One Nation is up. One Nation is down. What the weekly polling movements are actually telling us and what they are not.

In a single week, three separate polls, Resolve, Newspoll, and YouGov, told three separate stories about the state of Australian politics.

And yet, reading some of the commentary that followed, you would be forgiven for thinking Pauline Hanson’s movement had suddenly stalled, collapsed, or surging to new heights, depending on which headline you clicked on first. The dominant framing, as ever, was that something had shifted. Something was moving. Labor had “surged back.” The Coalition was “stable.” The insurgency was fading, or perhaps accelerating, or perhaps both, simultaneously, in different households, reading different mastheads.

Most of this is the reflexive reporting of survey-to-survey movement as if each wave were a referendum in miniature and it reflects a profound misunderstanding of what is actually happening to the Australian electorate.

We are not in a normal political period. We are in the middle of a structural realignment, and the tools built for normal periods, including the habit of reading each poll as a discrete verdict on the state of political play, are not fit for purpose. If you want to understand what One Nation’s vote is doing, you have to stop staring at the weekly numbers and start looking at the shape of the thing beneath them.

What the surveys are actually measuring

Start with a simple observation that very few columnists seem willing to make: when the electorate is in flux, different sampling methodologies, different weighting schemes, and different question orders produce larger gaps than they do in stable periods. This is not a flaw in the polls. It is a feature of the moment. The country is in transition.

In a settled two-party system, where the overwhelming majority of voters are either rusted-on or weakly attached to one of two brands, polling is a relatively straightforward exercise. You sample, you weight against known preference flows and demographic benchmarks, and you produce a number that will be close to what every other reputable pollster produces, give or take a couple of points. That was Australia from the 1950s through to 2022. That was the environment in which most of our political journalism was professionally formed.

That Australia no longer exists.

In the current environment, a single respondent’s answer depends on a cascade of variables that have only become decisive in the last five or six years. Whether they are prompted with One........

© Pearls and Irritations