In an article on Wings over Scotland, the importance of Glasgow for the independence campaign was dismissed: "[T]his site’s prediction remains the same [...] – if we get 38% in Glasgow come the autumn of 2014, we’ll win."
This might of course be true, but it didn't sound plausible to me, so I decided to have a quick look at the likely distribution of Yes and No votes in the case of a narrow win for the Yes side.
I first found the regional votes from the 2011 election (discarding all parties smaller than the LibDems):
|West of Scotland||117306||92530||35995||9148|
|South of Scotland||114270||70595||54352||15096|
|Mid Scotland and Fife||116691||64623||36691||15103|
|North East Scotland||140749||43893||37681||18178|
|Highlands and Islands||85082||25884||20843||21729|
I then got some useful figures from Ipsos Mori about the correlation between political views and independence voting intentions:
If we assign 90% of the Don't knows to Yes (this might not be realistic, but it's the easiest way to get a majority in favour of independence), we get this:
The result of referendum according to these distributions would be a very small Yes victory (51% Yes, 49% No). The geographical distribution would look as follows:
|West of Scotland||48.7%||51.3%|
|South of Scotland||46.7%||53.3%|
|Mid Scotland and Fife||50.7%||49.3%|
|North East Scotland||55.4%||44.6%|
|Highlands and Islands||54.1%||45.9%|
If my calculations are correct, it's likely the results in Glasgow will be extremely close to the results for Scotland as a whole, and it seems unlikely that 38% Yes in Glasgow would be sufficient to ensure that Scotland becomes independent.