Where has Labor’s rise in the primary vote come from? Across all polls now, it’s clear Labor has lifted its vote from the depths of despair of earlier this year, when it sank to a primary vote in the high 20s in some polls, up into the mid-30s.

If it gets to ~38%, the view within Labor goes, it’s game on. The pressure will start to mount on Tony Abbott, as long as Labor can keep its mistake rate down. All those months of Abbott being so far ahead in the polls has inculcated an expectation of victory in the Coalition, that the next election is in the bag. A change of tactics will be demanded. Some will want more than a change of tactics. They’ll look to Malcolm. Or maybe Joe.

That’s the theory anyway. There’s some validity to it. Abbott will remember the pressure John Hewson came under when the Coalition’s double-digit lead over Labor began disappearing in 1992 as Keating made good his promise to “do you slowly”. It didn’t remove complacency, it panicked Hewson into changing key elements of Fightback. It’s strange what expectations — your own and others — of victory can do to your judgment.

In cold, hard reality, Abbott would be prime minister with a huge majority if an election were held now. But the reality is also that Labor’s vote has lifted.

Essential Research’s polling had Labor bottoming out at 31% in July. This week it hit 36%. Essential runs a fortnightly rolling average, so yesterday’s poll reflected last week and the week before, across 2000+ respondents. It means it’s less prone to the normal fluctuations of polling, but also takes a while to pick up fast-developing changes, although I suspect they’re far rarer than a lot of commentators would have us believe.

The first thing to note is that it has accompanied a fall in the Prime Minister’s net disapproval rating, which as my colleague Possum pointed out, is critical to restoring a government’s fortunes.

This week’s poll, conducted in the second half of last week, showed a significant rise in the raw numbers for female support for Labor. In the space of a week, female support for Labor (excluding “Don’t Knows”, who are allocated by a separate process) went from 29% — it’s been below 30% for months — to 34%. Now, you could argue that’s because of the recent focus on Tony Abbott’s aggression, his use of “chairthing” decades ago, alleged wall-punching and so on, which is interpreted to be of special significance to women.

But it may just be polling static. We’ll know if the rise is sustained, over the coming weeks. What is clearer is the male voters appear to have warmed to Labor. Male support for Labor has lifted from 25% back in July to 31% now, and has been there for several weeks. So it’s not clear that there are any particularly female dimensions to the rise in Labor’s support — yet.

What else has lifted support for Labor? For all the focus on Queensland, NSW has strengthened for Labor: the Coalition led nearly two-to-one in July, 47-24% (remember, excluding “Don’t Knows”). But that lead has steadily shrunk, and this week is down to 7 points; last week it was 10 points. Victoria had a sudden Labor surge this week, but has otherwise been roughly steady. The gap between the parties in Queensland has remained roughly the same, around 11-12 points.

Beyond that, we get to sample sizes that are too small. The Coalition lead over Labor among middle-income earners — $52,000 to $83,000 — has steadily shrunk from around 8 points in July to zero this week. But that’s only with 230-odd respondents. The Coalition’s lead over Labor among 45-54-year-olds has steadily shrunk and been replaced this week with a handy lead, but again it’s only a couple of hundred respondents. They’re demographics worth keeping an eye on, however — especially the crucial middle-income earners.

What the data does tell us is that a couple of the frequently-discussed causes of Labor’s rise — Abbott’s reputation with women, and the actions of Campbell Newman — aren’t necessarily the reasons behind the current rise; it could be Barry O’Farrell, who has governed in a steady-as-she-goes manner, who is doing the damage. And maybe blokes dislike pugilistic alternative prime ministers too.

Then there are the Greens. The Greens have demonstrably lost a slice of their primary vote, which for a long period hovered at 10-12% but is now down to 9%. This might suggest Labor has had some success at pulling more progressive voters back from the clutch of the Greens. But the Greens’ vote had already hit 9% before Labor’s recovery began, so it’s hard to say that Labor’s primary vote rise is built on the Greens. And this week, the Coalition vote fell — not by much, but it’s the first time that the Coalition has fallen noticeably. Labor doesn’t need to win back Greens voters, except in a couple of urban seats where it’s in a real fight with them. It needs those voters who’ve switched to the Coalition to come back.

For over a year the polls have blandly shown a Coalition landslide. Now they’re on the move. The next few weeks will be interesting.