Three Election Observations

As someone interested in politics, who has studied it, and who somehow has ended up working in the Westminster Village, a few observations based on gut feelings and random thoughts.

1. 'It's the economy stupid'
The overwhelming factor in voter behaviour is who people think will make them better off. It's not as simple as straight cash and tax policy, but includes things like education, health and pensions - so it's a bit more nuanced than people think. However, all depend on the underlying economy and whilst I doubt many people appreciate just how dicey the economy is at the macro level, it will still be the largest deciding voting factor. The reason there is no clear party in the lead is that all have pretty similar economic policies - the differences are minor - hence voter uncertainty.

2. It's not the 'first internet election'
Labour may have lost one PPC due to poor tweeting, but the internet won't be a massive factor in this election. Compared to the US Presidential election the parties are raising very little cash via the internet and most of their publicity is still via the mainstream media and direct mail. One or two political blogs will be worth following - and may break a massive story - but it will largely still be a fairly traditional election campaign.

3. Opinion polls are interesting but ultimately useless
It's a cliche that the only poll that matters is the one on election day, but though opinion polls are mildly diverting, they don't actually tell us a huge amount. The following are some of the reasons;

• margin of error - this is usually +/- 2 or 3% - this can make quite a difference, esp. between an overall result and a hung parliament
• there are only a snapshot - it is how a person would vote at the time they asked the question. Polls, on the whole, don't measure depth or strength of support; ask the same person an hour later and the result could be different. A lot can change in a campaign - and voters are fickle
• the 'don't knows' - notice how most polls don't publish the percentage of 'don't knows'/'haven't decideds'? In the past I've been polled and you often don't get this as option. My hunch is that this election there are huge number who are don't knows - to be useful we need polls showing this
• won't vote - who know what the turnout will be, and more importantly how it will be distributed? This is as important as the 'don't knows'

Look at polls, but take the above in to account - polls are interested but are flawed.

Some predictions to follow...