Scottish Referendum poll tracker - read the .
Analysis
David Cowling, Editor, 大象传媒 Political Research Unit
Over recent weeks the gap between "No" and "Yes" has narrowed in the opinion polls and this has created its own excitement as the two campaign groups redouble their efforts to convince voters to support them.
As each side tries to create momentum, the polls are used to batter their opponents. But where does that warfare leave individual electors as they seek to discover possible outcomes?
Traditionally, we describe polls as temperature readings valid at the time they are taken.
Individual polls are important, but we should also be looking at any trend over time: is the direction of travel benefiting one side or the other?
However, it is also important to look beyond the headline poll figures for "Yes" or "No".
Most polls will ask respondents how certain they are to vote - are the supporters of one side more enthusiastic about voting than those of the other. Pollsters will often adjust their referendum voting figures to take account of any such differences.
The issues of what currency an independent Scotland would use, as well as the economic consequences of any independence have played a significant role in the campaign.
The campaign polls have also revealed significant differences between various groups: older voters are much more likely to oppose independence than younger voters; women are much less supportive of independence than men; and people living in Scotland but born outside it are less favourable towards an independent Scotland than native-born Scots.
Will the final polls show that these divisions remain, or will they become less stark?
The polls certainly suggest that the real battleground in this campaign is among Labour voters. Conservatives and Liberal Democrats are overwhelmingly opposed to independence but there is evidence that a significant minority of Labour voters support it and, I suspect, they will be the focus of the two campaigns right up to 18 September.
And this campaign will be a real test for the polling companies themselves.
You would think, given the choice is simply "Yes" or "No", that the challenge to produce accurate figures would be relatively simple. But predicting the same "Yes" or "No" choice in the 2011 UK referendum on the Alternative Vote system caused some pollsters serious grief.
The September referendum has no precedent for the pollsters to measure their 2014 poll results against. This issue cuts across gender, families, friendships and parties and that makes the work of the pollsters more difficult than simply measuring either "Yes" or "No" would suggest.