The emphasis on the relationship between Tony Abbott and women is fuelled by opinion polls that consistently show that he is less popular with women voters. Newspoll has him trailing Julia Gillard among women as better prime minister by 41–34 per cent. The same poll showed Gillard trailing Abbott among men by 42–36 per cent.
But gender is only one factor. Public opinion polls and academic voting studies reveal many ways in which support for political parties and leaders varies according to demographic categories. Other common newspaper poll categories include age, state and urban-rural location. Academic studies add other categories such as religious denomination and church attendance.
Some of these variations are fairly stable over time and don't vary much according to party leader. For instance rural voters, especially farmers, traditionally support the Coalition parties disproportionately.
The fact that Abbott and Gillard are now the leaders offers voters more variation than ever before with one of the leaders being a woman. In other ways, too, including religious belief, there are differences. Although both are city representatives of similar age, Abbott is from Sydney and Gillard is from Melbourne. All of these differences may produce variations in voter support for the leaders and for the parties they lead.
Abbott's relative unpopularity with women does appear to have a personal component, even though women voters have become increasingly attracted to Labor since 30 years ago when it was Labor not Abbott that suffered from a negative gender gap.
There are plausible reasons for his apparent 'gender trouble'. His personal style has long been characterised as macho and aggressive, a style that many women voters do not find attractive. Well before he became leader his role was that of his party's head-kicker. His personal policy preferences include emphatic opposition to abortion, a position that is not popular with many feminist women, but which has not become his party's policy.
It is plausible that this combination has contributed to his 'gender gap' with women voters. Each of these attributes are popular with some women, it is true, but probably not with a majority of women.
However, undue concentration on the role of gender in shaping voter choice serves to downplay other very interesting variations.
Age is one. Gillard is much more popular than Abbott among younger voters. The same recent Newspoll survey showed Gillard was rated better PM over Abbott among 18–34-year-olds by 41–35 per cent (almost the same as her gender advantage). Similarly Abbott was ahead of Gillard as better PM among those 50 and older by 40–36 per cent. Age is as big a factor as gender.
The same is true of the location of voters. The same Newspoll reported that in the five state capital cities excluding Hobart, Gillard led Abbott by a margin of 41–36 per cent. Similarly among non-capital city voters Abbott leads Gillard by about the same margin, in this case 41–34 per cent.
Finally, in Victoria, her home state, Gillard led Abbott by a huge 45–33 per cent. But in Abbott's home state of NSW, he led Gillard by a much smaller 39–38 per cent.
Some of these variations are no surprise and are part of the accepted wisdom of voter studies. Older voters and rural voters are disproportionately more politically conservative than younger voters.
These statistical differences may be less personal than gender and thus attract less attention. It is the duelling personalities of Abbott and Gillard that contribute especially to the focus on gender. But ultimately these other characteristics will be equally important in determining the outcome of the next federal election.
Gillard and Abbott should do all in their power to increase their popularity in the voting segments in which they are weaker. For Gillard her weakness lies among older, rural, male voters. For Abbott his weakness lies with younger, metropolitan, women voters.
John Warhurst is an Emeritus Professor of Political Science at the Australian National University and a columnist with the Canberra Times.