Showing posts with label Polls. Show all posts
Showing posts with label Polls. Show all posts

Monday, 12 April 2010

Two New Graphs...



These new graphs go back to 1st of March, and show the Labour and Conservative shares, together with the all-important lead. Error bars are also shown, indicating the 95% confidence intervals.

We can see that the lead was narrowest on the 5th March and 24th March, and highest on the 4th April. Since then, the lead appears to have narrowed...

The yo-yo effect on the Labour estimate is due to the wide variation in different pollsters' estimates, particularly as Harris and Opinium enter and exit the field. The pollsters are far more uniform in their estimates of the Tories' share.

Update: all three parties are now displayed on the same graph. At right there's an image illustrating how many seats the LibDems might win for various shares of the vote. It assumes whatever the LD vote share the Tories maintain a roughly five point lead over Labour.

Reblog this post [with Zemanta]

Sunday, 17 January 2010

I, the undermentioned, hereby give notice that...

Your Current Forecast



Why not play along with our seat-by-seat forecast? It works like this...

Friday, 15 January 2010

Introducing a New Graph...


Following on from the PB article in 2006, I have now refined the Battleground Graph to show the best visual estimate of election outcomes.

The differences are small but can be summarised as follows.

Sunday, 29 November 2009

The Hung Parliament that never was... [at least, not until later]


Can't see video? click here

Us old hands will remember the 1992 election, when almost all the polls pointed to a hung parliament, and even the network exit polls confirmed it. As the night wore on, the Tory projection increased, and the Conservatives were ultimately returned with an overall majority of 21, which, as it happened, proved insufficient to last through the parliament.

1992 is remembered now for being the "Waterloo of the Polls", and less for its real significance - the emergence of a strong anti-Tory bias in the electoral system, which has increased markedly since that date.

John Major's majority rested on just 1241 votes across 11 constituencies. Another 0.5% swing, and Labour probably would have forced a hung parliament.

Although virtually uncommented upon at the time, the electoral system had shifted dramatically in Labour's favour, and continued moving in the same direction at the two subsequent elections.
Reblog this post [with Zemanta]