Loyalties melt in election campaign heat

By AUDREY YOUNG

Almost a third of voters changed their minds about which party they would vote for during the last election campaign, a detailed study has found.

Almost half made up their minds during the campaign, and 12 per cent on election day.

The results have been reported in a new book, Proportional Representation on Trial, published b y the Auckland University Press and based on the research of Jack Vowles, Peter Aimer, Jeffrey Karp, Susan Banducci, Raymond Miller and Ann Sullivan.

"The campaign period is now more critical as people's voting choices have become more volatile from one election to the next," says Dr Vowles in a chapter titled "Did the Campaign Matter?"

He concludes that the campaign is potentially more important under MMP than First Past the Post.

Under MMP, a party's ability to cross the 5 per cent threshold can decisively affect the overall outcome.

The voters were interviewed daily throughout the campaign about their intentions and on how they voted. About 2400 took part in the survey, part of the 1999 New Zealand Election Study.

Of the 48 per cent who indicated they had made up their minds during the four-week campaign, 14 per cent said they decided in the last week and 12 per cent decided on election day.

The study shows that 31 per cent of the respondents cast a vote that did not match the intention they expressed during the campaign.

Five per cent of National's election-day vote came from people who, during the first half of the campaign, had intended to vote Labour.

Labour was better at converting voters - 9.4 per cent of its election-day voters had intended during the first half of the campaign to vote National.

The researchers say the two main reasons for switching allegiance are changing perceptions of party leaders or a party becoming favoured on an issue of concern to particular voters.

Labour had the advantage on education and health.

National's best issues were the economy and taxes.

Dr Vowels concludes that with the exception of Coromandel, there was no evidence that the results of public opinion polls had significant effects on voter choices.