One of the less obvious losers in this election looks like being the pollsters, and not for the reasons most people might suspect. It is not because it seems as if their results are inaccurate. There has been a broad consistency between them and, thus far at least, a broad constancy in their findings.

The main exception to this consistency has been the fact that, at times, Newspoll has had a touch of the Morgans. The Morgan poll results over the years had a tendency to mirror the volatility of the personality of its principal Gary Morgan. In recent times, Newspoll has started to emulate this volatility, sometimes in counter-intuitive ways. Newspoll had Labor’s support increasing in the first week when most commentators thought the coalition did well; then, when most scored the campaign’s second week to Labor, Newspoll recorded a substantial swing back to the government. No similarly sized movements were apparent in the other polls.

All newspapers have a vested interest in magnifying the importance of the expensive polls they’ve commissioned, and there is the usual tendency to wring as much newsworthiness as they can out of them and to gain mentions in other media. This has led, as usual, to the magnification of the importance of very small differences that are well within the usual range of sampling fluctuations.

The Newspoll results released on 6 November found all primary votes unchanged, except a one percentage point drop in the Labor vote and the same increase in the Greens’ vote. Newspoll translated this into a change in the two-party preferred vote to 53-47, a movement towards the opposition. The pollster distributed the second preferences of the 11 per cent who didn’t intend voting for the two major parties 6-5, or roughly 55-45, to Labor, whereas in the last election it was just over 60-40. In an earlier poll, Newspoll had done the rounding the other way, 7-4, or around 65-35 for Labor.

In other words, this apparent movement to the coalition didn’t even reflect sampling fluctuations. It was an outcome of borderline decisions about how to round results into whole numbers. The Australian reported the essentially static poll in a properly cautious way, but almost universally in the broadcasting media the headline reports hailed the poll as good news for the government and evidence that it was making up ground.

Apart from such lapses, it is not accuracy but a lack of penetration that is the pollsters’ problem in the current campaign. They use their usual bag of tricks with apparent competence. Apart from voting intentions we know about preferred prime minister, the approval ratings of the leaders, and who people expect to win the election.

These findings have some interest and value. It is interesting, for example, that compared with Howard’s defeat of Keating in 1996, when disapproval of both leaders was high, this time approval of both leaders is relatively high. It also cements the impression of Labor’s lead in the voting intentions question that a majority prefer Rudd as prime minister and also expect a Labor victory. These suggest at least a small degree of inoculation against some late change.

It is in two other areas that the polls are failing to penetrate.

The first is that they have never reported in a revealing way about the link between policy preferences and party vote. Although it is a staple of election polling to ask which party is stronger in a variety of policy areas, there is little sophistication in the interpretation or framing of these questions. They are reported like a scoreboard as if 50 per cent or a lead over the other major party constitutes a victory. They give a simple headline - typically, Labor ahead on environment, health and education and the Coalition on economic management and national security.

The accompanying stories never comment on two factors which contribute to these findings. First of all, people who favour one party or the other tend to extend a “halo effect” to individual policy areas, influencing their perception of that party’s performance across the board. Meanwhile, each party has a stereotyped image, which means that a large proportion of people tend to attribute each with contrasting strengths.

There are fascinating and important issues in exploring the connection between policy preferences and party vote, but the reporting in the press rarely illuminates their dynamic.

A peak of misleading scoreboard reporting was Malcolm Farr and the Daily Telegraph’s front page story in anticipation of an interest rate rise. On Monday 5 November their headline was “Not to blame,” as comforting for the government as it was misleading about the electoral implications.

The Galaxy poll offered a choice from four reasons for a rate rise - international factors, the Australian economy, the Reserve Bank and the prime minister, hardly a logically watertight or mutually exclusive list of possible causes. It also highlighted the fact that 49 per cent did not think Mr Howard had knowingly misled the public on interest rates three years ago, while 42 per cent thought he had. This too was interpreted as if it were a victory for Howard. Outside the Murdoch press few would believe that this gets to the core of any likely electoral impact of the rate rise announced two days later.

If only a fraction of these 42 per cent who thought that Howard deliberately misled them or the 20 per cent who blame him for interest rate rises changed their vote as a result it constitutes a terrible electoral albatross for the government. Moreover such adverse developments weave their way into the tapestry of political perceptions in ways that these questions are too crude to capture.

In the current election, the pollsters’ standard arsenal of questions has a particularly profound ring of irrelevance. None of the polls seems to give any insight into why such a crucially large section of the electorate seems to have turned its back on the government, and seems immune to all its blandishments. None of the questions they ask seems to capture key elements: whether there is an “it’s time” factor, whether people no longer trust the government, whether people are simply sick of the key personalities. The pollsters do not seem capable of coming up with new questions or approaches that give a sense of the dynamics of the swing to Labor.

At least so far in the campaign, the central conundrum is that news coverage has gone at least as much the government’s way as it could have hoped, but it has not yet gained much traction in the polls. The campaign has had a few high points which have gone to the opposition - the leaders’ debate, the cabinet leak on Kyoto and Tony Abbott’s day of own goals. But most of the time, the daily news agenda has, if anything, danced to the government’s orchestration.

The television news covers the highlights of the day, especially the visually interesting ones, and to some extent allows the parties to frame the debate with a measure of gamesman commentary built in, while the tabloid press is still strongly slanted in favour of the coalition.

Apart from any editorial bias, there are two main reasons for the government’s news ascendancy. The first is that “me too” is inherently less newsworthy. Throughout the campaign, the prime minister’s promises have received more prominence than the opposition leader’s, and especially so when the latter has been covered as simply saying snap.

The second, more sinister one is that one of the keys to the coalition’s success has been their well-honed outrage machine. This has been as evident as ever. The way that the chorus of cabinet ministers has howled at every apparent Labor gaffe, at every off-hand comment as allegedly showing sinister secret plans, has been as sustained as ever. It is still a very successful way of gaining prominent press coverage, much more so than the opposition’s counter claims.

But at least so far this technique does not seem to be gaining great traction with the public. Why not? Has the public heard them all once too often? Where are the polls that might illuminate such fatigue, to show that in key segments there might now be an immunity against what used to be effective attacks?