This is only the third update of this blog's Poll of Polls for the EU referendum - I haven't been able to produce new figures as often as I expected, partly because telephone polling has been surprisingly infrequent. The "Leave" camp has now practically drawn level on the average for online polls, and although the "Remain" lead on the telephone average is still hefty, it's been slashed by 5% since mid-December. That may simply be a reversion to the mean, because the telephone lead was actually slightly smaller in the first update back in October. Even so, the 50/50 average (ie. the mid-point between the telephone and online averages) shows a Remain lead of less than 10% for the first time.
SCOT GOES POP POLL OF POLLS
Should the United Kingdom remain a member of the European Union or leave the European Union?
50/50 ONLINE/TELEPHONE AVERAGE :
Remain 48.4% (-1.7)
Leave 39.0% (+1.4)
ONLINE AVERAGE :
Remain 42.3% (-0.8)
Leave 41.9% (+0.2)
TELEPHONE AVERAGE :
Remain 54.5% (-2.5)
Leave 36.0% (+2.5)
(The online average is based on seven polls - four from ICM, one from YouGov, one from Survation and one from Panelbase. The telephone average is based on one poll from ComRes and one from Ipsos-Mori.)
As we've discussed before, the two most plausible explanations for the enormous gulf between telephone and online results are that volunteer online polling panels contain a disproportionate number of politically committed people (which would imply that telephone polls are most accurate), and that some people may be too embarrassed to tell a telephone interviewer that they want to leave the EU (which would imply that online polls are most accurate). It could be a bit of both, of course, in which case the 50/50 average might be the best guide.
Just to muddy the waters further, Anthony Wells has pointed out that telephone polls are producing far fewer Don't Knows than online polls. That's an entirely counter-intuitive finding, because if online samples really do contain more politically committed people, you'd expect those individuals to know their own minds. Could it be that online respondents are thinking about the question more carefully and giving a considered response, while many telephone respondents are just giving a very casual answer? If so, we should be cautious about assuming that telephone polls are more likely to be on the money, because sooner or later the disinterested telephone respondents will start considering the issue in more depth, and might just reach the same conclusions as their online counterparts. Some would say that was precisely the explanation for the sudden convergence in the polls at the end of the independence referendum campaign, when the telephone and face-to-face polls showed a seismic shift, but the online polls barely budged. (The big exception was YouGov, an online-only firm that showed a massive swing to Yes - but that might be partly explained by their convoluted and secretive "Kellner Correction", which artificially generated No-friendly results earlier in the campaign.)
* * *
As you know, the uncertainty has returned in respect of present-day polling on independence - there have been just two 'real world' polls (ie. telephone or face-to-face) on the topic since the referendum, and both have shown significant Yes leads. All of the other polls have used volunteer online polling panels, and with the odd exception have generally shown narrow No leads. But our occasional commenter "Roger Mexico" confused the issue slightly on the previous thread by claiming there was a third 'real world' poll out there - an ICM telephone poll showing a No lead. Roger has a fairly obvious agenda, but he's far more knowledgeable about polling than the likes of Aldo, so I do start to question my own sanity when he says things like that. However, I've double-checked and triple-checked - I can only find one ICM poll since the referendum, it was commissioned by the Guardian, and they clearly stated that it was conducted online. So I'm pretty sure Roger has got this one wrong, and that (as things stand) the divide between real world and online pollsters is extremely clear-cut.
* * *
It's not often that I disagree with the editorial stance of The National, but I think they've got it wrong in their endorsement for a campaign that would criminalise the likes of the "pick-up artist" Roosh V for incitement of hatred against women. I've read the blogpost in which he argues that rape should be legalised on private property, and as idiotic, outrageous and offensive as it is, it doesn't actually contain any incitement to rape - in fact the opening part of the post contains a long explanation that he thinks rape is morally wrong. He's essentially arguing that the onus for reducing the incidence of rape should be entirely placed on women, rather than on men or on the law - and he claims to honestly believe that would be far more effective. You might think that his argument is a complete sham (it probably is) and that he really is as "pro-rape" as the newspapers have dubbed him. But in a society that is supposed to treasure the right to free speech, it's dangerous enough to throw a person into jail for his actual words. To incarcerate someone because of our contested interpretation of the hidden meaning of his words...well, that's crossing a threshold that we really shouldn't ever cross.