The mobile phone vote
Over at Comment Central Danny's sidekick Alice picks up the interesting fact that there is a signficant difference in US polls that include cellphones and those that do not - if the sample includes cellphones Barack Obama has a 4 point bigger lead.
Alice's conclusion is probably wrong - if the US pollsters are weighting their data correctly then this is not a sign that young people are coming out to vote for Obama. Both samples should be weighted to have the same number of young people - what it actually suggests is that young people who only use cellphones, or who are easier to reach by cellphone, are more likely to support Obama than young people who do not only use a cellphone.
It does raise an interesting question for the UK. None of the UK phone pollsters contact people by mobile phone, so people with only a mobile are excluded from polls (it isn't a direct problem for YouGov who don't use the phone, but obviously having no landline means you are less likely to have internet access, so there's a knock-on effect). This does not make phone polls skewed in terms of age since that is corrected for by weighting, but if people who use only a mobile phone have a significantly different outlook on life than those who have a landline then it would still skew the results. Could there be a big chunk of Labour, Lib Dem or Tory supporters out there without landlines who never get polled?
According to Andrew Cooper of Populus they and ICM have both looked at the issue in the UK and so far haven't managed to detect any differences between the political opinions or voting preferences of landline and mobile-only households. Currently the UK still has fewer "mobile-only" households than comparable countries anyway. In the USA about 15% are "mobile-only", in Italy it's a stunning 37%. For the UK the latest figures have 11% of households as mobile only; the proportion of homes with a landline has fallen to 88%, down from 92% in 2002. So while it isn't a problem at the moment (and may never become one, after all, only 62% or so of households have internet access and that seems to work), it's something that will need to be watched as the proportion of homes with landlines falls.
