We project Conservatives to be 100 seats ahead of Labour


2:38 pm - April 30th 2010

by Guest    


      Share on Tumblr

contribution by Nate Silver, Renard Sexton & Dan Berman

The Uniform Swing Calculator has its proponents, and it has the virtue of being fairly easy to calculate seats.

But, most recent elections in the UK have not been all that dramatic, with fairly minor shifts in the vote between Labour and Conservatives.

We’ve designed an alternative approach which, while also based on fairly simple assumptions, is potentially more robust.

With some new improvements to our UK General Election forecasting model, we project that this election bodes quite bad news for the Labour party.

We now show them holding on to only about 200 seats in the House of Commons, versus roughly 300 for the Conseratives and 120 for the Liberal Democrats.


As before, the basic premise of the model is to explicitly shift votes from one party to another — for instance, take a percentage of Labour’s votes and give them to Liberal Democrats — and then apply these vote shifts in a sophisticated way across the UK’s 650 constituencies.

The vote estimates included herein are not based on any hard-and-fast rules; rather, they are educated guesses based on an examination of cross-tabular results in recent several UK polls, with the stipulation that the overall vote share estimates must correspond to a rough average of recent polls.

Rather, the “value-add” of the model is in the way that it takes these vote share estimates and translates them into estimates in the number of seats that will be controlled by each party.

Several improvements have been made to the model since the original version that we posted on Sunday:

— It now allows voters to move into and out of the electorate, rather than simply from one party to another.
— It now has a more advanced process for assigning votes across individual constituencies, producing effects that to some degree resemble “tactical voting”.
— It now refines its results based on regional-level polling data — this is particularly important for Scotland.
— It now accounts for incumbent retirements.
— It now accounts for Scottish and Welsh regional/nationalist parties, in addition to those from Northern Ireland.

These will be described in more detail in subsequent posts.

Although we are proud of this product and believe that it will more likely than not it will me more accurate than simple uniform swing calculations (which show substantially more optimistic seat estimates for Labour), it is experimental and should be regarded as such.

———–
All three write for the US-based politics site FiveThirtyEight.com, where it was cross-posted from.

Renard Sexton has written more on this today on Guardian CIF. Renard Sexton is FiveThirtyEight’s international affairs columnist and is based in Geneva, Switzerland. He can be contacted at sexton538@gmail.com

    Share on Tumblr   submit to reddit  


About the author
This is a guest post.
· Other posts by


Story Filed Under: Blog ,Elections2010

Sorry, the comment form is closed at this time.


Reader comments


120 for the Liberal Democrats seems rather high – where are they going to win all these seats? Presumably on these figures mainly from Labour (whose seats have to go somewhere), but whilst I could see them picking up about 30 Labour seats and even a few Conservative held ones, I can see Conservatives gaining some seats where Labour majorities are split by Liberal Democrat resurgence (proving not only they can win there?). Also, there is some suggestion that Liberal Democrat held marginals may still be vulnerable to Conservatives in some places, so I think the model is overestimating Liberal Democrat seat count.

2. Flowerpower

Interesting approach.

But you’re still allowing Labour 27% of the vote. Probably a bit on the high side.

“120 for the Liberal Democrats seems rather high – where are they going to win all these seats?”

Regionally. Look at cruder data than this and Lib Dems are currently on course to take 100 seats, with a further (upto) 30 being extremely close races that could go either way. Throw in a few odd constituencies that seem to be uncharacteristically favouring Lib Dems nationally and 120 is realistic based on the current data.

However if Tories do retain a mid-high 30’s then it won’t be that 120, but we need to get a consistent trend on Tory numbers which we’ve not had for over a week.

They’re doing an average of the polls, assuming it is correct. I see no reason why it should fall below 27% to be honest. That is actually lower than the percentage of ppl who are sympathetic to Labour, but as I say Brown has pushed many away.

The reason why Libdems are projected to do so well is because they’ve taken a significant percentage of previously (2005) Labour voters, if I’m reading correctly) which gains them seats.

There can only be one reason: this Labour party turning right-wards on many issues.

Well, it’s pretty-looking, but it doesn’t seem really practical to make it work with polling because you need that breakdown of voters. Multiple polls across that model with accuracy could be tricky, particularly given not all polls actually breakdown their voters like that and the huge margin of error on said calculations.

A very simple system that works well is the ‘ratio’ system, where you just change the quantity of votes in a constituency based on the party vote last time compared to this time. So, for example, if the votes had shifted like this compared to last time:

Party 1: 40% -> 35%
Party 2: 30% -> 50%
Party 3: 30% -> 15%

then constituencies would be calculated like this:

Party 1: 10,000 * (35/40) = 8,750
Party 2: 20,000 * (50/30) = 33,333
Party 3: 70,000 * (15/30) = 35,000

Mercifully, the UK Elect software actually has a ratio option already for its calculations so there’s no need to build a new system. The methods aren’t quite as fancy as Nate’s, but it’s far more practical IMO at least.

If you’re wondering, BTW, UK Elect on ratio settings gives nearly identical results to Nate’s model: Conservatives on 296, Labour on 198, Liberal Democrats on 119.

“Also, there is some suggestion that Liberal Democrat held marginals may still be vulnerable to Conservatives in some places”

This is extremely unlikely as it stands, Tories simply aren’t performing well in areas that Lib Dems hold marginal seats (like the South West), in fact they’re performing worse than 2005 at last count.

7. Gaf the Horse

This is what I’ve been scared of all along. I’m a LD activist and I’ve enjoyed leafleting all the more recently with the thought that we might actually hold the balance of power after the election. But I’ve always had at the back of my mind the nagging doubt that the Tories might actually be doing better than they seem. I suppose on figures like that it might be possible for Labour and the LDs to get together with someone and form a coalition with a tiny minority, but its going to be very difficult to do and the Tories are in poll (sic) position really.

Lets hope that the model is completely wrong 🙂

Am I the only one thinking “I don’t really care what the numbers are, as long as it results in neither Labour or the Tories having a majority” ?

Altho’ come to think of it… the further either of them are away from a majority the better!

“120 for the Liberal Democrats seems rather high – where are they going to win all these seats? ”
In my model at least, after they’ve tapped the South West by outgrowing the Tories the inner cities are next – three seats in Edinburgh, three in Newcastle, three in Birmingham, a bunch in north London. I could provide a map and/or list if you’d like.

so labour and libdem together have more seats than the tories(after labour have been in power for 13 years)? doesn’t sound great news for dave.

i agree galen

@ 9

Be interesting to see. Marginal information I’ve seen (trying to place where) has Liberal Democrats possibly losing several seats across eastern Wessex area, although south-west and London seats seem safer.

Sunny,

“There can only be one reason: this Labour party turning right-wards on many issues.”

Brilliant prediction. People leave Labour to support the Nick Clegg-led Liberal Democrats because Labour are not to the left enough?

I think you are too stuck on the left-right scale. Voters may actually not self-identify that way (as my wife keeps reminding me), but rather select a party they like for a number of reasons. Don’t confuse what is essentially a one-dimensional model of politics with reality.

People leave Labour to support the Nick Clegg-led Liberal Democrats because Labour are not to the left enough?

Yes, have you seen Libdem policies? Especially with younger voters

Yes, have you seen Libdem policies?

If you have, it will be more than the majority of Lib Dem voters have.

@12

Alright. Bear in mind here that I’m not accounting for any regional or local changes here (even for Scotland, which IIRC isn’t seeing the same LD surge as the rest of the country), and this is just a very quick mock-up of the percentages presented in the post (i.e. load my program up and bang in the figures). Here are the extra 59 seats that they’d win, coloured by the party that hold them notionally:

http://img153.imageshack.us/img153/3772/ldgains.png

TimJ: You might well be statistically right, but manifestos have outsold Harry Potter these last two weeks. Bet that’s not happened before.

“If you have, it will be more than the majority of Lib Dem voters have.”

Typical snide remark from the tory troll.

Why aren’t the tories winning this election by miles seeing as (according to them) Britain is Broken, broke and the tory press is giving Dave the biggest free ride of all time? Add to that all the money they have thrown at this election and really the tories should be on 50% of the vote.

All those advantages and so far the can just about get mid 30s .

Interesting. Similar results here: http://bit.ly/dzHmFK

20. East Londoner

This analysis, like any analysis based on simple statistics is flawed because it does not take into account the regional and turnout variations across the country nor the statistical fluke that means that Labour support is concentrated in a few places. Very crudely it matters little if the tories get huge numbers of votes in Surrey and Labour very few because the tories already hold all those seats, though it does push up their national percentage and depresses Labour’s (a fairly detailed explanation of this written by Peter Kelner can be found here http://today.yougov.co.uk/commentaries/peter-kellner/uphill-battle-conservatives). The current electoral system is broken but that is the one that will be used next Thursday, it will almost certainly deliver a result in terms of seats that bears little resemblance to the national percentage share. Hopefully this will be the last time this will happen.

Hm. But does it account for tactical voting?

I suspect Tory support v vote is underestimated, LD overestimated. Yet….

“Hm. But does it account for tactical voting?”
No model is likely to in this election because tactical voting will probably be limited and not in one coherent direction (for/against one party).

@ Watchmen

“Sunny,

“There can only be one reason: this Labour party turning right-wards on many issues.”

Brilliant prediction. People leave Labour to support the Nick Clegg-led Liberal Democrats because Labour are not to the left enough?

I think you are too stuck on the left-right scale. Voters may actually not self-identify that way (as my wife keeps reminding me), but rather select a party they like for a number of reasons. Don’t confuse what is essentially a one-dimensional model of politics with reality.”

Who have you spoken with?

I was hardcore left-Labour, Socialist for decades, just in my 20’s but you get the point-and my friends were as well.

I’m now nd have been for now a year and abit voting Lib Dems and so are my friends.

Why? Because we bothered to actually read their policies.


Reactions: Twitter, blogs
  1. Christian DeFeo

    RT @libcon: We project Conservatives to be 100 seats ahead of Labour http://bit.ly/dcd8iY

  2. sdv_duras

    RT @libcon: We project Conservatives to be 100 seats ahead of Labour http://bit.ly/dcd8iY

  3. andrew dalby

    These are reliable pollsters RT @libcon: We project Conservatives to be 100 seats ahead of Labour http://bit.ly/bBadJl #ge2010

  4. johnhalton

    is beginning not to like Nate Silver any more. Another depressingly convincing analysis of Labour's prospects: http://bit.ly/923SfW

  5. Stephen Duffy

    Gin & Tonic anyone? RT @libcon: We project Conservatives to be 100 seats ahead of Labour http://bit.ly/dcd8iY

  6. andrew

    Liberal Conspiracy » We project Conservatives to be 100 seats …: This post was mentioned on Twitter by Liberal C… http://bit.ly/9onAg6

  7. Liberal Conspiracy

    We project Conservatives to be 100 seats ahead of Labour http://bit.ly/dcd8iY

  8. Tweets that mention Liberal Conspiracy » We project Conservatives to be 100 seats ahead of Labour -- Topsy.com

    […] This post was mentioned on Twitter by Liberal Conspiracy, johnhalton, andrew dalby, sdv_duras, Christian DeFeo and others. Christian DeFeo said: RT @libcon: We project Conservatives to be 100 seats ahead of Labour http://bit.ly/dcd8iY […]

  9. Arun Sundar C

    "We project Conservatives to be 100 seats ahead of Labour" and related posts http://ow.ly/17eq8A





Sorry, the comment form is closed at this time.