I realise it is a bit of a cliche: disgruntled left-wing activist takes to the internet to yell at the media for their incorrect reporting of the Labour Party. But here I go anyway.

The way the Labour Leadership contest is now being framed is that Cunliffe has won, as he has the Unions, and the Membership, but the nasty caucus might go against the Will Of The New Zealand People and install Grant Robertson anyway. There is no doubt that Cunliffe has the advantage, and is scoring well in these polls – however, the value of these polls is next to nothing.

On Friday we had the 3 News Poll. It asked 500 people – though from the story, I can’t tell what they actually asked. It was effectively a 40:30:30 split, which isn’t the death knell that Gower editorialised it to be. The poll was conducted at the height of TV3’s Jones-fest, which has been criticised by all bar one of this country’s top media commentators. The One News poll that was on Q&A yesterday didn’t even ask the right question, as far as I can tell. Cunliffe came out tops in the poll, which asked the question “who could beat John Key in an election”. Now, I don’t want to split hairs, but that actually isn’t the same question as “which of the three Labour candidates would you vote for in the leadership election.” Even if that was the question that they had asked, it doesn’t reflect the complexity of the preferential voting system which is being used.

Opinion of the general public is useful, I guess, but both polls narrowed things down to Labour Voters. Neither news outlet released the breakdown of how many people “voted Labour” in their poll groups, but I seem to recall it was around 150 in the 3 News one. Again, I realise this may seem pedantic, but even a poll of people who “voted Labour” is not the same as a poll of people who are Labour members, ie the only people who’s votes count in this election. While Labour don’t release the number of Party members, if we generously assume it to be 60,000, then around 1 in 10 people who voted Labour at the last election is actually a member of the party. That means that of the ~150 people who would vote Labour in these polls, around 15 of them might actually be members. Do I need to point out the statistical irrelevance of polling 15 people?

Further to this, some in the media are reading more into the announcement that 4 of the 6 unions are supporting Cunliffe than they should. It’s clear that Cunliffe will take the majority of votes from these 4 unions – but it’s not correct to say that they are block voting. The executives of these unions have recommended to their delegates that they vote in a certain way, and you would assume that many of them will. However, the delegates are not actually compelled to vote this way, and may be lobbied by members of their unions to vote otherwise.

It seems that the narrative is firming around Cunliffe, but I’d argue that that was based on statistically meaningless polls and a misunderstanding of how the unions are voting. Cunliffe came out of the blocks fast, and dominated the first week of the campaign. TV3’s love-in with Jones, and the subsequent backlash, was the story of the second week. The only time Robertson has really been mentioned was when discussions about his sexuality came up. He needs to make a move, today or tomorrow, to take back some momentum in this last week. He should have been buoyed by the last few days, which have taken in his base (Wellington), his home town of Dunedin, and finishes tomorrow with what will be a blockbuster of a meeting in Christchurch. Robertson is likely to take 7, if not 8, of the 10 South Island MPs, and I would imagine that they will be encouraging their members to vote the same. To make any inroads, Robertson will have to overcome a: the political media getting bored of this contest and b: none of them having much interest in what goes on in the South Island. So that’s why I’m asking, I assume in vain, that the media refrain from calling this contest until it’s actually done.