Datablog badge new 620

Scottish independence: advantage no, but we need more polls

Since August 1, 17 polls have been released, of which only one had yes ahead. With less than a week to go it’s still advantage no, but a narrowing gap and polling uncertainties make it a close race

  • theguardian.com,
  • Jump to comments ()
Former Labour Prime Minister Gordon Brown making the case for Scotland staying in the UK.
Former Labour Prime Minister Gordon Brown making the case for Scotland staying in the UK. Photograph: Murdo MacLeod/Murdo MacLeod

A poll released overnight by YouGov has the no to independence camp back in the lead, ahead by four percentage points. A week ago YouGov placed yes ahead by two.

This is a relatively sizeable change in just a few days. What’s going on?

Here’s a different question. Since August 1, 17 polls have been released publicly, of which only one had yes ahead. Guess which poll received the most attention?


This chart shows the average no lead over the past year. The topline trend is quite clear, and the underlying trends moving towards yes are discussed here. This doesn’t mean yes is ahead, but that a 20 point gap a year ago, crept to 15 at the beginning of 2014, was 10 points a month or so ago, and is now averaging less than five points.

Polls released since the second debate show the following no leads: +5, +6, -2, +4, +1, +6, +6. In August, the gaps were +14, +13, +20. +13, +9, +4, +13. Once margins of error, fluctuations within polls taken by the same company in short succession (the latest YouGov isn’t too dissimilar to figures released just before the poll showing yes ahead), the overall figures and trends are in fact relatively consistent. In addition to focussing on single standout polls, there is actually a small number of polls being carried out. If there were 50 sets of polls released, the perception, and relative proportion, of outliers would likely be different. Furthermore, when companies release figures on a monthly basis, a double-digit change is perceived and headlined differently, and more dramatically, to smaller regular changes visible within data released over more frequent intervals.

So, what’s going on in Scotland?

The gap between no and yes has narrowed significantly over the past few weeks, and most of the movement has been towards yes. Although with a much smaller margin than just a month ago, no still leads though. It is also worth keeping in mind that about 18.5% of the 4.2 million registered voters are postal, and many of these ballots will have been returned since they were sent out on Aug. 26.

Yet, with less than a week to go important uncertainties remain:

  1. Turnout: a record 97% of eligible voters are registered to vote. This anticipates what is likely to be an historic level of participation, far higher than the turnout levels of regularly held elections. Many that have never voted, or haven’t done so in a long time, will head to the polls next Thursday.
  2. Polling referendums is inherently complex, and pollsters have often been wrong due to the lack of historical data, and uniqueness and diversity of questions.

Polls rely on past behaviour to interpret the present, and historic data in this case is scarce. This difficulty is compounded by the small number of polls being carried out in Scotland. Given the trends over the past month, the narrow gap, and these challenges, it’s advantage no, but the final whistle is still some way away.

Until then, what we really need are more polls.

Today's best video

;