Ever since the Newspoll results on Saturday, I’ve been wracking (and wrecking) my brain trying to find a logical answer to the conflict between two opposing political manifestations: the anti-Andrews protests v the pro-Andrews polls.  

The Newspoll, published in The Weekend Australian on Saturday, has Labor’s primary vote at 44 per cent, up from the 42.9 per cent recorded at the “Danslide” election in November 2018. 

On a two-party preferred basis, Labor is sitting at 58 per cent, up from 57.3 at the 2018 election, while the Coalition is on 42 per cent, marginally down from 42.7 per cent three years ago.  

It’s unlikely that the Newspoll is a rogue poll. Two less publicised Morgan Polls, one out yesterday and the other from November 12, both give Andrews similar leads. And Victorians vote in exactly 12 months today.

Some observers have claimed the anti-Andrews protests as the largest since – and perhaps larger than – the Vietnam war. Can both reflect political facts on the ground?  

As everyone knows, Victoria is not a single electorate. The poll is therefore not a reliable representation of voter intentions electorate by electorate, even as a snapshot in time.

For one thing, as previous polls have shown, electors may say one thing to pollsters reflecting their feelings about parties at the time, but that may not translate into actual votes at the ballot box. ‘I am fed up with the Liberals… on Tuesday, but on Saturday when I get the pencil in my hands, I cannot cast a vote for Labor’, sort of thing. 

Secondly, in a safe Labor seat, say, even a whopping increase of 10% in intended Labor votes would drive the poll numbers — but the seat would remain just one Labor seat. If that were a pattern, the polls would pick up the increase for Labot but it would be meaningless in terms of seats (if intentions were translated into votes). 

Maybe that’s why at the last election, the polls pointed to Coalition hell, but the ballot box delivered a miracle.