Yesterday's Scottish referendum poll conducted by YouGov has caused real panic amongst those supporting the better together campaign after the opinion poll placed the 'Yes' vote in front for the first time. According to YouGov's latest findings, the number of people planning to vote for independence on 18th of this month now rests at 51% while the number of people panning to vote 'No' is 49%.
Due to how close current polls are, many have called the referendum 'too close to call', however, by using polling data from previous democratic processes, it is still likely that going by results of YouGov's latest poll, Scotland will vote in 10 days' time to remain a part of the United Kingdom.
Through looking at historical opinion poll data from previous British votes, it can be seen that how people claim they will vote on polling day can be dramatically different to the actual result. The best example of this is the 1992 general election in which the majority of pollsters predicted a very tight Labour victory; only for the Conservative party to comfortably win the election with a 7.6% lead. Although an inquiry set up by the Market Research Society into why the polls were so wrong found several errors in the way polls were conducted, the British General Election Study 1992, found that many people who had told pollsters that they 'didn't know' who they were going to vote for ended up voting Conservative.
Equally, it was discovered that those who refused to disclose how they planned to vote also disproportionately voted for the Conservative Party. Given the unpopularity of the Conservative government at the time, it likely that many of these 'don't know's or 'prefer not say's did not want to admit that they were considering voting Tory because they feared that the majority would disapprove of their decision.
Also in the 2010 general election opinion polls in the lead up to the vote indicated a different outcome to how the electorate cast their ballot. In the final opinion polls of 7 highly regarded pollsters, all of them over estimated the success of the Liberal Democrats by an average of 4.4 percentage points. Further, all of these pollsters underestimated the share of the vote that the Labour party would receive by an average of 2.3 percentage points.
In both these examples, opinion polls have overestimated the support for change.
Therefore, although YouGov's most recent poll places 'Yes' ahead by two percentage points, the poll also indicated that eight percent of those eligible to vote still 'don't know'. By looking at the historical evidence from the 1992, those who are unsure tend to be cautionary in their voting behaviour and vote for the status quo certainty instead of change. This is likely to be because: if a voter votes for the status quo, they know what to expect; however, if they vote for change they vote for a path untread and paved with uncertainty.
Further, the evidence from the 1992 and 2010 general elections indicate that opinion polls normally overestimate the support for change. Consequently, given that YouGov's most recent poll still only places the 'Yes' campaign 2 percentage points above the 'No' vote, the chances are that the actually figures are more 50/50.
Once the 'not sure' and 'prefer not to say' votes are added in to the equation, it is most probably that going on YouGov's most recent data ,the Better Together campaign are still narrowly in the lead.