I’m traditionally a Labour voter, so I do have a bias towards

Comment on Weekly General Election Poll Results by David.

I’m traditionally a Labour voter, so I do have a bias towards them, but you can see from the poll results if I’m trying to manipulate my visitors to vote Labour I’m doing something wrong with the main poll!

I’m very anti-Tory, but you can see up until the TV debates the Conservatives was in the lead with around 31% of the vote. I could never have seen that coming and planned to manipulate the poll three weeks before the election to put the Lib Dems in the lead.

Don’t forget the polls got none of the above and floating voter which has tended to be around 6% or more of the vote. As I understand official polls, don’t knows are screened out. That 6% depending on where it goes on May 6th would account for the difference between my poll and official ones.

Then we’ll probably have some % difference caused by having UKIP, Greens, BNP and Plaid (surprisingly Plaid has remained quite stable, similar to the 2005 general election result and it would have been the easiest poll option to manipulate). If a small percentage of supporters from one of the small parties decided to manipulate the poll, they could throw it off relatively easily (see BNP info below).

I would say UKIP, Green and Plaid support is going to be similar to what we’ll see on May 6th (maybe my poll will have them a little higher than they should be). There’s been no evidence whatsoever of their party supporters trying to manipulate the poll.

The BNP’s vote however has been artificially increased because 1st the BNP support though relatively small are VERY, VERY vocal online. BNP supporters have been commenting on my site in bulk since September last year, while general public support didn’t kick in until the beginning of April (day the election was called).

In hindsight I wish I’d have done what all the pollsters do and just have the three main parties and Other for the rest. But then even though I knew my site would be popular by now (getting search engine traffic is what I do for a living), I didn’t anticipate 100,000 visitors a week. So I didn’t think about anything like this when I created the poll (at first it had 25 options).

Anyway, last week my site received around 1,500 visitors from the BNP official website, this week it was around 900 and I think it’s fair to assume the majority voted BNP in the poll since the pages the BNP supporters came from were told to come here and vote BNP. This traffic direct from the BNP official website pushed the BNP vote share to around 9% for the last two weeks. To balance the poll for this sort of manipulation I’ve removed 2/3rds of the votes that came from the traffic from the BNP official site (~1,000 the first week and ~600 this week) and still BNP support is higher than it should be. That still leaves quite a lot of votes that I could have rightly removed and for that reason I think the current BNPs vote share is significantly higher than we’ll see on May 6th (we’ll see on May 7th).

I’ve found no serious evidence of manipulation from any of the main parties supporters. There’s websites saying come here, vote all sorts of parties, but looking at my logs only the BNP have generated a fair amount of traffic from their official site (it’s a popular site) other sites it’s a drop in the ocean relative to traffic from search engines) over 100,000 visitors a week).

This means of over 100,000 visitors who visit the site a week the vast majority (well over 95%) have found their way here by typing things like General Election Poll, Election 2010 etc… in search engines like Google.

That throws the question open, is an average search engine visitor who finds their way to my site an average British voter?

I’d say no.

Younger people are going to be more likely to use the Internet than the old. So the site is probably not getting it’s fair share of OAPs.

I think the real value in this poll is tracking the weekly changes not the overall result. What’s very clear with regards the Lib Dem vote is it started low with the first 40,000 votes in the poll (up until 10th April) showing the Lib Dems on around 16% of the vote (similar to official polls).

There was also very little interest for the Lib Dems in the comments, you’d get the odd comment about them, but not a lot (now there’s quite a bit).

After the first live TV debate Lib Dem support sky rocketed and this week has remained high at 35% despite some of the Nick Clegg hype dying down.

If this support stays high and it’s only going to change if something really anti Clegg happens I expect to see strong support for the Lib Dems on May 6th (I would not be surprised if they gained over 30% of the popular vote).

Although I own this site I’m no different to any of you. I’ve looked at the Lib Dems as a wasted vote, a party I’d have considered voting for IF they could win, but they wouldn’t win, so why waste my vote. Things changed after that 1st live TV debate, many people believe the Lib Dems can win (highly unlikely if you look at what they’d have to do, but the belief is there). This cancels out mine and other negative impression that a vote for the Lib Dems is a wasted vote, it might be a long shot, but the Lib Dems could win now, many might argue it’s less of a long shot than Labour winning a majority again.

I don’t know if the official polls are manipulated in the favor of a particular party, but what is interesting is after the 1st live TV debate a poll was released saying something like 33% plan to vote Lib Dem before the poll had been manipulated to take into account the 1,000-2,000 people they are polling aren’t a cross section of British voters. After the manipulation the Lib Dem support dropped to I think 26% (I forget the exact %s)!!!

You could make the argument, if they manipulate a poll that much they aren’t doing a particularly good job of finding a representative sample of British voters and as such the official polls are not very accurate! It should be a case they’d selected the people to poll so they do get a representative sample of voters and if they do there’s no reason to change the numbers.

The only number that’s been changed on my poll is the BNP vote share because of obvious manipulation.

Shorter answer is I’m not sure :-)

David