Iwantareferendum Results
They aren't representative polls in a strict sense, but I though the results of the private referendums organised by iwantareferendum.com were worth a comment.
Private referendums are rare creatures, not least because they are very expensive to carry out (the only really notable one prior to this was Brian Souter's 2000 private referendum on Section 2a in Scotland). In practice they are a campaigning technique for groups, being able to say that hundreds of thousands of people have voted to support you gives a group a certain moral argument. For those of us interested in what they say about public opinion however, their problem is that they are neither fish nor foul - they are not a representative opinion poll, but neither are they the equivalent of an actual referendum.
Private referendums certainly shouldn't have a problem with inviting a biased or skewed group of people - they invite everyone in the country or constituency where they are holding the referendum (though in the case of iwantareferendum ballots could only go to people who chose to make their details available on the edited electoral register).
The problem is that the people who chose to return their ballot papers may not reflect the opinions of those who didn't. The turnout for the iwantarefendumcampaign was surprisingly good - on average 36.2%, higher than many actual local elections - but we still don't know what the other 63.8% of people think (actually, because not everyone was on the edited register, the proportion of people who actually took part is lower than that). What if the people who voted were mostly young, or mostly Tory, or richer, or skewed in some other way?
The risk with private referendums is that they are invariably organised by, and for the benefit of, people on one side of the argument. Brian Souter was in favour of keeping Section 2a, so it's easy to imagine that people who didn't support the legislation not wanting any part of his referendum. Equally, it doesn't take a rocket scientist to work out that iwantarefendum.com are in favour of a referendum, so there was the chance that only those who agreed with them would take part. This potential skew is made worse when one side of the argument deliberately tries to get their supporters not to participate - in several constituencies including East Renfrewshire and Eastleigh the incumbent party advised supporters to boycott the referendum.
In practice Souter's referendum did suffer in this way - he found 86.2% in favour of keeping Section 2a, when opinion polls at the time showed only 60% in favour. Iwantareferendum seem to have suffered this sort of skew to a far lesser extent, they found 87.9% in favour of a referendum. This is higher than in representative opinion polls, but not by very much. Looking at recent polls that asked if people wanted a referendum and repercentaging to exclude don't knows we get figures like 75%, 83%, 76% and 86% in a Populus poll for the BBC back in October.
This does underline the problem with private referendums as a way of judging public opinion. I can say it looks like it did get a good reflection of public opinion on the question of the referendum because it's in the same sort of region as the polls... but if it hadn't, I'd have been saying that those participating where obviously horribly skewed. It's only because we have the polls to back it up we know what to make of it.
Of course, in a real referendum or a real election the fact that people who don't vote might disagree is irrelevant: decisions are made by those that turn up. Unfortunately private referendums aren't a proxy for real ones for several reasons. Firstly they have no consequences - one may vote diffrently when one's decision actually matters. Probably more important is the fact that real referendums or elections have campaigns - iwantareferendum sent out pro-and-anti arguments in leaflets with ballot papers, but the campaign beyond that was one sided. Conservative PPCs and MPs were campaigning to get people to vote, while in many seats Labour and Lib Dem incumbents were encouraging people not to participate in the ballot. In a real referendum, both sides would have been encouraging their supporters to vote, with whatever eventual impact on the result.
Not a representative poll and not a real referendum. What do private referendums tell us then? Well, in this case not much more than that 133,251 people in 10 marginal seats would like a referendum, have voted to express that opinion and have likely been told their current MP is ignoring them (though 2 of the MPs have since said they will vote in favour of a referendum). For those local areas and the eventual election results there that is no small thing.
