2015 General Election - a record breaker

[Deleted]

Jan 13, 2016 0:59:53 GMT Davıd Boothroyd said:

There probably isn't much point calculating swing at the Scottish Parliament elections, because it doesn't represent how voters behave.

That's convenient

[YL]

Obviously the Lab/SNP swing is more meaningful than the Lab/Con one in Glasgow NE, but David does have a point that it's a bit dodgy to compare it with swings between other pairs of parties, as it's not the same sort of voter move. I'm happy with calling it the biggest Lab to SNP swing ever.

When I've done maps of swings I've used the same pair of parties for each constituency/ward (Lab/Con for general elections, Lab/LD for Sheffield council). However, the extreme cases can be a bit weird, as with the 1970 Merthyr record, which was not in any way an impressive performance by the Conservatives: their vote was down on 1966. The constituency I live in produced the largest Con to Lab swing in 2015, but I don't think anyone thinks a Con to Lab shift was what was really going on.

The "swing between the top two parties" can give silly results too. Belfast South in 2005 is a well-known example, solemnly recorded on the BBC website as a 13.4% swing from SDLP to DUP.

[YL]

Here's the Yorkshire & the Humber Lab/Con swing map I made:

(base map due to Sibboleth ).

The four darkest red seats are Sheffield Hallam, Bradford East, Bradford West and Leeds North West, i.e. three Lib Dem defences challenged by Labour and one law unto itself.

[Davıd Boothroyd]

Jan 13, 2016 8:07:16 GMT doktorb🏳️‍🌈🏳️‍⚧️ said:

Jan 13, 2016 0:59:53 GMT Davıd Boothroyd said:

There probably isn't much point calculating swing at the Scottish Parliament elections, because it doesn't represent how voters behave.

That's convenient

Damned inconvenient actually. It means it's very difficult to predict accurately how SP voting intention translates into seats. The calculation of Conservative-Labour swing relies on the fact that voters don't move proportionately - so a Conservative vote change of 5% up or down is the same change whether the Conservative vote started at 10% or at 70%. That does not apply to votes moving between Labour and SNP, as the table showing SNP and Labour change ordered by Labour vote in 2010 showed.

[froome]

The concept of swing is meaningless because it is very difficult to predict accurately how any voting intention translates into seats. That is just as true of Lab/Con swings at constituency level.

It is quite possible for there to be a swing from Labour to Conservative in a seat that Labour gain, and for there to have been no voters there (or very few) who actually did vote Labour in one election and Conservative in the next.

Predictions of seats changing hands relies of the nuances of votes in seats between all the parties, which the idea of a swing cannot show.

None of this undervalues having a statement of how the votes of Conservative and Labour compare, but using a swing to describe it misrepresents it to voters.

[carlton43]

Jan 13, 2016 10:25:30 GMT froome said:

The concept of swing is meaningless because it is very difficult to predict accurately how any voting intention translates into seats. That is just as true of Lab/Con swings at constituency level.

It is quite possible for there to be a swing from Labour to Conservative in a seat that Labour gain, and for there to have been no voters there (or very few) who actually did vote Labour in one election and Conservative in the next.

Predictions of seats changing hands relies of the nuances of votes in seats between all the parties, which the idea of a swing cannot show.

None of this undervalues having a statement of how the votes of Conservative and Labour compare, but using a swing to describe it misrepresents it to voters.

Well, after laboriously working my way from the least marginal Liverpool seat down the list for 100 instances, I still think there is much to learn from the swings discovered. The Classic Swing between Conservative and Labour is the king of trends and as instructive as ever. I hope David would agree that the essence of his position is two-fold with the essential truth that only two parties can shape an administration being the core of it all; so the movement between those two parties in all seats where both make contest has a real and basic meaning in the context of the whole contest. Those two parties have been at the centre of the national contest for 100-years. The second point is that all other contests, whilst being interesting and often very dramatic, and even when materially affecting the national result (SNP in 2015) are still peripheral within the national and the historic context. I am now convinced that David is quite right and I have therefore altered my vote above.  

[carlton43]

To address the Scottish Problem I show the result in the 'Carlton method' from Dundee East being the most safe seat in Scotland and the most safe seat held by the SNP. As an aside it is with a certain irony that one notes that the 2010 most marginal seat for Labour in Scotland is the only seat that they now hold!

DUNDEE EAST
TO 72%
Majority Margin 39.7%
SNP 59.6 + 21.9 LAB 19.9 - 13.4
Classic Swing (Con/Lab) LAB 6.6% To CON.
Current Swing (Between Winner and Runner-Up) LAB 17.7% TO SNP.
This was an SNP Hold with no gain and no change of Runner-Up to occasion any more swing calculation.

I think that provides a succinct, meaningful and useful summary. I see both swings to be useful and necessary.

[carlton43]

Let me provide one more example of the 'Carlton Method' to show an oddity of a result in the very safe seats, No.8 Manchester Gorton.

MANCHESTER GORTON
TO 57.6%
MAJORITY MARGIN 57.3%
LAB 67.1 + 17.0 GREEN 9.8 + 7.0
CLASSIC SWING (CON/LAB) CON 9.2% TO LAB
CURRENT SWING (Between Winner and Runner-Up) GREEN 5.0% TO LAB
HISTORIC SWING (Between Winner and 2010 Runner-Up) LD 22.7% TO LAB
That provides a full picture of the utter dominance by Labour in a compact and succinct format.

[johnloony]

When considering individual constituencies (rather than broad national changes) I sometimes think of the "swing" as being the one between whichever two parties that had the biggest changes. In the case of 2010-to-2015, that is often from Lib Dem to UKIP. I sometimes (in my own mind) call it a "shift" rather than "swing".

This is a subject in which various different definitions and terms are useful, depending on what one is wanting to do. Using the Butler/Boothroyd Con/Lab swing to predict the overall result of a general election is only useful if the average national swing is fairly close to what happens in the broad range of marginal constituencies (one can ignore what happens in very safe seats). The U.K. Is still broadly a two-party Con/ Lab country (with the two main parties broadly geographically spread) despite the recent upsurge in support for UKIP, SNP and Green Party. If it were not - if it drifted towards a more fractured multi-party system like India or Belgium - it wouldn't be.

[mick745]

if a party wants to win any seat it doesn't hold, the task facing it is the 'swing' needed. therefore a swing can be expressed between any party or parties, and represents the shift of voters from one to the other. the highest swing represents the most successful shifting of votes from one party to another in any particular seat.

[Davıd Boothroyd]

Jan 14, 2016 22:17:06 GMT mick745 said:

if a party wants to win any seat it doesn't hold, the task facing it is the 'swing' needed. therefore a swing can be expressed between any party or parties, and represents the shift of voters from one to the other. the highest swing represents the most successful shifting of votes from one party to another in any particular seat.

There are several problem with that analysis.

1) A party which wants to win in a seat which it does not hold can do so without winning any extra votes, but just by having the other parties divide their votes up so that none of them is larger.
2) Of course a calculation can be made of 'swing' between any two parties. You could also calculate the square root of their votes. Doesn't make it a meaningful calculation.
3) You seem to be arguing that a calculation of swing must necessarily mean that there has been a shift of actual voters. It doesn't. Whichever parties you use, swing is going to be the overall result of many tens of thousands of individual changes - between parties, of voters deciding to participate in the election or deciding not to participate, of voters dying or moving, of voters coming of age etc.
4) The point of swing is to be able to compare two different areas. Your calculation makes it impossible to do so.
5) You seem to have forgotten the politics. It is a great deal easier for voters to shift from the Liberal Democrats to Labour or Conservative, and vice versa, than it is for voters to shift directly from Labour to Conservative. It's also a whole lot more meaningful.
6) Most crucially, if you just take swing as a simplistic calculation of half the biggest rise and biggest fall in party percentages, then you are led down the blind alley in which a '10% swing from Liberal Democrat to Conservative' is treated as being the same magnitude as a 10% swing from Labour to Conservative when it bloody well isn't!

[carlton43]

My trawling through the swing of all constituencies has got to Number 165 and at Number 137 is Richmond (Yorks). I advance my stats for this in my own defence as much as for any other reason.

RICHMOND (YORKS)
TO 64,7%
CON 51.4 - 11.4
UKIP 15.2 (First candidature)
Classic Swing (Con-Lab) Con 4.7 to Labour.
Current Swing (Winner-Runner-Up) Con 13.3 to UKIP
Historic Swing (Winner-2910 Runner-Up) LD 0.7 to Con

Now, why have I singled this out? Well, because it is a stolid, rural, deeply conservative (lower case) seat. For such a Conservative seat with a very settled-state electorate, the TO is amazingly low. The heavy fall in the Conservative vote 11.4 is way out of character for such constituencies in 2015. The swing from LD is surprisingly low and the Classic swing to Labour rather high for the type of seat. Why this thusness?

My interest is pushing this is a serious slur called upon me in the GE run up for suggesting on the constituency thread that the choice of candidate here at Richmond was a very serious error of judgement. It was suggested that I had made such comment because I didn't like the candidate for personal racist reasons. Like virtually everyone here I am at heart a bit of an anorak and most of my factual comment is 100% informed and guided by genuine interest and usually dispassionate analysis. It is not Jones-like invective or polemical rhetoric. I was both hurt and aggrieved at that time to have had my forecast traduced by a personal attack.

So, here is the clear evidence of what a party can do if it chooses a clearly unsatisfactory candidate 'in the eyes of a significant number of the electorate'. That was not my view, my hoped for result, my attack upon the candidate. It was me a fellow anorak opining a view of a possible outcome. To have chosen a very 'foreign', very southern, very urban, very grey suit banker, very thin and very young man for an expansive very rural, hunting fishing, beer drinking, agricultural, northern seat seemed barmy. And it was. In a rather good year the party suffered a considerable reverse, not because I willed it, but because I was just plain right about the choice. Parties can be silly and the Conservative party made some very duff choices in 2015. This was one of them. This is what happened!

[Davıd Boothroyd]

Jan 17, 2016 14:17:49 GMT carlton43 said:

My trawling through the swing of all constituencies has got to Number 165 and at Number 137 is Richmond (Yorks). I advance my stats for this in my own defence as much as for any other reason.

RICHMOND (YORKS)
TO 64,7%
CON 51.4 - 11.4

Now, why have I singled this out? Well, because it is a stolid, rural, deeply conservative (lower case) seat. For such a Conservative seat with a very settled-state electorate, the TO is amazingly low. The heavy fall in the Conservative vote 11.4 is way out of character for such constituencies in 2015. The swing from LD is surprisingly low and the Classic swing to Labour rather high for the type of seat. Why this thusness?

My interest is pushing this is a serious slur called upon me in the GE run up for suggesting on the constituency thread that the choice of candidate here at Richmond was a very serious error of judgement.

Is there not an explanation which goes along these lines:

1) Richmond is a safe Conservative seat generally, but in Parliamentary terms it suddenly became a massively safe one in the 2001 general election.
2) That was also the election in which the sitting Conservative MP stood for election as the Conservative Party leader.
3) The phenomenon of party leaders doing exceptionally well in their own seats is very well known.
4) The sitting MP then ceased to be party leader but did not cease to be a prominent politician, and their vote did not fall back.
5) After 26 years, the sitting MP had amassed a considerable personal vote anyway.

Therefore:

6) The striking fall in the Conservative vote would have happened anyway, whoever the Conservative association had chosen as their candidate.

[greatkingrat]

Had they selected a local candidate I'm sure they would have done a bit better, but I don't think it had as big an impact as you are claiming. William Hague no doubt had a significant personal vote, plus there were two independent candidates (including the local council leader) who got nearly 10% between them, so it is not surprising the Conservative vote dropped more here than in neighbouring constituencies.

[carlton43]

Jan 17, 2016 14:36:49 GMT Davıd Boothroyd said:

Jan 17, 2016 14:17:49 GMT carlton43 said:

My trawling through the swing of all constituencies has got to Number 165 and at Number 137 is Richmond (Yorks). I advance my stats for this in my own defence as much as for any other reason.

RICHMOND (YORKS)
TO 64,7%
CON 51.4 - 11.4

Now, why have I singled this out? Well, because it is a stolid, rural, deeply conservative (lower case) seat. For such a Conservative seat with a very settled-state electorate, the TO is amazingly low. The heavy fall in the Conservative vote 11.4 is way out of character for such constituencies in 2015. The swing from LD is surprisingly low and the Classic swing to Labour rather high for the type of seat. Why this thusness?

My interest is pushing this is a serious slur called upon me in the GE run up for suggesting on the constituency thread that the choice of candidate here at Richmond was a very serious error of judgement.

Is there not an explanation which goes along these lines:

1) Richmond is a safe Conservative seat generally, but in Parliamentary terms it suddenly became a massively safe one in the 2001 general election.
2) That was also the election in which the sitting Conservative MP stood for election as the Conservative Party leader.
3) The phenomenon of party leaders doing exceptionally well in their own seats is very well known.
4) The sitting MP then ceased to be party leader but did not cease to be a prominent politician, and their vote did not fall back.
5) After 26 years, the sitting MP had amassed a considerable personal vote anyway.

Therefore:

6) The striking fall in the Conservative vote would have happened anyway, whoever the Conservative association had chosen as their candidate.

In a word David.....No!

Did those circumstances have any effect at all? Almost certainly yes. But diminished swing from LD and heavier swing to Labour....No. 11.4% drop in share......No. 

[carlton43]

Jan 17, 2016 14:40:44 GMT greatkingrat said:

Had they selected a local candidate I'm sure they would have done a bit better, but I don't think it had as big an impact as you are claiming. William Hague no doubt had a significant personal vote, plus there were two independent candidates (including the local council leader) who got nearly 10% between them, so it is not surprising the Conservative vote dropped more here than in neighbouring constituencies.

But would that have happened with a more appropriate candidate in first instance?

[Davıd Boothroyd]

Compare with some other cases where a former Conservative Party leader stood down and was succeeded by a new candidate. In 2010, the Conservative vote nationally increased by 3.7%, but in Folkestone and Hythe, it fell by 4.5%. In 2001, the Conservative vote nationally increased by 1.2%, but in Huntingdon, it fell by 5.4%.

[carlton43]

Jan 17, 2016 15:02:47 GMT Davıd Boothroyd said:

Compare with some other cases where a former Conservative Party leader stood down and was succeeded by a new candidate. In 2010, the Conservative vote nationally increased by 3.7%, but in Folkestone and Hythe, it fell by 4.5%. In 2001, the Conservative vote nationally increased by 1.2%, but in Huntingdon, it fell by 5.4%.

Good stuff David and not deaf to your point, but they were different elections in years with a quite different dynamic. Sedgefield has gone off but not just because the nice Mr. Blair is no longer there!!

[mick745]

Another interesting point is that there was a 5% difference in turnout between England and Scotland, with Scotland ahead. In England it was 66% and Scotland a 71% turnout. I put it to you that this a record differential between the two countries.

[Deleted]

Jan 17, 2016 14:17:49 GMT carlton43 said:

My trawling through the swing of all constituencies has got to Number 165 and at Number 137 is Richmond (Yorks). I advance my stats for this in my own defence as much as for any other reason.

RICHMOND (YORKS)
TO 64,7%
CON 51.4 - 11.4
UKIP 15.2 (First candidature)
Classic Swing (Con-Lab) Con 4.7 to Labour.
Current Swing (Winner-Runner-Up) Con 13.3 to UKIP
Historic Swing (Winner-2910 Runner-Up) LD 0.7 to Con

Now, why have I singled this out? Well, because it is a stolid, rural, deeply conservative (lower case) seat. For such a Conservative seat with a very settled-state electorate, the TO is amazingly low. The heavy fall in the Conservative vote 11.4 is way out of character for such constituencies in 2015. The swing from LD is surprisingly low and the Classic swing to Labour rather high for the type of seat. Why this thusness?

My interest is pushing this is a serious slur called upon me in the GE run up for suggesting on the constituency thread that the choice of candidate here at Richmond was a very serious error of judgement. It was suggested that I had made such comment because I didn't like the candidate for personal racist reasons. Like virtually everyone here I am at heart a bit of an anorak and most of my factual comment is 100% informed and guided by genuine interest and usually dispassionate analysis. It is not Jones-like invective or polemical rhetoric. I was both hurt and aggrieved at that time to have had my forecast traduced by a personal attack.

So, here is the clear evidence of what a party can do if it chooses a clearly unsatisfactory candidate 'in the eyes of a significant number of the electorate'. That was not my view, my hoped for result, my attack upon the candidate. It was me a fellow anorak opining a view of a possible outcome. To have chosen a very 'foreign', very southern, very urban, very grey suit banker, very thin and very young man for an expansive very rural, hunting fishing, beer drinking, agricultural, northern seat seemed barmy. And it was. In a rather good year the party suffered a considerable reverse, not because I willed it, but because I was just plain right about the choice. Parties can be silly and the Conservative party made some very duff choices in 2015. This was one of them. This is what happened!

Look at Hexham, another highly rural seat - a very poor Tory performance when the sitting MP retired in 2010, followed by a much much improved result in 2015.  Was the new candidate poor - no, it's just because incumbency is very important in highly rural seats. That combined with the leader effect.

Same will happen here - I bet this seat is in the biggest 20 increases in the Tory share in 2020.