Daniel “Dan” Andrews has a problem. The Victorian Opposition Leader, who tried to get everyone to call him Dan in recent election advertising, isn’t very well liked. According to yesterday’s Newspoll current Premier Denis Napthine leads Andrews 47%-34% in the preferred premier stakes. Last week’s Fairfax Ipsos poll isn’t any better — 45% consider Napthine the preferred premier to Andrew’s 36%. Regardless, the unliked and relatively unknown Andrews is on track to be Victoria’s next premier, as he leads a party with a commanding two-party preferred vote of 54%-46% according to Newspoll, or 56%-44% according to Ipsos.

If Andrews wins, it wouldn’t be the first time an unpopular leader has led a party to victory. The metric has a history of throwing up misleading and counter-intuitive election commentary, frequently being in opposition to, or failing to reflect, the depth of the direction of the two-party preferred vote. Kevin Rudd had a lead on Tony Abbott in the preferred prime minister stakes right up until the last few polls before the election, even though Abbott’s Liberal opposition was ahead on the two-party preferred vote the whole time. It was a similar case when Julia Gillard was in power — she was more personally popular despite floundering in two-party preferred polling. And in 2007, many commentators argued John Howard still had a chance because he was only marginally behind Kevin Rudd as preferred prime minister — needless to say the election result proved otherwise. Going back even further, Kim Beazley led Howard on preferred prime minister in the 1998 election campaign. Paul Keating was more popular than Howard — that didn’t stop him losing. In 1993, John Hewson was the preferred prime minister over Keating, but, again, that proved a poor indicator of the final result (though last-minute gaffes may have had something to do with that).

So does having a popular leader matter, or does having an unpopular one hurt? Sure, to some extent, but the popularity metric is overblown in importance in the Australian press, ABC election guru Antony Green tells Crikey. “It always astonishes me when newspapers lead with the preferred PM rating. They’ll focus on it sometimes even if it’s at odds with the two-party preferred vote. But that’s by far the more important one.”

The metric always benefits the incumbent, to a far greater extent than the two-party preferred vote, says psephologist Kevin Bonham.”The opposition leader will usually take the lead only if the two-party preferred heavily favours the opposition,” he said. Incumbency will generally get you around 16 points on the preferred leader rating federally, while the benefits of incumbency are less pronounced but still significant at a state level.

“I think what happens is that if people like the PM and aren’t sure about the opposition leader, they say the PM is their preferred leader. But if they dislike the PM and aren’t sure about the opposition leader, they will often say they’re not sure,” Bonham said. It’s harder for state leaders, he adds — no one tends to ever know much about the local opposition leader. And Napthine is doing better than you’d expect given his two-party preferred score.

The preferred prime minister score can be useful between election periods, when people aren’t thinking much about their vote in an upcoming election. In those times, it can be a leading indicator, Green says. “Kevin Rudd’s popularity started to fall before the government vote started to fall,” he says. But he strongly cautions against overanalysing the popularity vote. “If you want to know how people will vote, ask them how they will vote,” he said.

Some systems, like the presidential system in the United States, see votes strongly influenced by the popularity of the leaders. But Australia’s party system is very strong, and remains the greatest determinant of how people will vote. “The fundamental reason people vote the way they do in this country is party allegiances or attitudes to parties,” Green said. “Two-thirds to three-quarters of people will vote that way. Beyond that, the next thing is party leadership, and beyond that, you have local issues and candidates. But overwhelmingly people vote on parties. So lots of people who say they like the prime minister will never, ever consider voting for them, even if they like the prime minister more than the opposition leader.”

“So the preferred leader score is a useful bit of information,” Green concluded. “But it doesn’t tell you much about who’ll win the election.”