BREXIT: The 2 reasons why Leave must approach recent polls with caution
The world of EU referendum opinion polls took a stunning turn on Tuesday when ICM's most recent phone poll indicated a four-point lead for Leave.
Remarkably, this meant that there had been a 14-point swing in favour of a Brexit since ICM's previous phone poll which was conducted just two weeks before.
The result also meant that the recent trend of surveys conducted over the phone consistently showing comfortable leads for the Remain campaign was broken.
It would be unwise for Leave campaigners to get overly excited about the poll, though. That's because there are 2 important issues to consider when assessing its significance.
These issues were explained by poll analyst Matt Singh in his latest post on the Number Cruncher Politics blog.
Polls conducted following key events often produce overshoots
As Sign explains in his post, polls produced following major events often produce dramatic but short-lived results.
This was the case immediately after the draft version of Britain's renegotiated EU membership was published - certain polls showed huge leads for the Leave campaign, which didn't last.
It is important to note that ICM's latest phone poll was conducted shortly after the latest net migration figures were released.
These figures boosted Leave's cause as they showed that David Cameron was failing to fulfill past promises on reducing immigration to Britain.
Of course, it is possible that public opinion is shifting significantly in favour of Britain leaving the 28-nation bloc. But this conclusion can only be confidently reached if this trend sustains in polls published in the coming weeks.
Until that happens, the possibility remains that ICM's latest poll was merely an overshoot born in the immediate aftermath of the net migration figures.
.jp2)
Experts are suspicious about polls conducted over bank holiday weekends
The second reason why the ICM poll should be approached with caution is that it was based on data that was gathered on a bank holiday weekend.
Phone polls conducted on these weekends in the past have thrown up really odd results. Singh believes that this might be because an abnormally amount of people are away from their homes, meaning samples can be skewed.
An example that he cited was when general election voting intention polls conducted in December 2014 would randomly diverge at least partly because they fell on Christmas holidays.
But the 'bank holiday weekend effect' theory requires more evidence to be developed further. YouGov's Anthony Wells said the impact it has on polling remains unclear in a post published on the UK Polling Report.
However, it is clear that the remarkable results of ICM's latest poll can only really be understood when assessed alongside surveys that will be published in the coming days and weeks.