Skip to main content
ABC News
Selection bias in UK polling (Part 1): Cell phones

We discussed the issue of cell phones and cell phone-only voters here previously in the context of the 2008 US Presidential election, where Nate’s analysis determined that excluding mobiles from the sample negatively impacted Obama’s polling numbers by 2-3 points — a small but significant drop.

This analysis, which compared the topline results of pollsters who included a cell-phone sample in their data with those that did not, was backed up by late 2008 data from the CDC that confirmed that more than 20 percent of US households are cell phone-only, in addition to another 14.5 percent who have a landline phone but receive “all or almost all calls on wireless telephones.”

In the UK, at least 13 percent of households fall into the mobile-only category as of 2008, with lower income Britons far more likely to be mobile-only (23 percent) than wealthier people (8 percent). In addition, younger people are far more likely to live in mobile-only households, including more than a quarter of the youngest demographic group (15-24) and a fifth in 25-34 year olds.


What does this mean for voters this year? To begin with, young voters are the volatile of the voting subgroups in terms of turnout, while older voters are more consistent in voting patterns.


Since 1997, turnout among the youngest voters has fallen significantly, with the 18-24 group the only one that did not have an uptick from 2001 to 2005.

This is actually fairly different from what has been seen in the United States, where the swing in voter turnout among young voters has been even more extreme, and largely dependent on whether a presidential campaign is being waged.

Main takeaways:

1. If youth turnout continues to fall or only moderately rebounds, the impact of not capturing cell phone-only voters may be small enough not to have a significant impact in overall numbers or seat allocation.

2. However, if youth turnout jumps by 10-20 percent and shifts hard towards a particular party, the impacts could be undercounted by telephone polls. Given the Liberal Democrats surge following the debate among young voters, this could be a realistic scenario, meaning that the Lib Dems could be in a stronger position than it seems.

3. Young American voters are a pretty volatile bunch, and overall much less likely to vote (average of 37 percent over last three elections) than British youngsters (45.5 percent over previous three elections).


Renard Sexton is FiveThirtyEight’s international affairs columnist and is based in Geneva, Switzerland. He can be contacted at sexton538@gmail.com

Comments