It appears Scotland in Union have been systematically withholding results from their own polls that are unfavourable to the anti-independence cause

Just a little postscript to my blogpost yesterday about Survation's latest propaganda poll for the anti-independence pressure group "Scotland in Union".  I've just noticed that the Twitter account British Electoral Politics posted a tweet about the poll that gave different percentage changes on the Westminster voting intention numbers than the ones I gave in my own post.  The SNP are listed as being down three percentage points rather than one, and Labour are listed as being up seven points rather than four.

I initially wondered if I had made a mistake and used the wrong Survation poll as the baseline for the changes - ie. I thought there may have been a more recent Survation poll with Westminster voting intention numbers that I had overlooked.  But the date given by British Electoral Politics for the baseline poll is 3rd May, which was the date of the previous poll in the series of Survation propaganda polls for Scotland in Union.  I went back and checked the data tables for that poll, which are available on both the Survation website and the Scotland in Union website.  There appears to be no trace in those tables of Westminster voting intention numbers.  The most likely explanation is that a Westminster question was indeed asked in that poll, but Scotland in Union instructed Survation to withhold the results because they were too favourable for the SNP and too unfavourable for unionist parties.

Now, to be clear, withholding results on a particular question is not against British Polling Council rules.  Scotland in Union were entitled to do it, but by the same token we're entitled to draw our own conclusions from the fact that they did it.  It may well mean that a large number of other inconvenient results have also been edited out of their polls.  It also means, going forward, that if parliamentary voting intentions are missing from Scotland in Union polls, it's probably to hide a good result for the SNP.

Where there may be more alarm bells ringing in terms of the rule-book is if Scotland in Union themselves have now suddenly decided to make the Westminster results from the May poll known, with a view to making the percentage changes on the new poll look more flattering for Labour.  If that's the case, we're now entitled to see more complete datasets from the May poll.

I'm also fascinated by what this episode tells us about the internal politics of Scotland in Union.  The Westminster numbers from the new poll are moderately good for Labour but an unmitigated catastrophe for the Conservatives.  The fact that Scotland in Union have departed from their apparent normal practice of keeping the results secret may suggest that they're serving the interests of Labour and not the interests of the Tories.