Looking at the YouGov poll for 4th – 8th March, which suggests a 2 point lead for NO and a 3 point fall in SNP support, I was browsing the M/F and age ratios to see they made sense. They do, but when I looked at the sample in terms of region, something struck me. It may or may not have been a worthwhile strike.
One of you will know.
Anyhow, here is the distribution and the Yes/No ratio:
Region Electors Sample Yes No
North East Scotland ? 164/164 44 56
Highlands & Islands 206K 109/113 49 51
South Scotland 208K 167/178 43 57
West Scotland ? 112/111 51 49
Central ? 126/127 47 53
Mid-Scotland & Fife ? 132/117 54 46
Lothians 650K 183/184 48 52
Glasgow 462K 108/106 59 41
Why is the strong Yes region of Glasgow sample (106), much smaller than the one for No majority region of Southern Scotland (Borders, Dumfries & Galloway) (178), yet it has more than twice the electorate?
In YouGov’s November poll the Glasgow sample was 136 and the South Scotland sample was 154.
More interesting, the Ipsos MORI poll in October which recorded the highest lead for Yes of 52 to 39, had a Glasgow sample higher than the South Scotland one by 156 to 125.
Also, the Savanta poll in December which returned a lead for Yes of 52 to 38 had a Glasgow sample of 139 compared to a Southern Scotland one of 83.