I was reasonably encouraged in late March when a Panelbase poll for the Sunday Times suggested that support for independence had only slipped from 52% to 49%. I would have expected a bigger drop in a time of crisis when voters are concentrating on managing from day to day, and when big constitutional projects might seem like a luxury. However, late March was fairly early in the epidemic, and I still thought the Yes vote would drift further downwards - and actually I wasn't overly concerned by that prospect, because past experience suggests that shifts in public opinion during a crisis are not a reliable guide to what will happen when things get back to normal. But, as it turns out, I was wrong anyway. A new Panelbase poll suggests Yes support has nudged back up.
Should Scotland be an independent country? (Panelbase, 1st-5th May 2020):
Yes 50% (+1)
No 50% (-1)
This is the Panelbase poll that I heard was in the field a few days ago. I initially suspected it had once again been commissioned by the Sunday Times, because Panelbase's other frequent Scottish client is Wings Over Scotland, and the questions that I'd been told about didn't sound very 'Wings-like'. But then someone mentioned that there were also lots of oddly-worded questions about the trans issue, and at that point it became clear it had to be a Wings poll! Stuart seems to be underwhelmed by the independence results, but I think he's losing his sense of perspective somewhat due to his disillusionment with the SNP leadership and his "something needs to change" narrative. In truth, this is one of the best ever Panelbase polls for Yes. There have only been a tiny handful of occasions when Panelbase have reported Yes in the lead or level-pegging. In recent times, they had become one of the more No-friendly firms and repeatedly showed Yes stuck in the 43-45% range. To now be at 50%, and particularly in these unpromising circumstances, should be regarded as an exceptionally good result in my opinion.
That said, virtually all polling firms have changed their methodology recently to bring in weighting by recalled 2019 vote, and given that the SNP did well in the general election, that may be boosting the reported Yes vote. So post-election and pre-election polls are not directly comparable, and it may well be that 50% in a Panelbase poll now is the rough equivalent of 48% or 49% before the election. But in a sense that doesn't really matter, because the new weightings are likely to produce more accurate results anyway. It's entirely possible that we were slightly underestimating support for independence during 2018 and 2019.
Why would the popularity of independence be so resilient during the worst pandemic since the Spanish flu of 1918-20? I can only guess, but the perception that the Scottish government is handling the crisis more competently than the UK government may be a significant factor.