What's in a number? How (and why) measuring research productivity in different ways changes the gender gap

A persistent finding in studies of research productivity is the 'gender gap', where men seem to publish more academic research than women. However, this gap varies widely from study to study, and little has been done to explore how these claims might be sensitive to what is being measured and how. Using a dataset of publications statistics spanning five years for a Norwegian social science research institute, this paper looks at how (and why) measuring productivity in different ways provides different pictures of the gender gap. Based on the situated context of the institute, we also disaggregate the data by staff category, methodological orientation, and language background, and consider the impact of leaves of absence.We find widely varying measures of the gender gap depending on how we measure and disaggregate, and argue that different bibliometric indicators capture different aspects of research performance, including diversity of output and collaboration, which reflect different publication practices that are both gendered and situated. We suggest that looking at academic writing as a situated - and gendered - social practice offers a potential for deriving more theoretically consistent explanations for both the seeming persistence of the gender gap and the wide contextual variations.