Sharpe's Opinion

As someone who evangelises about the Internet and its game-changing potential in politics, I’ve been told several times that local parties don’t worry too much about their websites and internet strategies, because they know that leaflets work. Why invest in a website when your tried and tested methods of getting the word out still work perfectly fine?

Shooting At Straw Men Is Fun!

Costigan Quist this morning is confident about the effectiveness of leafleting:

The Tories put out a load of leaflets and their vote held up. The Lib Dems put out a load of leaflets and our vote held up. Labour and the Greens both put out fewer leaflets and their votes both fell sharply.

Really? Quite apart from the obvious criticism of that particular argument, how much of an effect did the leafleting really have on the result? What would the result have been, for example, if leaflets were 100% effective? If the electorate could be entirely predictably swayed from their initial voting intentions simply by receiving leaflets through their door?

For once, in the wake of a rather singular by-election, we can take a look. In the North Norwich by-election, only one opinion poll was taken, right after the by-election was called. Since then, a whole range of different parties used a whole range of different campaigning techniques to put their message across. So let’s see how effective they’ve been – out comes the spreadsheet application, lets crunch some numbers and make some pie!

Thanks to the on-the-ground reports from Nich Starling we also know how many leaflets each party was handing out – his final tally was as follows:

• Lib Dem 14
• Tory 13
• Labour 5
• UKIP 4

• Green 2

So if those leaflets were 100% effective, we might end up expecting to see something like this, right?

UPDATE: Just to be clear, the above equation is not mathematical, and shouldn’t be taken seriously. (That was sort of the point – see the comments thread for more.)

It’s particularly interesting, isn’t it, to look at the Lib Dems and the Greens. The Lib Dems, in the words of Iain Dale, ‘threw the kitchen sink at this campaign’, whilst the Greens did awfully little to avail themselves of the public. You can see in the above pie charts that the Lib Dems and Greens start off very close, and end up with a Lib Dem landslide. The Tories start strong and get stronger. Labour come a distant third.

Compared to the actual results of course, that projection is near enough laughable:

Labour ended up with exactly the predicted vote share, so hooray for Labour’s leafleting strategy. Nobody else did. The Lib Dems, in particular, put out nearly 40% of the leaflets and dropped a percentage point in the voting.

So is Costigan right? Well, it’s certainly true that the Conservative and Lib Dem votes ‘held up’, but can that really be attributed to the leaflets? I’m going to have to say NO.

The final results didn’t just fail to follow a ’100% effective leaflets’ projection, there’s barely any significant correlation at all between the number of leaflets handed out and the change in vote share from the opinion poll to the voting booth. Even if we say that UKIP were outliers, it’s still hardly a picture of the great effectiveness of leafleting. Forget the straw man – have a look at this scatter graph, which just drives home the the point:

I’m sorry, but you think leaflets are effective as campaigning tools? Are we really to believe that the Tory and UKIP leaflets were so much more effective that they made the difference? I just don’t see it.

That’s particularly true when you consider that the change more accurately reflects the opinion changes on a national level – the right are resurgent and the Lib Dems have been holding up their vote share, whilst Labour’s voting base has collapsed and ‘Other’ has been gaining steadily more and more votes. I just can’t see how anyone could look at this election result and say that leafleting has anything more than a pathetically negligible effect on voting habits.

Okay, but What Else Are We Supposed to Do?

I’m not the first to see this and I’m not the first to say it. Even just this morning, Irfan Ahmed is arguing the same point, although his ‘solution’ of trying emails instead of leaflets is pretty lacklustre. The truth is, online campaigning has also failed horribly at the Norwich North by-election. Take the Libertarian Party, for example, wildly over-represented in the blogosphere. While they don’t ‘campaign’ so much in the traditional sense of the term, they have well-read blogs and send out emails and use Twitter to get their point across, and look what happened to them – 36 votes.

The Freakonomics guys claim (H/T Charlotte) that the influence of campaigns rarely accounts for more than 1% of the election result. Decent campaigns can get you a short way in politics, but to go the distance what you really need is to have ideas that people want to vote for.

Advertising a movie can bring a lot of people into the cinema to watch it. An effective trailer can account for a huge amount of ticket sales, and a well recognised cast, writer, director or even producer can get bums on seats for the opening weekend. What sells films in the long term, though, is the reviews, and you only get good reviews by having a good film in the first place.

I can’t help but think it’s got to be the same with election campaigns. You can do all the advertising, get your message out, and get hundreds of leaflets through doors, but the vast majority of elections there’s only three things that are really going to make a difference to the result: the popularity of the incumbent, the popularity of your party (especially on a national level), and what extent the public agree with your ideas.

But, you know, I guess that’s why they call it democracy.