If the social media platform didn’t exist, where else might reporters find another such source of easy-to-write stories about the alleged mood of the electorate, as represented by the entirely unrepresentative gaggle of tweeters? Why, they would have quote the scribbles on toilet doors!
The Wednesday before the 2001 election The Bulletin published a Morgan poll putting Labor ahead, with an election-winning two party preferred vote of 52.5% to the Coalition’s 47.5%. The day itself produced a different result. A tad over half the votes ended up flowing John Howard’s way. The Coalition was returned with 50.98 per cent of the vote.
A pugnacious Gary Morgan argued it was quite possible that the raging debate on asylum seekers – this was Tampa time, remember – caused a last minute shift. The Australian Electoral Survey shows a relatively high 35% of voters only made up their minds which way to go during the campaign. But his arguments cut no mustard with Kerry Packer. Soon after The Bulletin and Roy Morgan Research ended a long association.
Why retell this unfortunate tale? Because it’s fascinating to see how the bar has been lowered in the age of social media. Twitter is not a polling company. It is, at best, a source of aphorism mingled with a little news, but it functions mostly as a digital dunny door on which oddballs can satisfy their urges to scribble. But that doesn’t stop Twitter taking itself very seriously indeed. Every day of the campaign its media minions have been issuing a whole range of metrics. This has included a “who had had Twitter talking”, a measure of the percentage of Tweets mentioning each party handle, as if that is some indication of voting intention.
It placed Labor, on 48%, in the lead for the week to yesterday. The Coalition was well down, on 38%, and the Greens on 14%. A poll aggregate for the same period shows a very different result indeed: the Coalition’s primary vote at around 40%, Labor on 33% and the Greens on 10%.
But we’re in the age of feelpinions, where facts are an unnecessary restraint. Twitter is subtly conflating the entirely objective utterances of the self-selecting sample of the population that uses its product – many not even old enough to vote – with the supposedly scientific models of the pollsters.
The result is more junk data in an election where the wobbly products of robo-polls already abound.
Come back, Gary Morgan, all is forgiven.