As I've spent a fair bit of time recently calculating poll averages, I thought I might take it a step further and do the same thing for what is probably the closest international parallel to our own independence referendum, namely the Quebec referendum of 1995. I'm sure most of you know that, in the early stages, the Yes camp in Quebec found themselves in a seemingly hopeless predicament, before recovering with a few months to go to roughly where we in Scotland are right now on a mean average of the polls - 40% Yes, 60% No (with Don't Knows excluded). They then made further impressive gains which took them all the way to a narrow lead by the close of the campaign.
That part of the story is well-known, but what might surprise you a little more is that an average of the polls from September 1995, the month before the referendum, shows that Yes were still trailing by more than five points...
SEPTEMBER 1995 AVERAGE (excluding Don't Knows) :
An average of polls conducted during the month of the referendum itself shows that the position had been dramatically reversed...
OCTOBER 1995 AVERAGE (excluding Don't Knows) :
Yes 50.7% (+3.3)
No 49.3% (-3.3)
Then of course came the famously cruel twist in the tale that has poisoned Quebec politics to this day. Fieldwork for the final poll was concluded three days before the referendum, too late to pick up any late swings. The No campaign indulged in some rather questionable practices in a desperate effort to rescue the situation, and it appears that this was sufficient to generate a small but decisive late swing back to No. However, this didn't reverse all or even most of the gains that Yes had made between September and October, as a comparison between the final result and the September average will demonstrate...
REFERENDUM RESULT, 30 OCTOBER 1995 (changes from September polling average) :
Yes 49.4% (+2.0)
No 50.6% (-2.0)
I'd say those numbers fairly definitively give the lie to the hoary old myth that you'll still sometimes hear otherwise intelligent people trot out, namely that a Yes campaign is bound to suffer a seepage of support during any referendum campaign, and that undecided voters are bound to break for No. More specifically, it suggests that our own Yes campaign in Scotland could still hope to win even if it remains a few points behind next August. It currently stands at 40.1% in the Poll of Polls with Don't Knows excluded, so if the Quebec precedent is anything to go by, it would probably need to increase its support by at least 7% or so over the coming seven or eight months. For my money, the crucial period will be from May until early August, because it's in May that the official campaign period starts, and from that point on the broadcasters will finally be obliged to treat both sides of the debate equally (and it'll be fascinating to see how faithfully they live up to that obligation).
NOTE : Apologies if I got my French grammar wrong in the title of this post. If so, hopefully Tris will correct me!