by Martin Odoni

Okay, the tongue-in-cheek opinion poll I put up before the weekend has gone past 1,001 votes now, so that technically means it is now large enough to be considered ‘statistically representative’. Admittedly, nothing else about the poll really conforms to correct procedure, but then it was of course only meant as a rebuke to the arrogance of certain Labour peers who were poised to launch a ‘Motion-of-no-confidence’ in party leader Jeremy Corbyn. I was in effect saying, “How dare you use a procedure that is not even defined in the party’s rules to undermine a democratically-elected leader, especially when your peerages were not democratically-awarded?” And sure enough, the outcome is very decisive – at the time of writing, those voting No Confidence in the Labour peers total an almighty 990. Those who have expressed Confidence in them total a dismal 23. That is just shy of ninety-eight per cent declaring they have no confidence in the Labour peers.

NB: You can in fact still vote in the poll if you have not done so already. I have no plans to take it down, or attempt to ‘halt’ it. I am simply drawing a line under it now as it has had a large enough sampling size.

Now, the peers called off the motion over the weekend anyway, which suggests that they already realised they were inviting more trouble than the move was worth. And with the intended rebuke in mind, the way I worded the questioning of the poll beforehand was blatantly skewed in one direction, so we cannot consider the vote in any sense to be ‘scientific’. If readers wish to condemn me for that, I shall sleep none-the-worse for it. But in any case, supposedly ‘reputable’ pollsters have a history of gathering opinions in ways that are no better.

Take YouGov, for instance, who consistently under-estimate Labour support levels (exception; a polling model they used for a one-off survey around a week before the 2017 General Election correctly predicted a Hung Parliament), and therefore, despite excessive credence given to their data by mainstream media, tend to offer abnormally high numbers of outliers.

Why do they do this? Well, while I am so far unaware of any response or explanation offered by YouGov that might mitigate it, this image may give us a handy clue; –

YouGov blatant leading poll question

Is this massaging-of-questions the reason why YouGov keep predicting an apocalypse for Corbyn’s Labour that no other polling company can even detect?

It appears to be exactly the same old problem of massaging the questions to draw the desired answer. The only difference is that I cheerfully admit to doing it as my aim was satirical, not information-gathering or actual propaganda. YouGov, by contrast, claim that they do it for purposes of impartial research.

And if this is a typical example of how YouGov lead their respondees, well, they are not exactly ‘subtle’ about it, are they?