Three political opinion polls were published during the week, and all showed a lift to federal Labor's election chances under Mark Latham.

Well, actually, the polls themselves did no such thing. But why should facts intrude on a political honeymoon?

"Latham gives Labor poll hopes a lift," announced The Sydney Morning Herald's front page. "Voters embrace Latham gamble," claimed The Australian. And "Latham already swaying voters," we learnt from The Sun-Herald.

Let's start with the two released on Tuesday. ACNielsen was published in The Age and Sydney Morning Herald. The previous Nielsen, in late September, had the government in front 52 per cent to 48per cent two party preferred. And this week's? The same: 52 to 48.

Then there was Newspoll, in The Australian. If we ignore (as surely we must) the four days of limbo between Simon Crean's resignation and the Labor caucus leadership vote, Newspoll's final Crean survey registered 50:50 after distribution of preferences.

This week the pollster had the government ahead 51 to 49 per cent. So one survey was status quo, Labor went backwards in the other, but both were written up winningly for the new leader.

The third was a Sun-Herald Taverner poll that surveyed "five crucial western Sydney electorates". The headline was "Latham already swaying voters".

Taverner found primary support for the ALP at 42 per cent and 45 per cent for the Liberals. The paper concluded: "After preference distributions, which always strongly favour Labor federally, the result would comfortably hand the western Sydney politician the keys to the Lodge." The most obvious point about that is that no five electorates, even allegedly crucial western Sydney ones, can be extrapolated to the 150 nationwide.

But the Sun-Herald's conclusion is invalid within any parameters. At the last election, primary support for the Liberal Party in the five seats (Greenway, Lindsay, Lowe, Macarthur and Parramatta) was 44.2per cent, for Labor it was 40.2, and after distribution of preferences, the Liberals led 51.3 to 48.7.

According to Taverner, Labor's vote has gone up by two points and the Liberals' by one. In the absence of minor party data, a reasonable two party preferred estimate rounds to 51 to 49 in the Liberals' favour.

Which shows no movement. No seats change hands. No keys to the Lodge.

All right, we know that opinion polls aren't particularly accurate. It's still early days, and Australians are yet to make up their minds about the new alternative. Latham is outperforming his predecessor on such esoterica as preferred prime minister and approval rating.

But come election time, the Labor Party will need votes - two party preferred ones, preferably in marginal seats.

Both Alexander Downer (1994) and John Howard (1995) achieved immediate and huge (approaching 10per cent) vote leaps in their first Newspolls. Latham got no bounce at all.

That's what the papers should be saying.

Peter Brent is editor of mumble.com.au, a website that looks at electoral behaviour.