Another day, another poll showing the Victorian ALP in the box seat for this month’s election. Newspoll today puts Labor’s lead at 54%-46% on two-party-preferred; down a point since the last poll (although The Australian‘s story neglects to mention that), but at the high end of other recent results.

Politicians and pundits continue to tell anyone who will listen that the voters are not yet paying attention. That can be a self-serving line; it might just mean “they’re not paying attention to me”. But even assuming it’s true, it has two conflicting interpretations.

Sometimes people are disengaged because they just haven’t yet turned their minds to a particular election. It might be seen as unimportant, have a relatively low media profile, or there may just be a lot of other things on (like a Spring racing carnival). That’s the line being pushed by the Liberals; Premier Denis Napthine told The Age last week that “voters had only just started to focus”.

Other times, a lack of attention means that people have stopped listening because they’ve already made up their minds. That usually correlates with stability in the opinion polls, which is what we’ve seen lately — a predictable election is a boring election.

The first possibility should be reflected in a high “undecided” response in the opinion polls, the second in a low one. But unfortunately there’s no standard measure of undecideds; different polls report them in different ways, and different questions can get radically different responses. And as I put it back in 2007, “the problem is metaphysical as much as epistemological: just what is it that these questions are (badly) measuring?”.

There’s some evidence that the number of undecideds this time is higher than usual; Newspoll says that 14% say “they could vote this way but there is just as much chance they will vote for someone else”, up from 9% in 2010. Those who respond that way are slightly more likely to be Labor voters.

But results like that should be taken with great care. The notion of “late deciders” is one of those where politicians and media have a common interest: they both like the idea that voters are more flexible than they really are. It sells papers and it soothes political egos.

So even an election that is pretty obviously a foregone conclusion, like the 2007 federal election, gets plenty of commentary to the effect that people don’t make up their minds until the last minute. In reality, of course, the Howard government was dead in the water months out.

Observers on both sides are starting to talk about the Victorian election in the same sort of terms. I confess I’m not convinced about this. There seem to be at least three reasons to think that this is a different situation from 2007, and that the Victorian election is still open. One is that the polls, although stable, are well short of landslide territory. A gap of 53%-47% is bridgeable in a way that 57%-43% is not, at least in Australia. It’s normal (although certainly not universal) for an incumbent government to improve on its poll results in the final week or two.

A second reason is the difference between state and federal polls. State voters who are ignoring state issues have a default option in the shape of their federal voting intention; what they tell pollsters may simply reflect, in this case, their poor opinion of the Abbott government. As state governments suffer from declining powers and a dwindling talent pool, it’s reasonable to think that trend might be stronger than ever.

Third is the problem of fixed-term parliaments. There’s little hard data on this, but anecdotally it seems that the lack of an announcement (with associated speculation) about the election date deprives voters of an occasion to focus on state politics, giving more importance to the final period of campaigning.

For all these reasons I’m inclined to think that the Napthine government still shouldn’t be written off. But with just four weeks to go, on the evidence of the polls, you’d rather be in Labor’s position.