In a forthcoming article on the relationship between candidate gender and electoral success in the British Journal of Political Science, Schwindt-Bayer, Malecki, and Crisp look at how gender affects candidate performance in three contexts where PR-STV systems are employed: the lower houses of Malta and Ireland, and the Australian Senate.
A useful aspect of their research is that it separates the nomination of candidates from their performance; they examine whether ‘once a prospective candidate has decided to enter a race, does the candidate’s gender systematically work against (or for) her?’ (p. 1). They hold an array of individual, party, and constituency-level variables constant in their analysis – including incumbency, electoral experience, and previous party performance in a constituency.
Their findings suggest that PR-STV does not always disadvantage women: in Malta there was no discernible negative or positive relationship between gender and electoral performance, and in Australian Senate races they found that female candidates do slightly better than male candidates. However in Irish elections female candidates receive, on average, 2% fewer first preference votes than their male counterparts. Furthermore, female candidates in Ireland get less of a point bump in vote share for levels of electoral experience or incumbency than male candidates. The authors state that: ‘in Ireland, we found that voters discriminate against women, controlling for all else’ (p.11).
Overall, the effects of voter discrimination on female representation are certainly not as dramatic as the problem of under-nomination. In a context where women have never represented more than 20% of nominated candidates, and where the number of female candidates has actually fallen since the ‘high point’ of nearly 20% in 1997, attention has rightly been focused on increasing the number of nominated female candidates.
The findings are nonetheless rather suggestive – the implication is that PR-STV per se is neither encouraging nor inhibiting of female representation. However,PR-STV may be somewhat inhibitive (compared to, for example, closed lists) for female representation in Ireland where, for whatever reason, female candidates appear to systematically receive fewer votes than their male counterparts.
Irish Politics Forum 
Interesting analysis, Matt. STV’s poor performance on the representation of women has been one of those curious questions. The great proponent of this system, Enid Lakeman, made great play about how STV gives voters the perfect opportunity to vote on a gender basis. The fact that this never worked in Ireland and Malta — the only two STV cases and both bottom of the league in the representation of women — was at best inconvenient, with some arguing that this said more about political culture than the electoral system — the argument being that the problem was that conservative party selectorates were not nominating enough women candidates. And analysis by the likes of Bob Darcy seemed to bear this out. This research you cite does too, but it does add an interesting wrinkle in the Irish case about a conservative tendency among some Irish voters. Interesting.
This is interesting, although I have not yet read the paper and look forward to seeing the data and methods used. It is notable that the findings are more in line with older aggregate level research such as Carty (1980, 96) who found that “the electorate appears reluctant to return women to the Dáil” and Marsh (1987) Gallagher (2003) who concurred with this finding and suggested that “male candidates fare better than women, other things being equal,” (p. 91). On the other hand, Laver, Galligan, and Carney found that “there is no direct prejudice against women candidates on the part of Irish voters” from their analysis of the 1997 election (1999, 122). However, in a more recent paper utilising richer and more nuanced data including the Irish National Election Study , as well as aggregate evidence on background characteristics and election expenditure and crucially the electronic voting returns in Dublin West in 2002, McElroy and Marsh (2010) found no evidence that female candidates were discriminated against. Rather the suggestion was that women were less likely to put themselves forward to run and were less likely to be recruited by the parties.
It might be interesting to test the hypothesis that women, in general, are far too sensible to put themselves in a position where they are used as lobby fodder by the party hierarchies and have to succumb to the mind-numbing task of speeding up the delivery of public services for some of their constituents if they have any hope of being re-elected. I expect most women find they have far more useful, important and life-affirming things to do. I know I’ll probably get into trouble here, but I find that the (relatively few) women who decided to play this “boys’ game” with the objective of winning (e.g., Golda Meir, Margaret Thatcher, Angela Merkel) found it fairly easy to beat their male counterparts and opponents into cocked hats.
I haven’t seen the paper, but it’s hard to see how they can draw any conclusion about STV’s effect on gender representation when all three countries in the study use STV. I suppose if there’s a great deal of variation on the representation of women in these three countries it might suggest that STV is not important, but one would still want to compare levels of representation to those in other countries.
To be fair to the authors, they don’t attribute any causal effect to PR-STV compared to other systems, indeed they state that: ‘it is inappropriate to attribute a universal effect to STV rules – they merely translate voter’s sincere preferences’ (p. 2).
My reasoning was simply that, if we believe the authors’ claims that female candidates systematically receive fewer votes that male candidates in Ireland, then the goal of female representation may be better served by a system that did not allow voters to exercise intra-party choice, such as a closed list.
This is interesting and I’m looking forward to reading the paper. The conflicting evidence in the literature on voter bias is quite surprising. For example, John Lane (1995: 9), looking at all Maltese elections held between 1947 and 1992, concluded that “there is no evidence that Maltese voters have significantly or considerably favoured male over female candidates when they cast their votes and ranked their preferences among candidates of their party”. Lane instead suggested that the problem of the considerably low numbers of women in parliament lay with conservative candidate selectors.