Survey research offers an important mirror on society.
Occasionally, opinion polls can conjured into existence — to suit partisan leanings. This article looks at a recent example with a supposed survey of Scottish adults.
In response to a ComRes internet panel poll conducted on 24–25th July, a Twitter account claimed there was a Scottish opinion poll. This survey of Scotland apparently showed Labour overtaking the Conservatives:
New Scottish poll has Labour overtaking the Tories…
Tories would be wiped out in Scotland…
There are multiple avenues of suspicion: no company has been named,the central estimates are missing (what share did they change to?), and the fieldwork dates are unknown.
Importantly, neither the position nor change of the Scottish National Party has been stated.
How to check
As a general secondary source, Wikipedia offers a good first port of call for checking about a vote intention poll. The page is divided into national poll results, and those surveys of Scotland, Wales, Northern Ireland and English regions. We should remember that sub-samples of national polls are not internally weighted.
Looking at Scottish poll results: we see no survey results in July 2019.
Another aggregation of Scottish polling results is done by What Scotland Thinks. Again, no Westminster vote intention poll in July 2019 has been recorded.
There are numerous polling aggregation accounts on Twitter, such as Britain Elects, Election Maps UK and Complete Politics. None of these accounts reported a new Scottish Westminster vote intention poll.
Mr James then responded (and blocked) that:
Lab 19… Con 16…
The co editor of PB reported in it last night..
Moreover, these claimed central estimates allow us to disprove the existence of this ‘poll’. We can work the changes (CON -2, LAB +3) backwards, giving us a supposed poll of Scottish adults with:
No such poll exists.
There were no Westminster vote intention polls of Scottish adults in July 2019. Fabricating polling results for partisan reasons should be excoriated.