When politicians are asked about unflattering polling figures it’s a tricky question to answer – what can you say? “The poll shows no one likes me and I’m going to lose”? Most politicans resort to the time honoured cliches of there being lots to play for or “the only poll that counts is on polling day”. The latter grates on my nerves, but I can see that there’s not much good you can say.
An alternative of course, is to shoot the messenger and try and claim that the poll is all wrong and really you are in front. That’s when Ken Livingstone has tried to do today… not, it has to be said, very successfully. Let’s see what he said:
1) YouGov’s data does not accurately weight the proportion of black and Asian people in London [...] Yougov’s Internet panels therefore do not accurately reflect London’s population, most notably significantly under-counting ethnic minority Londoners It’s true YouGov don’t weight by ethnicity: no pollsters do in normal political polls. There is no particular reason why they shouldn’t, in fact, it tends to have a good correlation with voting intention and individuals don’t change their ethnicity over time, so it would be quite a good weighting variable – but no one does it. What is bizarre is where on earth Livingstone got the idea that the YouGov sample didn’t include enough ethnic minority voters, just because you don’t specifically weight by a variable, unless there is a bias in your sample it doesn’t follow it will be unrepresentative – in fact, polling is based on the assumption that if a sample is representative in terms of age, gender, class, etc, etc, it will be representative in other ways, like who people would vote for or what they think about policies.
There are no published figures for the ethnic make up of the sample, so there is no way it can be known and nothing for Livingstone to base his criticism on. It looks from this Guardian report that they’ve got the strange idea that YouGov’s sample included the same proportion of ethnic minorities as in the rest of the UK. This is just rubbish, why would it? In the last YouGov poll of London around 24% of respondents were from ethnic minorities. That probably isn’t high enough – the proportion of ethnic minorities in London between the ages of 18 and 65 is around 29%, so even once you add in all those white pensioners it’s still going to be slightly higher than 24%…but really not very different. So first criticism: FALSE, there is no reason to think YouGov underestimate white Londoners by a signficiant degree.
2) “YouGov has a record of significantly underestimating Ken Livingstone’s vote in London. On the day before the last Mayoral election YouGov put put Ken Livingstone just two per cent ahead of the Conservative [...] even though Livingstone actually won by nearly 11 percent. We’ve had this wheeled out before, and it’s still false. In 2004 YouGov published two sets of figures, one was based on only those likely to vote, and overestimated Steve Norris’s support. However, the figures in today’s poll are NOT based on only those certain to vote, they are based on ALL respondents. YouGov’s 2004 figures conducted on this basis predicted that Livingstone would win by 55% to Norris’s 45%. The actual result was Livingstone 55% to Norris 45%, so YouGov got it exacly right. So, second criticism: FALSE, YouGov’s final comparable poll in 2004 got the result correct.
3) YouGov today claims to find that Ken Livingstone will get no more votes for Mayor than Labour will win in the London Assembly elections – both will get 37 per cent. In the 2004 election, Ken Livingstone polled 11% higher than the Labour Party in the Assembly elections and all other polling, plus the Labour Party’s canvass returns, has shown Ken polling significantly ahead of Labour for the London Assembly. Labour’s canvass returns are, of course, not available for us to use as evidence and there have been no other published polls that asked about voting intention in the London assembly. While Ken Livingstone did indeed poll above the Labour party in 2004, there is no particular reason to think this is automatic. Peter Kellner’s explanation, that it shows exactly what it appears to: Ken used to be more popular than Labour but isn’t anymore, seems perfectly feasible. Third criticism: FALSE, there is no other publically available contemporary data on voting intention in the London Assembly to compare this too, and no reason to think it wrong
4) Today’s YouGov polling on issues reveals results far out of line with normal telephone polling. Normal telephone polling carried out by IPSOS MORI only two weeks ago found Londoners supporting the £25 a day CO2 charge on gas guzzlers by more than two to one at 61% per cent to 27%, whereas YouGov reports that this lead is only 41% to 39%. The two questions here weren’t comparable. YouGov asked a single question about supporting “increasing the congestion charge to £25 a day for higher polluting cars”. The MORI poll referred to is presumably this one, in which people had previously been given a list of examples of cars that would be affected by the £25 charge and had been told about the 100% discount for cars producing the lowest emissions, which YouGov didn’t mention. That’s not to say there is anything wrong with MORI’s poll, just that when people were given more information about the policy and the discount that’s included they are more positive about it. So, fourth criticism: FALSE, the questions weren’t comparable.
5) Today’s poll shows minority parties (that is all parties other than the three main ones) receiving only 5% of the vote [...] This is totally implausible . On this criticism I have some sympathy, in fact, I’ve been saying similar things in response to earlier mayoral polls. At the last election others got 18%. A lot of that was probably due to the European elections being on the same day and UKIP’s resultant strong showing…but even then, UKIP only got 6% so the 5% still seems very low. It may be because they’ve had minimal media coverage and minor parties may creep up in the polls once the mayoral manifestos are sent out and people become aware of them, it may be because YouGov prompting only by the big three parties is producing an artificially low level of support. As a caveat though, we thought the same thing prior to the Scottish elections and Peter Kellner even altered the prompting in YouGov’s questions to be more friendly to minor parties… the end result was to over-estimate them. It turned out that the collapse in minor party support was genuine and it might be here too. So, fifth criticism: FAIR POINT, but it’s not just YouGov, Ipsos MORI too had the others on only 4%
Interestingly enough there is another legitimate criticism that Livingstone missed – the poll was done over the Easter weekend. It made not have been such a problem this year since many school holidays weren’t actually at Easter, and the weather was so hideous most people were probably cowering at home, but traditionally long bank holiday weekends produce strange polling results because of people being on holiday, out with their families, visiting relatives and so on.
In situations like this I do always hope that politicians who poo-poo polls they don’t like don’t really believe what they are saying. Polls do set the agenda, so it is perfectly rational for politicians to try and disparage polls that put them in a bad light (though obviously, as a supporter of better public understanding of polls, I do hope the public see through any such pathetic, base claims and run the culprits from office with flaming torches), but if they actually believe such things in the privacy of their campaign headquarters and dismiss any unwanted messages as flawed they are in deep trouble.
Filed under: London