Mid-career field switches reduce gender disparities in academic publishing

Abstract

The continuing low proportions of women in most STEM fields in many countries is an ongoing concern, with no agreement about the fundamental causes or effective remedial actions. One previous study has found that professional women are more likely to switch from a (not necessarily academic) STEM career than professional women in comparable non-STEM jobs, reducing the overall numbers of STEM women. This study investigates whether the same is true for long term academics, and hence could partly account for current gender disparities. Based on the Scopus subject categories of the first and last five publications 2001–2018 of people in 31 countries with publishing careers starting after 2000, female researchers switching fields mid-career tend to move to fields with fewer women, relative to men switching fields mid-career. Thus, mid-career field changes within academia do not help to explain continuing gender disparities in publishing and other explanations must be sought.

Appendix: Gender differences in academic publishing during the career of Judit Bar-Ilan

Professor Judit Bar-Ilan lived and worked in Israel throughout her career. Within Israel, the proportion of female first-authored articles increased from 26.3% (2.8 male articles per female article) in 1996–98 to 39.6% (1.5 male articles per female article) in 2014–18. The female first authored article proportion increased in all broad fields over time, but at varying rates (Fig. 

Fig. 13

Percent of female first-authored journal articles out of all Scopus-indexed journal articles 1996–2018 with a first author affiliation of Israel, by broad Scopus field

13). There were relatively small increases in Mathematics, Multidisciplinary and in Biochemistry, Genetics and Molecular Biology but much larger increases in others, including more than doubling the proportion of female first authored articles in Business, Management and Accounting, and almost tripling it in Energy. Thus, Prof Bar-Ilan entered a highly gendered profession and experienced a gradual but uneven reducing of gender disparities.

Professor Bar-Ilan’s undergraduate background was in mathematics, which is the least female area in Israel, in terms of publishing. Most areas of mathematics experienced only small increases in the proportions of female first-authored research (Fig. 

Fig. 14

Percent of female first-authored journal articles out of all Scopus-indexed journal articles 1996–2018 with a first author affiliation of Israel, by narrow Scopus mathematics field

14).

Professor Bar-Ilan’s PhD was mathematical computer science, another a male-dominated area. Within computing, only Human–Computer Interaction and Information Systems had strong increases in the proportion of female first-authored articles (Fig. 

Fig. 15

Percent of female first-authored journal articles out of all Scopus-indexed journal articles 1996–2018 with a first author affiliation of Israel, by narrow Scopus computing field

15).

Professor Bar-Ilan moved into library and information science shortly after her PhD, although she continued to conduct mathematical computer science research related to library and information science, with the topic of information retrieval. The social sciences in 1996–18 was one of the most gender-balanced areas and increased to achieve almost gender parity in publishing by 2014–18 (Fig. 16). Library and Information Sciences is about average for gender parity amongst all social science fields in Scopus with a slow increase in the proportion of female first-authored articles.

Fig. 16

Percent of female first-authored journal articles out of all Scopus-indexed journal articles 1996–2018 with a first author affiliation of Israel, by narrow Scopus social science field

In summary, Professor Judit Bar-Ilan physically (in terms of department) and partially intellectually (in terms of publishing) moved from the male dominated subject of Mathematics to more gender balanced Library and Information Sciences (45% female first-authored articles) during her career. Her physical move echoed the overall increase in gender balance of research in her home country, even though the fields that she worked in tended to have static gender compositions, at least in terms of publishing.

Data and methods

Journal article records were extracted from Scopus for all papers with an institutional affiliation of Israel, and articles where the first author affiliation was not Israel were filtered out using the free Webometric Analyst Software.

The first names of the first authors were extracted and submitted to Gender-API.com to estimate their genders. First names reported by Gender-API.com to be used at least 90% of the time by one gender were retained as gendered first names when the result was based on at least 10 gendered records found by Gender-API.com. The remaining first names were not used. Since male and female names may have different degrees of gendering and scarcity, a correction factor was calculated for any systematic bias. This was based on estimating the true proportion of male and female first authors by multiplying the number of articles with each first author name with the percentage of people having that first name for each gender, as reported by Gender-API.com. Totalling these gave an estimate of the total number of male (215,142) and female (118,508) first authors. Correction factors were then calculated by dividing this number by the number found using the gendered first names for males (173,501) and females (88,585), respectively. Thus, each gendered article found was multiplied by 1.240 for males and 1.338 for females, compensating for female names being marginally harder to identify.

Articles were categorised by their Scopus broad and narrow fields. The start year was chosen to be 1996 since the coverage of Scopus expanded in this year, although Prof Bar-Ilan began academic publishing in 1989. Scopus data from before 1996 is of a different character due to the lower coverage and so comparisons with later data could be misleading.