimates taken from our computer prediction system. The system has an estimated error in the figures shown of +/-1.1%. We are quoting the median figures in the 1st Preference vote column, as this assists calculation of the transfers and eliminations as the count process progresses.  

10% to 15% of the transfer at each count stage, is assumed to be non-transferable. E.g. Martina Anderson (SF) has a 1.2% surplus, but only 1.0% of this is transferred to other candidates.

Eliminations are treated singly i.e. one at a time. This is to display the projected transfer patterns more clearly. In the real count process it is probable that 2 or more candidates will be eliminated as one block, at some stage in the count process. As an example, in our projections in the Table our 4th and 5th counts would be combined as one count stage in the actual count, as even if you add the total NI21 vote to the Green Party vote at that stage, it doesn’t reach the next highest candidate i.e. Alliance.

In terms of transfers, these have been modelled on previous patterns in European and Council elections. Transfer % have been rounded-up and rounded-down i.e. in the real count process every candidate usually picks up some transfers from all candidate eliminations, but if we have estimated that this transfer will be less than 0.05% then this was rounded down to zero, and if between 0.05% - 0.1% this is rounded up to 0.1% etc.   

Count process – After 1st preference Vote count:

Stage 1: - Transfer of Anderson (Sinn Fein) surplus: We are projecting that Martina Anderson (Sinn Fein) will be elected on the first count with a small surplus of 1.2%, and this surplus will be transferred first. Sinn Fein had a small surplus in 2009 after the 1st count, but this wasn’t transferred, because it wouldn’t have changed the order of the candidates in the vote, at the bottom of the candidates list, and in fact the surplus was never used in the entire count. We are not projecting that Anderson will necessarily have a much higher surplus this time, but because of the higher No. of candidates we are assuming that her surplus will have to be transferred first, as some of the candidates at the lower end of the 1st preference vote table, may be so close to each other in terms of their 1st preference vote count, that her surplus may change the order of the candidates in the poll. If this is the case, then the Sinn Fein surplus must be transferred before any candidate eliminations. As such, this will probably give a small initial boost to Alex Attwood the SDLP candidate. We project that there will be a 1.2% surplus, with 1% of this being transferable – see note above. Of this 1% we project 0.8% will go to Alex Attwood, and 0.1% each to Anna Lo and Tina McKenzie.

Stage 2: - Elimination of Brotherston (NI Conservatives): As no-one will have reached the Quota after Stage 1, then the Conservative candidate Mark Brotherston would be eliminated next, as he would have the lowest score at that stage of the vote count at 0.8%. We project that there will be 0.7% of his vote that will be transferable, with 0.1% non-transferable – see note above. Of this we project 0.1% will go to Allister, 0.1% to Dodds, 0.2% to Anna Lo, and 0.1% each to McKenzie, Nicholson, and Reilly.

Stage 3: - Elimination of Reilly (UKIP): As no-one will have reached the Quota after Stage 2, then the UKIP candidate Henry Reilly would be eliminated next, as he would have the lowest score at that stage of the vote count at 2.2%. We project that there will be 2.0% of his vote that will be transferable, with 0.2% non-transferable – see note above. Of this we project 1.4% will go to Allister, 0.2% to Dodds, and 0.4% to Nicholson.

Stage 4: - Elimination of Brown (Green Party): As no-one will have reached the Quota after Stage 3, then the Green Party candidate Ross Brown would be eliminated next, as he would have the lowest score at that stage of the vote count at 2.3%. We project that there will be 2.1% of his vote that will be transferable, with 0.2% non-transferable – see note above. Of this we project 0.1% to Allister, 0.4% will go to Attwood, 1.0% to Anna Lo, 0.4% to McKenzie and 0.2% to Nicholson.

Stage 5: - Elimination of McKenzie (NI21): As no-one will have reached the Quota after Stage 4, then the NI21 candidate Tina McKenzie would be eliminated next, as she would have the lowest score at that stage of the vote count at 3.8%. We project that there will be 3.5% of her vote that will be transferable, with 0.3% non-transferable – see note above. Of this we project 0.2% to Allister, 0.8% to Attwood, 0.1% to Dodds, 2.0% to Anna Lo, and 0.4% to Nicholson.

Stage 6: - Elimination of Lo (Alliance Party): As no-one will have reached the Quota after Stage 5, then the Alliance Party candidate Anna Lo would be eliminated next, as she would have the lowest score at that stage of the vote count at 9.2%. We project that there will be 8.7% of her vote that will be transferable, with 0.5% non-transferable – see note above. Of this we project 0.9% to Allister, 4.5% to Attwood, 0.5% to Dodds, and 2.8% to Nicholson.

Stage 7: - Elimination of Allister (TUV): As no-one will have reached the Quota after Stage 6, then the TUV Party candidate Jim Allister would be eliminated next, as he would have the lowest score at that stage of the vote count at 10.5%. We project that there will be 9.7% of his vote that will be transferable, with 0.8% non-transferable – see note above. Of this we project 0.2% to Attwood, 3.6% to Dodds, and 5.9% to Nicholson.

Stage 8: - Election of Dodds (DUP): At Stage 7 Dodds achieves the Quota with 25.3% and is declared elected. As Dodds’s surplus is less than the gap between Attwood and Nicholson at that stage of the count, then Attwood is eliminated as he has the lowest poll-score at this point, and Nicholson is declared elected sub-quota.

Election Result – 3 Candidates elected: Anderson (Sinn Fein), Dodds (DUP), Nicholson (UUP)

As such, we are predicting the same three parties returned again i.e. Anderson (Sinn Fein), Dodds (DUP), and Nicholson (UUP). However, as can be seen, the battle for the last two seats could be a close contest between the DUP, SDLP, and UUP. Our projections are only a model, and are meant to provide a ‘feel’ for a possible way the count could go. It is critical how the various transfer scores are applied, plus the level of non-transferable votes at each stage of the count. There are many different scenarios, with most of these showing SF, DUP, and UUP being the favourites to retain their seats. However, this time, undoubtedly the SDLP are also in with a real chance.


Belfast Telegraph Digital


From Belfast Telegraph