h1

Rallings and Thrasher on the VIPA seat predictor

October 4th, 2009

****For discussion on the Greek General Election Click here****
****This thread for UK Politics only****

What do this pair think of the new system?

In view of the heated discussion on the previous thread I thought I would publish comments by Professor Colin Rallings and Professor Michael Thrasher of the University of Plymouth and Co-Directors of the Local Government Chronicle Elections Centre.

They are also the academics who produced the Press Association guide to the results of the last election which was based on old boundaries. Their detailed figures will form the basis for much of the coverage of the next general election on all channels and in all the papers that carry results.

They are the nearest thing to “official psephologists” that there is.

When VIPA, which was developed by my son, Robert, came out they asked for evaluations copies and have since come back with their observations.

This was what they wrote: “We’ve had a look at this now. Seems interesting and well put together. Its weakness is of course the reliance on polls, especially being tied to ICM (who do admittedly have good record) and the problem of the reliability of people’s recall. But the essential premise – that votes are likely to disaggregate in different ways in different types of constituency- is well taken.”

Their point about the past vote recall data is clearly valid because there is an issue with people misremembering what they did. To deal with this the pollsters have different approaches and this is what is used in VIPA.

What is note-worthy about the Rallings and Thrasher response is that there is no mention of proportional swings – which has been the main criticism by one commenter on this site and Peter Kellner of YouGov.

Please can the discussion take place without personal abuse.

Mike Smithson