KNOWN to practitioners as "the quant" and "the qual", they are the yin and yang of the modern research that political parties use to target election campaigns and hone messages to voters.

Quantitative research relies on numbers. Large, statistically representative samples of voters from all over the country are surveyed by telephone for their opinions and voting intentions.

Qualitative research relies on language. A moderator gathers about a dozen or so people for a free-ranging group discussion to tease out their feelings and attitudes.

In the lead-up to last year's federal election, John Utting's UMR did Labor's quantitative research while Tony Mitchel- more's Visibility ran the focus groups.

Rather than conducting national polls like those published in the news media, UMR tracked voting intentions in about 20 electorates. On each of the last 40 days of campaigning, it questioned a "rolling sample" of 1500 voters from these target electorates. A one-page summary of results went daily to Kevin Rudd.

Meanwhile, Mitchelmore travelled around the country carrying out focus group discussions with people selected because they were swinging voters. He wanted strategic insights into their thinking and the phrases that sprang to their lips when they pondered bread-and-butter issues and broader aspirations.

Opinions differ on the merits of each method.

One experienced consultant says political parties always rely on hard numbers when deciding where to spend their money, and how much. "Focus groups help develop a theme but it is the quant that tests it and determines the strategy," the consultant said.

But, says another veteran political strategist: "Numbers can help you see whether it's worthwhile going hard in particular seats and lay out the sphere of possibilities, but they can't really help you in a strategic way. The person most valuable to a political leader at 5am every morning is a great qualitative researcher."

What makes a good focus group researcher? An academic expert, Richard Carson, from the University of California, San Diego, says the challenge is to work out what people want to say without leading them in the client's direction.

It's not about extracting the opinions of the articulate and confident, but of "ordinary people who often are not very conversant but care about the issues you are talking about", says Carson, who is visiting the University of Technology, Sydney.

Mark Davis