Gender differences in research productivity : a comparative analysis of Norway and Australia

This thesis is a comparative and empirical analysis of gender differences in research productivity among Norwegian and Australian academics. This study uses the Norwegian and Australian data from the Changing nature of the Academic Profession (CAP) project to examine the size of gender differences in research productivity and the correlates of productivity. Secondary sources are used to indicate how gender differences have changed over time. Research productivity is calculated as “article equivalents”, which is a weighted sum of journal articles (1 point), books edited (2 points) and books authored (5 points). Extensive bi-variate analyses are conducted on each of the hypothesised determinants of research productivity and separate multiple regression analyses are made for men and women in both countries. The major findings from this thesis are that Norwegian women averaged 21 percent fewer article equivalents than Norwegian men, while Australian women averaged 26 percent fewer than Australian men. There is little evidence of an overall reduction in gender differences in research productivity in Norway, but this is partly due to an increase in male research productivity among a small group of prolific publishers. It is far less clear how differences in gender-based research productivity have changed over time in Australia, but female participation in research has risen dramatically since 1993. The multiple regression analysis explains considerably more of the variation in individual research productivity in Australia (R2 = 0.42 women, 0.31 men) than in Norway (R2 = 0.21 women, 0.14 men). The strongest correlate of research productivity across all staff groups is academic rank, which is a particularly strong in the Australian sample given the more hierarchic nature of the Australian academic career structure. International collaboration also exhibits a strong effect size for all staff, while time spent on research is significant for most. The institutional variables included fail to generate large effect sizes or significance. Marital and family statuses also fail to account for gender differences, which may be due to imprecise questioning in the CAP survey