Gender Gap in Natural Language Processing Research: Disparities in Authorship and Citations

Gender Gap in Natural Language Processing Research: Disparities in Authorship and Citations Saif M. Mohammad National Research Council Canada Ottawa, Canada saif.mohammad@nrc-cnrc.gc.ca Abstract Disparities in authorship and citations across gender can have substantial adverse conse- quences not just on the disadvantaged genders, but also on the field of study as a whole. Mea- suring gender gaps is a crucial step towards ad- dressing them. In this work, we examine fe- male first author percentages and the citations to their papers in Natural Language Process- ing (1965 to 2019). We determine aggregate- level statistics using an existing manually cu- rated author–gender list as well as first names strongly associated with a gender. We find that only about 29% of first authors are female and only about 25% of last authors are female. No- tably, this percentage has not improved since the mid 2000s. We also show that, on average, female first authors are cited less than male first authors, even when controlling for expe- rience and area of research. Finally, we dis- cuss the ethical considerations involved in au- tomatic demographic analysis. 1 Introduction Gender gaps are quantitative measures of the dis- parities in social, political, intellectual, cultural, or economic success due to one’s gender. Gender gaps can also refer to disparities in access to re- sources (such as healthcare and education), which in turn lead to disparities in success. We need to pay attention to gender gaps not only because they are inherently unfair but also because better gender balance leads to higher productivity, better health and well-being, greater economic benefits, better decision making, as well as political and economic stability (Skjelsboek and Smith, 2001; Woetzel et al., 2015; Hakura et al., 2016; Mehta et al., 2017; Gallego and Guti´errez, 2018). Historically, gender has often been considered binary (male and female), immutable (cannot change), and physiological (mapped to biological sex). However, those views have been discredited (Hyde et al., 2019; Richards et al., 2017; Darwin, 2017; Lindsey, 2015; Kessler and McKenna, 1978). Gender is complex, and does not necessarily fall into binary male or female categories (e.g. nonbi- nary people), and also does not necessarily corre- spond to one’s assigned gender at birth. Society has often viewed different gender groups differently, imposing unequal social and power structures (Lindsey, 2015). The World Economic Forum’s 2018 Global Gender Gap Report (which examined data from more than 144 countries) high- lighted the gender gap between men and women in Artificial Intelligence as particularly alarming (WEF, 2018). It indicated that only 22% of the professionals in AI are women and that this low representation in a transformative field requires urgent action—otherwise, the AI gap has the po- tential to widen other gender gaps. Other studies have identified substantial gender gaps in science (H˚akanson, 2005; Larivi`ere et al., 2013; King et al., 2017; Andersen and Nielsen, 2018). Perez (2019) discusses, through numerous ex- amples, how there is a considerable lack of disag- gregated data for women and how that is directly leading to negative outcomes …

Content truncated to 500 words