Gender counts : men, women and electoral politics, 1893-1919.
Degree GrantorUniversity of Canterbury
Degree NameMaster of Arts
Gender has seldom been considered in accounts of electoral politics and voting in early twentieth century New Zealand. This thesis approaches the question of gender and electoral politics in three ways. The first is a case-study of the 1893 election campaign in Christchurch based on qualitative data. Gender threaded through both political organisation and debates in this election campaign. Men and women organized separately and invoked gender difference in the discussion of election issues. The second approach is a quantitative study across time and space comparing men's and women's participation rates in general elections from 1893 until 1954. Women's turnout was significantly lower than men's in the 1890s, but the difference had largely disappeared by the late 1940s. Moreover, although broad social changes increased women's participation relative to men's, factors such as party organisation and the nature and content of political debates were also important. The third approach is a statistical analysis comparing men's and women's voting preferences on the liquor issue and for the political parties at electorate level from 1893 until 1919. The analysis is of an ecological nature. It is designed to overcome the absence of individual-level voting data and to limit the ecological fallacy problem which is the error of assuming that relationships evident at the group level reflect relationships at the individual or sub-group level. The thesis reviews and trials five methods for ecological inference: Goodman's ecological regression, King's parametric and non-parametric methods, a semi-parametric method and the homogeneous method. King's non-parametric method is then used to estimate men's and women's support for Liberal, Reform and opposition candidates and for and against prohibition from 1893 until 1919. Significant differences between men's and women's preferences are revealed by the estimates. Together these three approaches indicate that gender was an important factor in election politics of the early twentieth century.