A new poll and a new Commons seat calculator

A new poll and a new Commons seat calculator

Today’s Populus Poll would probably carry a lot more weight if it was not so out of line with what the same pollster found just eight days ago. Within that short period:- Labour has jumped 7% from third to first place; the Tories are down 4% from first to second place ; and the Lib Dems are down from second to third place and are also down 4%

The figures are CON 28%(-4), LAB 35%(+7), LD 25%(-4)

For a good analysis on the latest numbers we commend the polling expert Anthony Wells

Meanwhile a regular user of the site, Andy Cooke, has produced a new House of Commons seat calculator which takes a totally different approach to the process of converting polls numbers to likely seats. This is a very welcome addition and no doubt will spark off a lot of discussion. The calculator can be down-loaded by clicking here or the first link on the top right of the screen.

    A main feature is that the electoral system is less biased against the Tories than has been suggested and that Michael Howard could secure a Commons majority with less than 37% of the vote – a share that is much lower than other calculations.

This is Andy’s introduction:-

The First Past the Post (FPTP) geographical constituency political system is the system used for General Elections in the UK. Under this system, it is very difficult to translate shares of the popular vote into seats in the House of Commons and thus the viability of a party’s chances to form a overnment

There are many ways to attempt to calculate this. The most common is that exemplified by the two most popular Internet-based seat calculators: The Martin Baxter Seat Calculator, and the Anthony Wells Seat Calculator. Both attempt to start from a known baseline (the 2001 General Election), apply a Uniform National Swing to the vote shares, adjust the number of votes received by each party accordingly. The winner in each constituency is then calculated and the party totals summed.

The drawback with this “Bottom up” system is that, firstly, the Uniform National Swing does not actually exist. There are inevitably large regional ariations in swing,even from one constituency to its neighbours. Secondly, the system has built in any and all biases that existed at the previous election. Such biases may be the ease or difficulty of winning seats for a third national party, the prevalence and effects of Tactical Voting against a given party, the popularity or otherwise of individual MPs who may or may not be standing at the following election, the inevitable change in bias provided by updates to constituency boundaries and numbers generated by reviews by the Boundary Commission, etc.

This is an attempt to provide a crude “top-down” seat allocation calculator. We recognise that there are different political “eras”, under which the seatallocation biases change noticeably. To create this calculator, the 16 General Elections in the UK since the Second World War were taken, and an attempt made to identify common eras among them, under which seat allocation appeared to follow consistent rules. Elections within those eras were then taken togetherand used to interpolate and extrapolate trends. A large measure of subjective estimation (ie guessing) was used for this.
The derived look-up charts were then made from a best-fit estimate.

The Eras fell very consistently within periods defined by major Boundary Commission changes of seat allocation, with the exception of “Era 5”. The overwhelming distortion within the period was due to the extensive and thorough Anti-Tory tactical voting. This era will cause a “hangover” into the next era, even with the removal of the Tactical Voting itself.

My main comment is that it’s quite complicated and underplays the Lib Dems ability to win seats through heavy targetting. Hopefully Andy will be able to join in the discussion.

picture http://www.tribuneindia.com/2001/20010923/w2.jpg

Comments are closed.