Lord Ashcroft’s recent poll suggested that the Lib Dems have opened up a significant lead over Labour in the race to win Cambridge at the General Election. Here are the headline poll findings:
But digging in to the copious data tables supplied with the poll suggests that there be more uncertainty than usual for one very important group of Cambridge voters – the 18-24-year-olds.
Pollsters try to come up with results that accurately reflect the population they are sampling by weighting the numbers on various characteristics, such as gender, age, social grade, and how people remember voting last time. For example, if they interview 600 men and 400 women in a particular constituency, they scale up the responses from the women and scale down the responses from the men to get a result that reflects the actual numbers of men and women across the constituency as a whole.
Thanks to the data tables that all reputable pollsters publish under the British Polling Council’s rules, we can see this weighting process at work. Here is the weighting applied to the different age groups for the constituency voting question, from table 6:
In the left-hand Total column we can see the overall result of the poll for the constituency voting question. This isn’t quite the same as the headline result of the poll, because pollsters apply a “spiral of silence” adjustment to allow for the fact that there’s a tendency for some people to be embarrassed about admitting that they vote Conservative. This has been standard practice since the 1992 General Election, a disaster for the pollsters who almost all predicted a hung parliament, rather than the Conservative victory that resulted. However, in this particular poll the “spiral of silence” adjustment isn’t very dramatic – it just transfers a single percentage point from the Greens to the Conservatives.
But the numbers I really want to draw your attention to are highlighted in red, and show the unweighted and weighted base for the 18-24 year olds. They show that this age group is represented in the poll by just 19 people, and their responses have been scaled up by a factor of nearly five. Meanwhile the 45-54 age group is represented by 134 people,and is scaled down by nearly half. Furthermore, the result for the 18-24 age group is very pro-Lib-Dem, with 55% supporting Julian Huppert to just 18% for Labour’s Daniel Zeichner. But there are a couple of factors that suggest that these numbers might not be very reliable – the small sample size, and the timing of the poll.
Because there are only 19 people in this age group, random fluctuation will play a large part in determining the result; it may simply be that the young people the pollsters spoke to happened to be more pro-Lib-Dem than average. But just as significant is the fact that the poll was conducted from Tuesday 24th to Saturday 28th March – meaning that many students will have been away from Cambridge at the time, because of the University vacations. And looking at the population structure of Cambridge, it’s clear that students are a very substantial part of this age group.
While there are certainly some students who are closely interested in politics, and who will be aware of the candidates and the campaign in the Cambridge constituency, there are many who are less so, and who will tend to vote on national issues – not least tuition fees. Judging by the local election results in the student wards in recent years, student political opinion is a lot more pro-Labour, and less pro-Lib-Dem, than the numbers in this poll.
So it’s entirely possible that if the poll had been conducted in term-time, and had managed to interview a larger sample of 18-24 year olds, that we might have seen a very different result for this age group, which would have cut the Lib Dem lead in the headline poll result significantly. In the three and a bit weeks of the campaign that remain, Labour will be doing all they can to get this age group to support them on polling day. There’s everything to play for yet.