Message from Robert: Well that was rubbish. We were trying to load-balance between the servers using perlbal. And perlbal performed so badly the load-balancer itself was unreachable. I’m going to try using nginx for load-balancing now. Hopefully there won’t be too many ill effects. If anyone has any experience implementing software load balancing on Linux, please email me: rcs1000 at g mail dot com
|Populus Apr 7 (Feb 7)||CON 2010||LAB 2010||LD 2010|
|Voted CON in 2005||90% (nc)||1% (nc)||4% (+1)|
|Voted LAB in 2005||14% (nc)||73% (+3)||10% (nc)|
|Voted LD in 2005||14% (-3)||8 (-2)||76% (+6)|
The above data is from the latest Populus poll with comparisons on the last published national survey from the firm two months ago.
One reason it was good to see Populus again is that it and ICM are the only two firms from the 2005 election that are operating in the same broad way now as they were then.
All the others – including YouGov, Ipsos-MORI and ComRes – have made significant methodological changes or else are new to regular UK political polling.
Populus also provide their data in a form that make it possible to track the churn and retention rates of sample members from the last election so that we are able to examine what can be good pointers.
In the table above we see that the Tories are still enjoying a significant retention and churn lead amongst 2005 voters although the gap has narrowed. This is because more Labour and Lib Dem voters from 2005 are now back with their party. But a significant part of the Tory current support comes from supporters of the other parties who have churned.
A big question is whether the three parties, on these measures, will get closer and whether Labour, in particular can win back much of the support that has gone the blue way?