Mr Quist has a thoughtful post up about the Conservative and Labour approaches to internet campaigning – thoughtful as much for his assessment of the Guardian’s “analysis” as anything else.

He concludes thusly:

It can be done, though. You can have excellent local websites in those constituencies. You can use the Internet as an additional tool to engage people locally. You can knock on people’s doors, get their email addresses and mobile phone numbers and then communicate with them quickly and easily. You can use local Facebook groups to reach a younger bunch of people.

That’s almost certainly the stuff that really works. It isn’t massively expensive, but it’s hard work and – whisper it – actually involves meeting people face to face and not just coming up with something really cool in your bedroom or office.

Broadly speaking, I’m in agreement with the tone of this. I’m a cynical friend of technology, and I think we (and probably the other parties) devote too much energy to doing web 2.0 stuff “because we can” rather than because there’s any evidence it works. Experimentation with no hope of return is the stuff of progress, of course, but that doesn’t mean that every cost-benefit analysis has to be thrown out of the window.

If only thirty-two people are joining every Facebook group you set up, and it’s the same thirty-two people, and one of them’s your little sister, it’s probably time to start saving even the slight energies required to set up Facebook groups. It just leaves bits of tat all over the internet, and enough tat, I reckon, can actually have an adverse effect on your party’s overall image. I get annoyed by the thousands of Facebook invitations to join groups with the format “So-and-so for Such-and-such a place” and I’m supposed to be a Lib Dem. What these people’s mates think of all this virtual clutter is anybody’s guess. If there was a virtual recycling box on Facebook, most political group invitations would go straight into it in much the same way most leaflets go into meatspace recycling boxes.

But I digress. In its essentials, Mr Q’s call is a familiar one. One hears it, with a greater or lesser degree of exasperation attached, from everyone who regularly hits the doorsteps. Less of this mucking about in your bedroom creating cool stuff, and more actual wear and tear on your shoes. Now, I’m a great believer in the sort of bone-deep expertise long-term campaigners have garnered. Some people can call elections correctly to within ten votes, and it’s probably not a bad idea to listen to them. Furthermore, it just makes sense to me that human contact is a better investment of time than online contact. The kind of balance between shoe-leather and modem that Costigan is advocating makes sense.

It’s in good company too. It finds support in something Karin Robinson, of Obama’s European campaigning team, said at the Lib Dem Voice fringe meeting in Harrogate. She said that they were very careful not to make the websites too good. They had to leave the user wanting something more, something s/he had to go out and knock on doors or join in social events to get.

So yeah, of course it all makes sense. But as any student of statistics knows far better than me, lots of things “make sense” and even “stand to reason” which nonetheless turn out to be totally untrue when you plot them on a graph. This is especially so when there might be variables at work you know nothing about – and what could be more variable than political campaigning? People’s political choices are affected by inputs as far apart as twenty seconds of a politician on TV, a duff bit of local officialdom (whether or not it actually relates to a political group) and the weather on polling day. And yes, even stuff they’ve read on the internet (which I reckon is far more likely to be informal opinion than official party bumf, but again, that’s only what makes sense to me).

We all know this, and yet from talking to some campaigners, you’d think certain techniques have the proven status of hard science. Nowhere is this clearer than in talking to advocates of negative campaigning. I’ve lost count of the number of times I’ve had it asserted to me that “negative campaigning works” with no attempt at back-up whatsoever. Works in comparison to what? The same election taking place in a different dimension on the other side of a wormhole with positive campaigning going on in it? You can’t institute a monopoly and then announce that it works perfectly.

There’s no winning with these people, either. If we win an election where their principles were put into play, it must be because of the negative campaigning. If we lose, it’s because we didn’t do enough negative campaigning. Heads I win, tails you lose. It’s so bananas that I am starting to nurse the horrid suspicion that many of the people who assert that negative campaigning works probably just like doing it.

It’s assertions like this that need, desperately need, to be tested as scientifically as possible against alternatives. And yes, doorstep campaigning does need to be tested against internet campaigning in order for the synthesis between them to be better understood. We can’t afford to spare people’s feelings from either camp. I mentioned cost-benefit analysis. No-one, to my knowledge, has ever come up with a set of roughly reliable equations that states any of the following:

  • x numbers of doors knocked on in month prior to election = x number of votes
  • x numbers of leaflets in month prior to election = x number of votes
  • x numbers of garden boards in month prior to election = x number of votes
  • some combination of the above that reflects their impact upon each other, beyond my tiny mathematical ken
  • all the above repeated, but for terms of three months, six months and a year prior to the election

Why don’t we have a national picture of this knowledge? Partly, it’s difficult to garner, certainly for national elections, because it doesn’t take account of ward targetting. The above equations for one particular ward in a seat may render vastly different answers from the same equations for another ward – but the results at the ballot box won’t be attributable to any particular ward. In local elections, of course, it is possible to separate out results by ward, and see what effect your activity has had there. But people vote differently in local elections.

Perhaps a better way of putting it would be, why aren’t we even trying to acquire this knowledge? Why aren’t we doing more with all the canvassing data and leafleting data local parties undoubtedly collect? Because until we do, and until (harder still) we develop equivalent equations which translate online campaigning activity into votes, then frankly the assertion that on-the-ground campaigning is more effective than internet campaigning remains just that, an assertion. As does the eminently sensible-sounding assertion that what we need is a balance between the two.

We need the beginnings of those equations. Yes, campaigning is an art, and there’s still room for endless art on top of them, but the number of significant campaigning decisions in front of us now are such that we simply can’t afford to charge off down the wrong path. If we do that, we’ll have no way of recognising the mistake we’ve made.  We won’t have anything to compare our performance against. We’ll stay stuck in the “Heads I win, tails you lose” pattern of proof-by-assertion that characterises all discussions on this subject at the moment.

We need a set of proofs, a set of principles, as scientific as possible, and applicable from region to ward-level, to be the foundation of the post-Rennard campaigning era. Assertion won’t cut it any more.