Malcolm Mackerras, my political science colleague, is about right (still 90:10 against Labor) in his assessment. At the very least the government should average the two polls (50:50 and 53:47 to the Coalition on two-party preferred votes) and recognise that it probably remains behind at about 48:52 at the moment.

This week's polls also showed a rise in personal support for Julia Gillard and a fall for Tony Abbott. In Newspoll, Gillard is now clearly ahead as better prime minister (46:32), but in the Herald/Nielsen poll she only leads narrowly (47:44).

These polls cap an improved three to six months for the government in public approval as measured by the polls. It has been on a steady upward path. Having two polls pointing in the same direction is always reassuring. But the next one or two polls will tell whether Labor's upsurge is reliable.

It may be only temporary relief and should be approached cautiously. The polls bump around as they have with the Greens recently. Newspoll reckoned the Greens went down massively from 12 per cent to 8 per cent and then back up to 12 per cent in the current poll. That pattern is implausible, yet commentators jumped on the 8 per cent figure and produced whole scenarios about the demise of the Greens.

Another troubling example is the measurement of the first preference votes for the Coalition. Newspoll said the Coalition dropped by five points (46 per cent to 41 per cent), while Herald/Nielsen reported that the Coalition was steady on 45 per cent. So polls give differing results.

The polls receive maximum front-page publicity but they don't deserve it. The main lesson is that polls don't speak for themselves; they have to be carefully dissected. They don't in themselves contain explanations for the patterns they contain. So a poll always generates lots of speculation and guess work.

The most reliable possible explanations are likely to be those that have been talked about for some time. Labor's fortunes have always been tied to its own ability to process issues successfully and to the general pattern of Australian federal-state politics. In short, Gillard Labor always needed to hang on long enough to put some runs on the board. It has almost done that. It also needed to hang on long enough to benefit from the arrival of numerous state Coalition governments.

These longer-term explanations are likely to be closer to the mark than latching onto something that has happened in the last couple of weeks before the sampling period. During this time Malcolm Turnbull has effectively criticised his own leader in the course of general remarks lamenting the standard of debate in Australian politics. Barnaby Joyce has unilaterally criticised the Chinese purchase of Cubbie station.

David Marr has published a critical study of Tony Abbott, Political Animal, which once again ignited debate about Abbott's attitude towards women. Julia Gillard mourned her father. Labor moved to deny permission to a super-trawler to fish in Australian waters. And of course there have been the harsh state budgets.

Any one of these things could have shifted voter intentions, though it has been the state budgets and the personal attacks on Abbott that have been seized upon by each side. We really don't know because the polls don't probe that deeply. Commentary is all speculation or educated guesses as a consequence.

Over recent months the federal government has not been a raging success but has had several achievements. First is the introduction of the carbon tax. This has been achieved without much if any measurable negative impact. The opposition has been reduced to claiming that the impact will be a slow burn. Secondly there has been the agreement with the opposition over off-shore processing of asylum seekers. While this has been unpopular in many circles and opposed by the Greens it is at least some sort of closure.

Finally the government has been on the front foot over education funding and national disability insurance. Nothing has been resolved and the potential investments are huge but it has set up a dialogue with state governments that has spread responsibility around.

These Coalition state governments are now clearly in the popular mind. There is little sign of Labor recovery at the state level, but that is not the point. Gone from the media are John Brumby, Kristina Keneally and Anna Bligh. In their place are Ted Baillieu, Barry O'Farrell and Campbell Newman. This is an important turn-around. The Labor governments in NSW and Queensland especially hurt federal Labor.

Any government attracts critics and disappoints groups of voters.

Decisions have to be made. This is the case even when it is business as usual. But the recent state budgets have not been business as usual. They have been budgets that have cut public servant jobs and cut or capped spending in highly sensitive areas like education. They have angered not just the usual adversaries but groups, like private school parents, more inclined to be allies of the Coalition. The heat has now been turned on the Coalition in a way that has advantaged the federal government.

These are my educated guesses. Only time will tell what has generated this apparent Labor revival. We don't really know yet though it appears that the pendulum may just be swinging back.

John Warhurst is an emeritus professor of political science at the Australian National University.

John.Warhurst@anu.edu.au.