As I posted a few days back, it would seem that Social Media trends do give a better indication of current voting trends rather than traditional polling.
Possibly
@GrandOldTeam (sorry if it wasn't you I was chatting to about it) had some different interpretation about it.
Nope - many moons ago I wrote a paper on this (that was presented in Russia.. with hindsight they might have been pooling research) that showed that social media was a particularly poor tool for looking into voting intentions.
The problems are threefold- firstly social media has a liberal/left wing bias in that those more politically towards the left are more likely to use it (will be exceptions obviously) so left wing belief/support will seem more prevalent.
Secondly the ease of which that bots/sock puppets can be set up means that something can seem far more popular of an idea than it is. I could create a post saying ‘let’s castrate Jebus in the town square tonight!’ Then get a thousand sock puppet accounts to like it. That makes that post look mega popular - and with the way that algorithms work it’s more likely to be seen.
Thirdly, the ‘shy Tory’ notion is in full effect, linked to the first point, if someone has a point that goes against the social media concensus then it likely won’t be posted. That missing post therefore doesn’t get counted. Plus do people always tell the truth on social media?
In short, social media is a laughable bad way of polling opinion. During the Cameron/Milliband debates one outlet used sentiment analysis to gauge public opinion pulled from Twitter. According to it Cameron got slated - when you actually pulled the data and looked at it, some young lad called Cameron from London was getting pelters from his GF on Twitter at the time - so her not being happy with her bf was somehow indicative of dislike for David Cameron.
Better days in a way.
