Thursday, April 07, 2005

Poll Dancing Turns Me On

I am really getting fascinated by opinion polls. There's an excellent blog dedicated to it, the UK Polling Report, and also an E-Politix blog from my hero, Bob Worcester.

There seem to be some fascinating intricacies that are worth noting, but the biggest point is a simple one: only compare polls from the same companies. One thing that really annoys me, though, and I believe I am right to query, is the use of national vote percentages. Under first-past-the-post, or any constituency-based system, the national swing is a rather useless forecast for the country's feelings. Constituencies are so often affected by local issues, on-the-ground campaigning and a whole host of intricacies, that the idea of such percentages, and the swingometer theory on which they are based, is fairly flawed. A party's decision to target a seat effectively, or the existance of a particular type of swing voter will be crucial. For example, moslem and student populations are going to be particularly volatile in 2005 given Iraq (in both cases) and tuition fees (for the students).

I suspect our pollsters will get the percentages correct, but when they make seat predictions, I think this year may require a particular liberality in applying a uniform national swing.

Me? As a result of my involvement in the educational simulation game PoliticsUk, I'm getting the pollsters' software from UK Elect in the next few days, and should be able to play with that to my little heart's content.