What do the academics say? How to improve your polling day operation

Welcome to the latest in my occasional series highlighting interesting findings from academic research. Today, a selection of findings from Sex, Lies and the Ballot Box: 50 things you need to know about British elections, edited by Philip Cowley and Robert Ford, which illustrate how election campaigns can maximise the impact of their efforts to get supporters to vote.

First, the evidence from Eline de Rooij which points towards targeting households with multiple occupants where the other occupants are not known opponents:

One study in the US found that a doorstep canvassing campaign targeted at individuals living in two-voter households also had a substantial impact on the turnout of the household member who was not canvassed: 60 per cent of the effect of the campaign was passed on to the second household member. Although similar studies conducted in the UK are yet to be published, preliminary evidence from one study during the Police and Crime Commissioner elections in 2012 seems to suggest that spill-over also occurs within UK households.

Note that in the UK evidence:

A personal reminder to a voter seemed to work pretty well at motivating their partner and children as well, and worked best in households in which members supported different political parties.

Hence the need to consider the voting intention of not only the person you’re trying to get out to vote but also of their other household members.

Moreover there’s the evidence presented by David Cutts, which points to targeting households where a supporter lives and no-one in the household is a regular voter because:

More than nine out of ten people who lived in a multi-person household where someone voted, voted themselves in the 2010 general election Less than one in ten did so when living with a non-voter.

Of course in practice campaigns never manage to identify all their supporters. But there’s a helpful family pattern that allows for useful extrapolation from known data – new, young voters are much more likely to vote the way of their mother than their father:

In studies where both the mother and father support the same party, the child is three times as likely to support it too. Where only one parent selects a party, the child is up to ten times more likely to choose the same party as the mother than that of the father.

(Another reason too, by the way, along with the facts that women are a majority of the electorate and that party polling has repeatedly found women to be a majority of Lib Dem swing voters, why the Liberal Democrats need to take appealing to female voters very seriously.)

For more about effective campaigning, of course, take a look at 101 Ways To Win An Election.

Note: I’ve summarised briefly the research above and hence not included the details about allowing for other factors. But all the research seems to be solid in concluding that the relationships I’ve mentioned are statistically significant even after allowing for other shared factors.

You can read the other posts in the What do the academics say? series here.