Dumfries & Galloway

2015 Result:
Conservative: 16926 (29.9%)
Labour: 13982 (24.7%)
Lib Dem: 953 (1.7%)
SNP: 23440 (41.4%)
UKIP: 1301 (2.3%)
MAJORITY: 6514 (11.5%)

Category: Semi-marginal SNP seat

Geography: Scotland, South of Scotland. Part of the Dumfries and Galloway council area.

Main population centres: Dumfries, Stranraer, Kirkcudbright, Castle Douglas, Dalbeattie, Wigtown, Newton Stewart.

Profile: A large rural seat that covers the south-eastern part of Scotland, primarily covering the old historic counties of Wigtownshire and Kirkcudbrightshire, as well as the majority of the town of Dumfries. Most of the seat is mountains and farmland, with small market and tourist towns dotted amongst the lochs and along the coast. Dumfries is by far the largest settlement, known as the "Queen of the South" and an administrative and economic centre for southern Scotland, followed by the ferry port of Stranraer at the western end of the seat.

Politics: Dumfries and Galloway was created in 2005 and won by the Labour party, who retained it in 2010 before losing to the SNP in 2015. The predecessor to the seat had been Galloway and Upper Nithsdale, historically a Conservative seat and won back by the Conservative at the 2001 election, perhaps helped by the outbreak of Foot and Mouth disease in the seat. Despite notional figures that in hindsight look rather questionable the addition of the town of Dumfries under the new boundaries in 2005 probably helped Labour take the seat.

Current MP
RICHARD ARKLESS (SNP) Born 1975, Stranraer. Educated at Glasgow Caledonian University. Former solicitor. First elected as MP for Dumfries & Galloway in 2015.
Past Results
Con: 16501 (32%)
Lab: 23950 (46%)
LDem: 4608 (9%)
SNP: 6419 (12%)
Oth: 695 (1%)
MAJ: 7449 (14%)
Con: 18002 (35%)
Lab: 20924 (41%)
LDem: 4259 (8%)
SNP: 6182 (12%)
Oth: 1524 (3%)
MAJ: 2922 (6%)
Con: 12222 (34%)
Lab: 7258 (20%)
LDem: 3698 (10%)
SNP: 12148 (34%)
Oth: 588 (2%)
MAJ: 74 (0%)
Con: 12825 (31%)
Lab: 6861 (16%)
LDem: 2700 (6%)
SNP: 18449 (44%)
Oth: 1183 (3%)
MAJ: 5624 (13%)

2015 Candidates
FINLAY CARSON (Conservative) IT consultant and former farmer.
RUSSELL BROWN (Labour) Born 1951, Annan. Educated at Annan Academy. ICI worker. Former Dumfries and Galloway councillor. Annandale and Eskdale councillor 1988-1996.MP for Dumfries 1997 to 2015. PPS to Lord Williams 2002-2003, PPS to Baroness Amos 2003-2005, PPS to Alistair Darling 2005-2007, PPS to Jim Murphy 2007-2010.
ANDREW METCALF (Liberal Democrat)
RICHARD ARKLESS (SNP) Born 1975, Stranraer. Educated at Glasgow Caledonian University. Solicitor.
Comments - 228 Responses on “Dumfries & Galloway”
  1. You are making this much more complex than it needs to be. Subsamples are not designed to be used in the way that you are doing, and polling companies do not guarantee that errors are removed at that level. Thus, we can work out an error size based on the subsample size, but there may well be other errors that we don’t know of. The suggestion that YouGov have too many people born in England in their Scottish voters panel is one such example. Obviously you would correct for this in a Scotland-wide poll, because we know that English-born voters vote significantly differently from the Scottish population as a whole. I don’t know if they would bother for a UK poll, or if this correction would only be done in the various weightings and so on (in which case the subsample figures might be very unreliable, but the poll as a whole would be fine).

    Also, there was no Ashcroft National Poll or Scottish polling that I can find that was released on the 10th October.

  2. @ Simon – regardless there clearly remains an underlying bias in favour of the SNP in his polls, he has never published a national poll in Scotland. His subsamples in national polls are unquestionably out of line with all other sample/subsamples, I don’t want to argue over your opinion on the merit and significance of his polling/subsample polls from national UK polls which is a different matter entirely. We have no real way of ensuring his constituency polls are reliable and reflect national polls in Scotland without using his subsample polls, which I feel are biased. Bias can be a result of various issues which different pollsters deal with differently. In my view subsamples have merit and I would think that they are correctly weighted – particularly due to the SNP’s influence on a UK-wide scale. Taking all his subsamples together also indicates an SNP bias which as I have raised before is not experienced in the bulk of other subsample polls when averaged up even with your hypothetical margin of error this is insanely out of touch. You can’t just take a sole pollster alone and claim that their results are more accurate than the rest when the general trend contradicts this, which is why I am spectical as to how reliable his constituency polls actually are particularly given that in the past his polls have generally overemphasised the level of support for the winning side: which is why I feel an average is a more accurate reflection of current public opinion. Time and time again these “random errors” have persistently shown themselves in his subsamples for Scotland, even the most recent polls show this. Margin of error should act as a guideline rather than a rule, and are reduced through averages: this is not the case for Lord Ashcroft! A last point- the subsamples for most other pollsters are all far more close together by comparison to Lord Ashcroft’s polls which have nearly always been well above average month after month.

    The Scottish polls were for the month of October average. The Lord Ashcroft poll was conducted in 10-12 October, values derived from page 5 table 3.

  3. NTY UK

    As James says above, it’s generally accepted and understood that Ashcroft uses polling companies that are members of the British Polling Council to set up and run his polls, before applying his own weighting once he has the results. If he did not use BPC pollsters, his polling would not be given the same importance by the polling community and would largely go unreported.

    It has also been observed that Ashcroft uses two different pollsters for his national polling and his constituency polling.

    Therefore, we should only be using his nationwide polls to help give us a picture of the nationwide situation. And his constituency polls are good indicators of local situations.

    There’s no reason to be looking at Scottish crossbreaks of Ashcroft nationwide polls if we have other polling companies conducting Scotland-only polls with a much smaller margin of error.

  4. Whether you think subsamples are correctly weighted or not doesn’t change the fact that they generally are not. Also, the poll that you mentioned had a subsample size in Scotland of 86, of which 28 were not usable participants (Don’t knows, won’t say, won’t votes). That leaves 56 people across Scotland who offered a voting intention. Drawing conclusions from that is like paying attention to the guy who claims to have conducted straw polls in seats across the country.

    In any case, the findings of Lord Ashcroft’s constituency polls are also broadly backed up by other polling work. Here’s a link to analysis by John Curtice, using work carried out by an entirely different pollster.


  5. Piemontis is right to say that Ashcroft’s Scottish crossbreaks for national polls needn’t be looked into too deeply. Indeed, there are perfectly sound reasons for slightly overweighting SNP and PC in nationwide polls which give the five GB parties as topline figures, as its a reasonably reliable way of factoring in the English “other”, which is usually harder to get right. By contrast, in a Scottish constituency poll getting the SNP right is presumably consideration number one.

    What’s less clear is how useful Ashcroft’s constituency polling is generally, due to a lack of evidence as to whether the questions and weighting are or are not reliable models to use in competitive and often crowded races.

    The best way to use constituency polls is to assume that any methodological issues are systemic and linear in nature – in other words that he has probably got each party’s prospects for vote share in roughly the right order, but may have used a method which consistently over- or under-weights individual parties, therefore mis-judging some seats by several points. That said, there’s little reason to conclude with any confidence that constituency polls are overestimating the SNP’s likely seat count. Or at least, no more than to assume that constituency polls might underestimate any difference in turnout between voters for UK-wide and Scotland-only parties.

  6. @ Simon – John Curtices observations differs from that of the Lord Ashcroft polls which suggest that Labour will be left with 5 seats in Scotland vs Lord Ashcroft’s 2. As he suggests the polls commissioned by Comres should be treated with some kind of caution, but what it does highlight is that the majority of seats look as if at the moment the SNP would win them, yet this could all change and I still hold some doubts over Lord Ashcroft’s opinion polls in general, but I accept that in constituencies with huge SNP leads it’s likely that the SNP will gain them – I’m just doubtful over his polls in marginal seats where vote share doesn’t match up with the national picture, across most of his constituency seats there’s been a general swing away from the Tories – I’m doubtful over how accurate this is along with questionable vote shares in some areas which don’t appear to be in line with even the referendum eg. Edinburgh SW vs Glasgow NE. Furthermore, there’s again quite a considerable difference in the findings of Comres and Lord Ashcroft in that Comres suggests far more Labour seats, this might not seem significant but under a FPTP system it makes all the difference.

  7. I am the candidate listed above my photo ex are not on your site please would you contact the LIB DEM office in Edinburgh and correct this

    Thank You

  8. UKIP PPC: Geoff Siddall (of Warrington).

  9. I wonder how much deposit loses will cost UKIP in Scotland.

  10. I think this seat will be an SNP or Con gain (probably with a recount).

    Any unionist tactical voting makes a Con win more likely, as does any swing from SNP back to Labour.

  11. @Andrew

    Do you think you will hold onto your deposit?

  12. Surely Lord Ashcroft has to have this one lined up….

  13. Yes, I would have expected Ashcroft to have done a poll here too.

  14. My bad, he did one here in March. I’m not quite sure how I managed to miss it…..


    SNP – 34
    Con – 30
    Lab – 28
    LD – 2
    Others 6

  15. There was some talk a couple of months ago on the Southampton Itchen thread (from myself and others) about how, if the Tories only took one seat from Labour this year, it would be Itchen. Now, however, Dumfries and Galloway almost looks like a better chance….

  16. If his other polls are anything to go by, the SNP will have gone up slightly since then. Best chance of preventing an SNP gain is large scale tactical voting. Only problem is that the Tories and Labour vehemently disagree on who should be the favoured ‘unionist’ party.

  17. My prediction is that the Lib Dems could win here…….a lost deposit – !!! đŸ™‚

  18. Yes JAMES I agree…

    the SNP and LAB vote would have to BOTH be at 30% or 31% for CON to come through (CON vote unlikely here to be 35%+ ) – not unforeseeable but looks unlikely, probably a 20% at best.

  19. “Only problem is that the Tories and Labour vehemently disagree on who should be the favoured ‘unionist’ party.”

    A very important point, which will boost the SNP’s chances considerably in seats like this, East Renfrewshire, West Aberdeen & Kincardine, Berwickshire, Edinburgh West and perhaps even DCT.

  20. Just to add, I think that whilst many Tories might agree to vote tactically for Labour, it will be far harder to get Scottish Labour voters to vote for a Tory….in many cases they might actually prefer the SNP.

  21. There’s been plenty of tactical voting by Labour and SNP supporters for each other in the Borders in the past.

  22. Exactly.

    In my own predictions I’ve assumed Labour will hold around 10 of their Scottish seats but I’m starting to think this may be way too optimistic.

  23. My experience from having been a candidate in Local Government elections since 1979, and having studied election results, is that, for example, in Ince ward of Wigan MBC in the early 1980s in a by election there were about 150 votes for Conservative ( or Independent on right wing). That was probably before the Falklands and before the Miners strike.

  24. My own prediction is for SNP to have 43-46 seats out of the 59 Scottish seats.

    I am assuming a small swingback to LAB enabling them to hold some seats in which, currently, SNP are moderately ahead.

  25. The bookies no longer offer the bet of Labour to lose seats in Scotland and the Tories to gain seats in Scotland. It was 12/1 IIRC before Christmas.

  26. The latest polls are indicating 51 SNP MP’s….if Nicola Sturgeon’s ratings improve further then a clean sweep is on the cards with the exception of course of Dumfrieshire, Clydesdale and Tweedale and Selkirk, Roxborough and Selkirk…..

    Imagine if Milliband wins a minority government situation and then has the added humiliation of having fewer or even the same number of Scottish MP’s as the Tories – !!!

  27. The latest polls show the SNP winning DCT

  28. So if Sturgeon does even better than now she’ll obviously be even further ahead there and quite possibly in Berwickshire as well.

  29. Maybe our bet is voided already….Milliband 0 – Cameron 0 – !!

  30. SNP to win ALL 59 Scottish seats is just a 3/1 chance with Ladbrokes (implied 25% chance)

  31. It’s interesting to ask which would be the last non-SNP seat in Scotland…ironically, most likely Orkney & Shetland

  32. The SNP are 3/1 to win Orkney and Shetland at Ladbrokes (or at least they were yesterday). That’s a better value bet than the SNP to win all 59 seats, which doesn’t account for upsets.

  33. What is the name of the seat which houses Faslane?

  34. Faslane is in Argyll & Bute which should be a very likely SNP gain.

  35. Labour hold by an increased majority, perhaps 9000. The polls are inaccurate, in that in the final few days people in Scotland will turn back to Labour in droves…SNP has never won this seat before.

  36. Do people still look this as the SNP in front or look at it as a 3-way marginal at this late stage?

  37. The SNP held Galloway and Upper Nithsdale, which makes up a large proportion of this seat, from 1974-79 and 1997-2001, so it’s inaccurate to say they’ve never won here.

    Woof – both? It’s a 3 way fight, but the SNP are likely favourites to win.

  38. Labour Hold. SNP 2nd.

  39. I travelled through this seat in mid-Apr and was taken back by a) the scenery and b) the number of Finlay Carson posters along the road at a time when I had barely seen any out on display back in my home constituency.

    Obviously I am aware that the number of posters bares no specific relationship to the number of votes, but I am also aware that some Scottish Tory friends are confident that the Tory vote is remaining relatively stable compared to 2010 in national polling and I canvassing returns.

    The interesting thing about the Finlay Carson posters is that you could barely see the “Scottish Conservatives and Unionists” label tucked away in the corner. Carson is clearly seen as the stronger pull in a rural constituency, similar to the manner in which Alex Fergusson has campaigned (successfully) in the constituency in Scottish Parliament elections.

    I’d like to see the Tories sneak through the middle, but can’t see how they’d achieve any more than 33% of the vote, even that will be difficult with a UKIP candidate attracting a derisory 1-2%. Any Unionist voting tactically will surely vote Labour as the best positioned party to defeat the SNP, and given the low SNP 2010 vote they’d need to take half of the 2010 Labour and Lib Dem vote which seems unrealistic but given the recent Scottish polling entirely feasible.

    SNP – 34%
    Conservative – 32%
    Labour – 30%
    Lib Dem – 3%
    UKIP – 1%

    Definitely a 3 way toss up dependent upon Unionist tactical voting and I’d hate to see the SNP grab a seat that is so staunchly Unionist.


  41. SNP gain. Con second. Labour third. I believe that, ironically enough, a number of Conservative voters tactically voted for Russell Brown (Labour). They won’t make that mistake again! Same happened in West Aberdeenshire & Kincardine where some Tories voted for LibDem MP Sir Robert Smith (who came third).

  42. @capercaillie

    I think I made that point about tactical voting here long before the election, i.e. that in the places where it was most likely to succeed (here, WAK, BRS), there was least likely to be agreement between the “unionist” parties as to who was best placed to “stop the Nats”.

  43. Very poor result for Labour here! Are they finished here for the foreseeable future?

  44. Alex Fergusson won’t be standing for the SP seat here in 2016, which means that it will be at best an uphill battle for the Conservatives to hold on. I am sure the Conservatives will make up the difference either on the list or by winning next door Dumfriesshire from Labour, and I would think that as in the GE the total Conservative vote should be up.

    Pity to see the man go, he was always very well regarded by people in the area.

  45. Russell Brown was beaten by the Conservatives by nearly 3000 is a massive turn around in his fortunes following his massive 7449 in 2010.

    What’s interesting is that Dumfriesshire was not historically a safe Tory constituency. Its majority was only 16.8% in 1959 when places like Edinburgh South, Glasgow Cathcart and Glasgow Hillhead were solid Tory seats.

    Hector Munro MP from a 1964 by election to 1997 (where he retired) seems to have offset the natural swing to Labour that occurred in the rest of Scotland. Brown defeated Munro’s successor on a huge swing and a majority of nearly 10000.

    When Brown won Dumfries & Galloway in 2005 on a larger majority than the 2001 notional the notional result was questioned.

    When Brown increased his majority in 2010 he may have made greater inroads into Galloway than consolidate support in Dumfries but his result was no greater than the strong Labour performance in East Renfrewshire or Ochil & South Perthshire (both like D & G, dramatically lost by Labour in 2015).

  46. This constituency is much more urban than the former Galloway & Upper Nithsdale.

    Russell Brown’s personal vote is also more likely to go to the SNP.

  47. The Tories could beat Labour in Dumfriesshire and Eastwood but be overtaken by the SNP. They are also not certain to hold Ayr and Galloway & West Dumfries but Berwickshire will be a shoe in.

    The Tories should win between 1 to 5 of the above constituency seats.

    I think that the Conservatives will fall further behind in Edinburgh Pentlands in the absence of David McLetchie and I fail to see why Ruth Davidson did not stand there instead of Edinburgh Central.

    If I was to place a guess at seats the Conservatives may close the gap on the SNP I would consider Perthshire North, Perthshire South & Kinross, Angus South, Angus North & Mearns and Aberdeenshire West.

  48. Apparently the Scottish Tories will be spending 80% of their campaign budget on list seats, and they expect to list all constituency MSPs except John Lamont.

  49. I was surprised at the extent of the collapse of Russell Brown’s vote given his solid win in 2010 and the relative strength of the Labour vote in Galloway & West Dumfries and the notional gain from the Conservatives of Dumfriesshire in the 2011 Holyrood elections.

  50. The notionals in this area have always been sketchy at best.

Leave a Reply

NB: Before commenting please make sure you are familiar with the Comments Policy. UKPollingReport is a site for non-partisan discussion of polls.

You are not currently logged into UKPollingReport. Registration is not compulsory, but is strongly encouraged. Either login here, or register here (commenters who have previously registered on the Constituency Guide section of the site *should* be able to use their existing login)