Survation’s New “Ballot Prompt” Voting Intention Method
How Does It Work and What Impact Does it Have?
In today’s Survation’s Daily Mirror poll on the penultimate campaign day before the general election, we have again used a new, innovative, methodology to supplement our usual voting intention questions, which we first released for our Mail on Sunday poll last weekend.
In addition to our usual voting intention questions, we showed respondents a replica of the full ballot paper they would see in the polling booth in their own constituencies on 7th May, based on the postcode they entered at the beginning of the survey. Below is shown an example from Prime Minister David Cameron’s Witney constituency.
“The general election takes place this coming Thursday, 7th May. Thinking specifically about your own constituency of Witney, who do you think you would be most likely to vote for if the election was today and you had to choose?”
The Importance of Candidates – Not Just Parties…
The 632 different versions of this ballot paper (one for every constituency in Great Britain) were then combined for reporting purposes into a single “ballot prompt” data table, allowing us to get an understanding of how voting intentions are affected by many of the elements not otherwise captured in usual voting intention polls.
This is not only the local candidate names, which can have a big impact where there are popular local incumbents, but also more “minor” details such as the order that candidates appear on the ballot (favouring those higher up), the party logos that appear next to their names and also an intangible factor of being presented with something that looks like the “real deal”, encouraging a more serious contemplation of the imminent nature of the election.
This method is also helpful in terms of obtaining a better estimation of national vote share for parties that are not standing in every constituency – how might a Green Party or, particularly, a BNP supporter vote, for example, when they realise that party does not have a candidate in their area? The BNP for instance have only been able to field a mere 8 candidates in the election due to their party’s financial collapse – down from 339 in the 2010 general election.
Impact on Results
In terms of what effect this had on voting intentions, our initial voting intention question had the following levels of party support:
Conservative – 32.6% Labour – 33.7% Liberal Democrat – 8.9% UKIP – 15.6% SNP – 4.3% Greens – 3.8% Other – 0.7%
However, once the ballot papers were shown and people were forced to choose between the specific candidates in their constituency, the voting intention was as follows, with changes shown since the first question:
Conservative – 31.0% (-1.6) Labour – 32.3% (-1.4) Liberal Democrat – 10.1% (+1.2) UKIP – 15.2% (-0.4) Green – 5.4% (+1.6) SNP – 3.9% (-0.4) Other – 1.8% (+1.1)
As can be seen, the main beneficiaries of this were smaller parties, particularly the Liberal Democrats (who are expected to benefit from significant incumbency factors in their strongest seats), the Greens and “Other” candidates, who no doubt benefit simply from people being reminded that they exist and are standing in their local area.
The big losers are the Conservative Party, slightly more so than Labour, with the result that the Labour lead increases a fraction compared with the first question. This result is particularly interesting as it calls into question the suggestion that the Conservative Party will enjoy a major first time incumbency boost thanks to all the seats they gained in 2010.
The other interesting result is the slight drop in SNP vote share. This is hard to draw many conclusions about in a Great Britain wide poll where Scotland is only a small subsample, but it may suggest that the SNP’s largely unknown slate of candidates in Scotland will face a slightly more difficult time than otherwise thought in unseating Labour and Liberal Democrat incumbents in Scotland.
Director of Research, Survation Ltd