Australian Financial Review

Lies & Statistics, pg63

Some how-to-vote cards matter more than others, writes Peter Brent.
26 August 2006

Last week's migration bill reminded us that, from time to time, Family First's Steve Fielding holds leverage in the Senate. But come next year's election his party's influence may be more significantly felt in the lower house. This is because of Family First preferences and their "directability".

There are many misconceptions about preferences. It might surprise you to learn that a House of Representatives preference deal between the ALP and the Greens is not worth much to the larger partner, because all the Greens can do is advise voters, using how-to-vote cards, as to who should get preferences; evidence suggests nearly all Green voters (about 96 per cent) ignore the advice.

Many Labor and Liberal voters, on the other hand, are "directable" on preferences, and so were One Nation's when they were a force.

Now, voting data suggests Family First, the new kid on the block, has a significant ability to influence its voters' preferences.

Family First's power base is South Australia, and in the state election this year it ran in 45 of the 47 electorates, polling 5.8 per cent.

The SA parliamentary library analysed the preference flows. Family First directed preferences to the Liberal Party in 32 seats, and the preferences duly flowed the Liberals' way, 65 per cent to 35 per cent for Labor. In six electorates Family First preferenced neither side but the Liberals still got most of the flow: 60 per cent to Labor's 40 per cent.

So far so good - we'd expect most Family First supporters to be conservative. However, in seven seats Family First preferenced the ALP, and the turnaround was dramatic: 61 per cent of votes now flowed to Labor.

That's a difference of 26 per cent between a Family First Liberal deal and a Family First Labor one.

We must take care not to overinterpret with these numbers. In particular, cause and effect are probably conflating - whatever influenced Family First to make exceptions for the seven Labor candidates (their positions on social issues, perhaps) would probably have led some voters to preference them in any case, with or without the party's imprimatur.

And extrapolating data from a particular time and place to another can be problematic. But this is what we will now do: apply that 26 per cent to the most recent federal election.

In 2004, Family First ran in 109 of 150 electorates, preferencing Labor in none. Family First's total national vote was only 2.1 per cent, but even so we can definitely say this: the coalition won four seats that would have gone to Labor had they received 26 per cent more of Family First's preferences. And four seats is a lot in a close election.

What we can't say categorically is that a Labor Family First deal would have led to this [subbed to 'the same'] outcome. But it does appear that this minor party's voters are relatively obedient how-to-vote card followers.

You can bet the major party strategists have taken note.

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