Past Scottish elections in today's council wards

[afleitch]

Sept 16, 2021 11:42:07 GMT hullenedge said:

Sept 16, 2021 10:28:28 GMT afleitch said:

It's getting the balance right. When I ran Gordon in 1992, one set of the results had the Lib Dems way ahead in Gordon proper and the Tories ahead in the Aberdeen part. Not only was this contrary to local election patterns but also from some of the articles/statements at that time about where Malcolm Bruce's team thought their vote was.

Angus is an interesting one. Because I have actual results from 2007 when Andrew Welsh stood, and as he was the candidate/victor on and off right through from the 70's, I haven't completely chucked out patterns of support that he clearly would have massed over several decades.

Glad to hear that you have retained that material.

Copied table 8.5 from The Boundary Commissions (Rossiter, Johnston & Pattie):-



There was an eight seat difference in 1992 for Con/Lab between the census method and R&T. Sadly in depth details not provided.

Danny Dorling used to host those figures on his personal site. He might still do; I'll see if I can find them.

I know in Scotland, they had Stirling notionally Labour.

[afleitch]

Here they are:

www.dannydorling.org/books/ukpopulation/Data.html

Excel: Elections 1832-2001. Party winners and votes. 1992 is the notional data and everything before is an attempt at historic results on those same boundaries.

Note: the figures taken together don't fully equal votes cast in some areas. There used to be a blurb explaining this; % vote shares were used to generate the raw vote totals.

[afleitch]

There's also another issue when looking at the visual impact of this; ward design.

The multi-member wards are actually well designed, usually containing definable communities (with south west Edinburgh perhaps being a bit of an exception) It is also in the nature of Scottish and UK geography that prosperous communities sit tooth by jowl with some of the most impoverished parts of the country.

One example, which you saw in my original estimations and will probably see again is in Renfrewshire. Areas where the Conservatives would be a fairly clear first place (due to West Renfrewshire and Inverclyde) get drowned out by being parts of wards with strongly Labour areas. Lochwinnoch and Kilbarchan are attached to the northern part of Johnstone. Houston with Linwood, Kilmalcolm with part of Port Glasgow. On the opposite side of that, Bishopton, Bridge of Weir and Langbank is drawn in such a way to be hyper Conservative with similar demographics to the richest parts of East Renfrewshire.

So you end up with potentially just one solid blue bolthole even when the Tories win the overall seat.

Around Strathclyde it happens in a few places; Tory West Kilbride in with Labour Dalry (though potentially about to be undone), Strathaven in with Stonehouse.

[aidanthomson]

Sept 16, 2021 9:26:26 GMT afleitch said:

Your analysis is spot on, and close to what I've determined (I have a copy of the alternate notionals you mentioned). The issue with Aberdeen North is as you say; amplifying strong Lib Dem results locally into a national election. There were of course Labour voters in Dyce, Bucksburn, Bridge of Don etc that while in the Gordon seat probably were voting Lib Dem, but Labour's presence in Gordon wasn't negligible between 1983 and 1992 and this shift wouldn't account for their rapid advance in 1997.

Thanks for confirming, and for passing on Danny Dorling's figures, which look very similar to an ex post facto R&T-style calculation I made based on the 1997 general election results. As I saw it, the local LD vote in rural Gordon in 1992 was an anti-Con vote, but in Dyce and Bridge of Don it was anti-Lab, and a lot of Con->LD local tactical support there unwound in the context of a Con/LD marginal in a general election.

Interesting that Dorling's figures also suggest that party support in Turriff was much more in line with the rest of Banff & Buchan than the R&T notional (and the 1994 Grampian region) result would indicate. I have a vague memory that the Con councillor in Turriff was a big figure in the local NFU, which would count for a lot in that neck of the woods.

There were some strange results elsewhere in the 1994 regionals. The Con vote in Bishopbriggs was very low for that area: 901 votes in a division where the party went on to poll 1611 in 1995. The turnout was also noticeably lower there than elsewhere in what is now East Dunbartonshire, and lower than it was across Bishopbriggs in 1995. I wonder if a bundle of Con votes (500? 1000?) got mislaid. It wouldn't have made any difference to the result in the division, but it might have made a difference to the notionals.

[afleitch]

Are you ready?

1992



First off the format. You'll see I've used a bit more colour. The shades represent the % of the vote the winning party got. I felt this was better than showing the difference between winner and second party given the four party system.

Secondly, we have 'abolished wards'. These are wards that as of the 1991 Census effectively didn't have enough voters to really demonstrate a particular political 'tradition', or had very few voters at all; Inverness South, Cumbernauld North and East Kilbride West. You can see these dotted out in white. They are essentially collapsed into the neighbouring ward. The choice of which ward, particularly in the New Towns is based on growth; so East Kilbride South was the growing part of the town before that growth spilled out etc.

We also have the new wards for North Ayrshire, and the Islands. Individual island data here is still derived from the 2007 box data. For more mainland non partisan Regions, there's more of a reliance on 2007 data mixed with demographic data then in the partisan Regions where Regional election data still plays a part.

How I originally did 1992 is on the first page so a little discussion about what has changed.

Highlands - abolishing Inverness South collapsed into Culloden etc get's rid of a Tory base that never should have been there. That flips Nairn to them.

Grampian - This is what some of the earlier discussion focused on; where the 'f' where the Lib Dem voters?

What was interesting about the demographic data (Dorling et al) was that not just in Scotland, but elsewhere in the UK, it worked naturally well when dealing with Labour v Tory contests, but Lib Dem (and SNP) wins weren't as demographically driven; they were local. And that localisation; the catalyst to winning a particular seat was rooted in local performance at local elections.

So Lib Dem strength in Grampian was in Aberdeen. The difference this time around is that they come top in Ellon, but not in North Kincardine. Aberdeen City doesn't change much.

Tayside - Again, here I relied on local results for the same reason as above. In this version the SNP top Brechin, but not Arbroath West. Carnoustie, while on paper should be Tory had a really really good result for Andrew Welsh in 2007, so that's factored in a little in the calculations going further back. In Dundee the SNP vote is more concentrated but the Tory vote less so they 'lose' a ward each. The SNP retain Kinross based on some of the local results, where demographically it should be Tory and in this version gain 'Highland' (Pitlochry, Aberfeldy) based on really quite resilient strength there.

Remember SNP support (thanks to those who contributed to some discussions earlier) in these rural areas correlated to industry and educational level as well as 'where they already had representation'; if the SNP won a random ED in 1978, 1982 or 1986 they were likely to hold it right through to 1994 even if it didn't make any sense based on the makeup of that area.

Central - No real change here, except Labour top in Stirling West.

Fife - Lib Dems lose out in Dunfermline Central with some population deletions.

Lothian - Similar picture, but Labour v Tory contests are a little better balanced. The Tories are ahead in Fountainbridge etc but Labour ahead in Dunbar.

Strathclyde - So here we have great regional results to use which aren't that far off from just going by demographics or by the 2007 base. Voting behaviour here was extremely settled. But there are some differences. Labour now top in Milngavie and with the 'abolition' of East Kilbride West which was packing in non-existant voters, the Tories now edge ahead in Avondale.

Renfrewshire is a nightmare because it's complicated by the Alliance being strong in Inverclyde which never really translated into anything at national level. This happened right through to 2003. Also that strong Tory presence in Bishopton, Bridge of Weir and Langbank is entirely borne out by local/demographics; it's packed in the best Tory parts of rural Renfrew into one ward. In this version I have the Tories ahead in Inverclyde West but Labour ahead in Houston/Linwood etc

Glasgow is fun, because we have unusual challenges to Labour in Govan and Pollok from different parties combined with quite vast wards. The Tories come out top in Pollokshields simply because I now weigh Tommy Sheridan's vote to the local results for Scottish Militant Labour around the time. There were also slightly too many Tories in Greater Pollok ward once I used older demographic profiles and deleted some of the peripheral new-builds.

Dumfries and Galloway and Borders - More reliance on demographics here, kicking the Tories down to second in two wards making it a Lib Dem sweep in the Borders and the SNP just sneaking ahead in Stranraer.

So that's that.

[Harry Hayfield]

Sept 16, 2021 13:52:10 GMT afleitch said:

Here they are:

www.dannydorling.org/books/ukpopulation/Data.html

Excel: Elections 1832-2001. Party winners and votes. 1992 is the notional data and everything before is an attempt at historic results on those same boundaries.

Note: the figures taken together don't fully equal votes cast in some areas. There used to be a blurb explaining this; % vote shares were used to generate the raw vote totals.

Thank you, that's the core data I have been using for my calculations post 2005 (but I couldn't remember where the original link was). Once the new boundaries have been agreed I shall rework those figures onto those new boundaries (and I will be interested to see what happens)

[afleitch]

So almost completed 1974-1987 as before. 1970 is in progress and I'm almost confident enough to go back to 1964 based on that data.

My hope is to go as far back as 1945, but this requires a lot of research on published voting behaviour (which I have) but also demographic change. Whole towns, suburbs have to be erased, cities repopulated, farm labour voting v 'market burgh' in rural areas explored etc. So that's my plan during early next year. If it's even possible!

[afleitch]

1983



Apologies for the closeness of the Alliance and Labour colour scheme in some places.

We have more abolished wards based on the 1981 Census v the 1991 Census, both located around Inverness. This creates an Alliance 'island' with Inverness.

A few changes from the original version. Here you can see where Labour fell back strongly and also where their vote remained resilient. The Alliance often found itself strongest in areas where there was an effective three way tussle for support. Indeed, one vote seperates the Tories from the Alliance in Dunfermline Central. Where a parties winning vote is less than 35% is usually indicative of a three way tie.

The Alliance manages to be ahead in two Paisley wards as before, and one Rutherglen ward. In Inverclyde (not especially clear with the colours) it is ahead in the majority of the wards and in total council vote share, due to strength in Greenock and Gourock split between two seats.

[Clark]

Upper Deeside the strongest Tory ward - based on Braemar, Balmoral and Ballater. No surprise there.

[afleitch]

1970 Draft



So here we have a year I've not posted before. It's not in it's final form, but I thought it would be worth posting.

Population here is based on the 1971 Census. Strictly speaking, the 'snapshot' of each census lasts 10 years; so 1971 would apply to the 1974 elections and 1979. But in this instance in order to 'wind back' I'm having to do a lot by hand so I'm using the 1971 census for this example.

As you can see, more wards 'disappear' particularly around the New Towns, but also settlements like Newton Mearns, Erskine, around Inverness and the Aberdeen periphery.

As before, the BBC/ITN splits will be used for 1979 and the 1974 elections in terms of how seats were split by district councils, but not necessarily internal splits. So Kinross and West Perth being split between Tayside/Perth and Kinross and Central/Stirling, but not Coatbridge (all in Monklands District) being split between Monklands West and East.

This does leave a few areas of concern which I've yet to address. Bothwell ward for example is 'Liberal' in 1979 based on exceptionally unusual results there in the 1980 Districts. But on balance, with regression etc the 1979 Notionals don't do a bad job. But I may make some adjustments where it wasn't strong. I've not fully decided one way or the other!

The 1970 results are in turn affected to an extent by the 1970 Notionals from the Nuffield(?) Study.

Anyway.

Here we see a huge swathe of blue. The SNP are standing in most constituencies at this point, winning one of them and in many seats scoring % shares similar to their 1979-1983 low point; but their support isn't concentrated enough in many seats to show here, barring Fraserburgh.

The Liberals had a tumble backwards from 1966; their support is still where you would expect it, just somewhat more patchy.

The Tories are showing a dominance that they couldn't manage in the 1980's. They hold every ward in 'Clydesdale'; though this may be due to how the 1970'S notionals dealt with the new East Kilbride seat. Winning Cathcart and being a hair behind in Pollok shows why the southern periphery of Glasgow is Tory blue. Outside of the city, the are ahead in places like Ralston, Bishopbriggs and in Burnside.

In Bearsden they are racking up 70% of the vote and while note winning, picking up to 45% in towns like Airdrie. So you are starting to see the strong class based voting that Scotland started to lose showing up here and some potential religous voting here. The religious aspect isn't as a result of any figures I have used; rather the constituency results hyper focusing them somewhat.

Labour are where you would expect them to be strong (even including Caithness and Sutherland) but also where they would't be for some time. The came top in a Perth ward for example. They have very strong showings in the north east fishing villages.

As mentioned, this is simply a draft; it might change up a bit but just thought I'd post something different.

[afleitch]

So it's been a year and things have got in the way, but I have had a look at ways to improve this. I've drafted a new version which should be uploaded soon.

Principles

1. Align with 'official' notionals. Yes there's a few 'misses' built in, but overall the 1979 and 1992 sets weren't too bad.

2. 1992 Notionals used the 1994 Regionals as a base and I have too, with a slight twist....

3. Communities. Regression is quite good for socio-economic derived results, but it does strain when dealing with a four party system. Using 1981 and 1991 Census data, I allowed each officially defined 'populated area'(settlement) of over 10000 to be treated as one unit and using the 1994 local results as the base. In urban (and partisan) District Councils this allowed for neat division with little left over 'rurality.'

The 1995 elections were held in wards (mostly) collapsable into the 1994 Electoral Divisions. The national results and the local pattern was broadly similar to the elections in 1994 so allowed for a subdivision of any larger ED's to create 'communities'

In rural District Councils where there were partisan elections, towns over 10k could be defined as a separate communities (St Andrews, Elgin, Peterhead, Perth etc). How important smaller rural towns were to their respective hinterlands is of course not necessarily based on population size. I used the official Travel to Work areas in these areas giving prominence to centres such as Oban, Huntly and the fishing towns in Grampian.

5. In rural Regions where District Council areas were defined as mostly non partisan (see Denver, Bochel etc), the Districts themselves were mostly used as blocs due to lack of results/contested wards within them. Even with this you will still see local patterns of support at the time (SNP in Wigtown, Labour in Lochaber etc). The exception to this was the town of Inverness which was sizable enough to be treated distinctly.

6. Cities. In the four main cities, consideration was given the the type of housing tenure/stock. So post-war estates were sizeable enough to be considered as separate and growing (or in most cases, contracting) communities. This will be more evident the further back you go (as is the previously discussed phenomenon of New Town politics in the 1970's/1980s.)

As with last time I'll uploaded 1992 soon. The differences aren't that drastic, but I feel it's more in line with the 'lay of the land' than over replying on the 2007 ward data and regression in a number of non partisan areas.

[afleitch]

SNP support of course remained dependent on areas where there had been prior SNP support and with respect to the Lib Dems, 'cleared it's orbit' as the major third party in those areas with the exceptions of Argyll and Bute and Inverness.

You can see the effects of the official notional results showing a higher concentration of support in 1997-2005 Banff and Buchan.

Strong support in Dundee West and Govan can also be seen.

Elsewhere you can see areas of local support in part based on the local results used to derive them but also linked to areas of SNP support in the 1970's; north Clackmannan, Stenhousemuir and Armadale/Bathgate. Also darker spots in Midlothian, Northfield in Aberdeen, Kilmarnock and Oban.

Again, these results are further from the 2007 source material that I initially used, but you can see not only where the SNP broke out in 1997, but also ten years later.

Very low SNP shares in parts of East Renfrewshire, Edinburgh, Helensburgh, Bearsden/Milngavie, Deeside, Berwickshire, St Andrews and Lochgelly etc due to different local factors.

Haven't quite finished off the Islands yet...soon!

[afleitch]

1992



Okay.

Questions?

The Rallings/Thrasher data of course shows in Banff and Buchan, where SNP strength is perhaps too strongly packed into the coast, but also perhaps not; support in Moray is pretty uniform.

A better adherence to local results in rural areas (and less reliance on 2007 data) putting Labour ahead in Dumfries, around Lochaber and the SNP in Wigtown

Party share maps are done and will be uploaded soon.

[afleitch]

Conservative Share 1992



Labour Share 1994


Liberal Democrat Share 1992



SNP Share 1992

[Georg Ebner]

So, on 1992 You have presented these 5 versions:

[afleitch]

That's correct. It's been a bit of an on and off project, which I've explained. The first example leads to the last; I've taken the time to use various regression techniques and local government results to try and hone each version.

And again; nothing is definitive

[afleitch]

On that note, I might rejig Aberdeen a little (I know). As I've used the Rallings and Thrasher notionals, Gordon, Aberdeen N,S and Kincardine were effectively all chopped into two and added up to create an additional seat. What the votes were in those bits were hard to workout and I got someone to run an algorithm. Doesn't make much difference.

1983 and 1987 just need to be mapped

1979 Notionals have been picked apart and can be inputted (they were just District based last time)

1974 can be ran off that.

1970 notionals are also deconstructed.

Post war demographics are sorted too.

So I can get back to 1966 out over the next month or so! Earlier than that, I've been dry running for ages.

[ntyuk1707]

2021


Support for the Conservatives grows in southern and north-eastern Scotland, while falling  back in the Central Belt. Labour come first in just 9 wards. The Liberal Democrats jump into first place in Caithness, eastern Sutherland and Howe of Fife.

Thanks to afleitch for letting me use his map format. 

[jamie]

Aug 28, 2023 17:20:01 GMT ntyuk1707 said:

Support for the Conservatives grows in southern and north-eastern Scotland, while falling back in the Central Belt. Labour come first in just 11 wards, most of which are located in the Dumbarton constituency. The Liberal Democrats jump into first place in Caithness, eastern Sutherland and Howe of Fife.

What’s the distribution of the vote like in Dumbarton constituency? I assume the SNP would be a lot stronger in Dumbarton than Helensburgh so I would have expected them to be ahead in at least 1 of its wards given they only lost by 4%. That said, even with tactical voting I'm guessing Labour don’t have much of a lead in Helensburgh, which presumably means they’re narrowly ahead in both ends of the constituency? 

[ntyuk1707]

Aug 28, 2023 23:16:08 GMT jamie said:

Aug 28, 2023 17:20:01 GMT ntyuk1707 said:

Support for the Conservatives grows in southern and north-eastern Scotland, while falling back in the Central Belt. Labour come first in just 11 wards, most of which are located in the Dumbarton constituency. The Liberal Democrats jump into first place in Caithness, eastern Sutherland and Howe of Fife.

What’s the distribution of the vote like in Dumbarton constituency? I assume the SNP would be a lot stronger in Dumbarton than Helensburgh so I would have expected them to be ahead in at least 1 of its wards given they only lost by 4%. That said, even with tactical voting I'm guessing Labour don’t have much of a lead in Helensburgh, which presumably means they’re narrowly ahead in both ends of the constituency? 

Well spotted - this was an error.

Jackie Baillie led in the Helensburgh & Lomond side of the constituency by over 20%, while the SNP topped the poll in the Dumbarton side of the constituency by approximately 7% of the vote.