«Sex differences in research funding, productivity ´ and impact: an analysis of Quebec university professors ` ´ Vincent Lariviere • Etienne ...»
Sex differences in research funding, productivity
and impact: an analysis of Quebec university professors
Vincent Lariviere • Etienne Vignola-Gagne • Christian Villeneuve •
Pascal Gelinas • Yves Gingras
Received: 14 October 2010
Ó Akademiai Kiado, Budapest, Hungary 2011
Using the entire population of professors at universities in the province of
Quebec (Canada), this article analyzes the relationship between sex and research funding, publication rates, and scientiﬁc impact. Since age is an important factor in research and the population pyramids of men and women are different, the role of age is also analyzed. The article shows that, after they have passed the age of about 38, women receive, on average, less funding for research than men, are generally less productive in terms of publications, and are at a slight disadvantage in terms of the scientiﬁc impact (measured by citations) of their publications. Various explanations for these differences are suggested, such as the more restricted collaboration networks of women, motherhood and the accompanying division of labour, women’s rank within the hierarchy of the scientiﬁc community and access to resources as well as their choice of research topics and level of specialization.
V. Lariviere (&) Á Y. Gingras ` Observatoire des Sciences et des Technologies (OST), Centre Interuniversitaire de Recherche sur la ´ ´ ` ´ Science et la Technologie (CIRST), Universite du Quebec a Montreal, CP 8888, Succ. Centre-ville, Montreal, QC H3C 3P8, Canada e-mail: firstname.lastname@example.org ` V. Lariviere School of Library and Information Science, Indiana University, Bloomington, IN, USA ´ E. Vignola-Gagne Fraunhofer Institute for Systems and Innovation Research (ISI), Karlsruhe, Germany ´ E. Vignola-Gagne Life-Science-Governance research platform, Department of Political Science, University of Vienna, Vienna, Austria C. Villeneuve Direction de l’Analyse et de la Recherche Institutionnelle, ´ ´ Universite du Quebec (UQSS), Quebec, QC, Canada ´ P. Gelinas ´ ` ´ Direction des Politiques et Analyses, Ministere du Developpement Economique, de l’Innovation et de l’Exportation, Quebec, QC, Canada ` V. Lariviere et al.
Keywords Sex Á Research funding Á Research productivity Á Research impact Á Colla
Introduction When the ﬁrst woman to receive a Masters’ degree from McGill University, Harriet Brooks, got married in London in 1907, a brilliant research career came to an end. Since women of that era were often forced to leave their jobs after marriage, Brooks was forced—though not without a battle with the administration of Barnard College at Columbia University—to abandoning her position as a young professor (Rayner-Canham and Rayner-Canham 1992; Rossiter 1982). Before getting married, this atomic physics specialist had been mentored by none other than Ernest Rutherford and Marie Curie, with whom she worked on—but never completed—her doctoral studies.
Although women in Western societies in general, and within the scientiﬁc community in particular, have made great strides since Harriet Brooks, several studies have demonstrated systematic differences between the sexes within scientiﬁc and technological ´ ﬁelds, and within the research community as a whole, both in the province of Quebec ´ (Conseil de la science et de la technologie du Quebec (CST) 1986; Heap and Sissons ´ 2010; Lasvergnas-Gremy 1984) and elsewhere (Xie and Shauman 2003; Zuckerman et al.
1991). Despite the fact that there is an increasing proportion of female professors in
´ ´ ´ ´Quebec (Conference des recteurs et des principaux des universites du Quebec (CREPUQ) 2010), one may still ask whether or not this progress in terms of workforce composition ´ has led to a greater presence in the research sphere. Based on the entire set of Quebec professors, the present study analyzes the correlations between sex and research funding, publication rate, as well as scientiﬁc impact. Since age is an important factor in research (Feist 2006; Gingras et al. 2008; Simonton 2004) and the population pyramids of men and women are different (see Fig. 1b), the data presented here are also broken down according
to the researchers’ ages. Furthermore, we analyze trends in each of the three broad ﬁelds:
health sciences, natural sciences and engineering (NSE) and social sciences and humanities (SSH).
After surveying the main published studies pertaining to the place of women in the scientiﬁc community, we present the sources of data and methods employed, followed by our main results. The discussion presents the various interpretations that may help
Fig. 1 a The number of researchers, broken down according to gender and ﬁeld, b the distribution of researchers according to gender and date of birth Sex differences in research funding, productivity and impact elucidate the trends observed while the conclusion reﬂects on the possibility of future changes in the basic values that underlie the present hierarchy of disciplines which favors male contributions to science.
Literature review A survey of the vast majority of studies completed since the 1990s clearly shows a gap of approximately 30% in research productivity between men and women, as measured through the number of publications. In other words, women publish between 70 and 80% as many articles as men (Fox 2005; Prpic 2002; Scheibinger 2003; Xie and Shauman 1998, 2003). This is a marked improvement over previous disparities: Zuckerman’s review (Zuckerman 1991) found that women published, on average, 40 to 50% fewer articles than men. Results for the cases of the United States (Etzkowitz et al. 2000; Fox 2005; Leahey 2007; Xie and Shauman 2003), Canada (Nahkaie 2002) and elsewhere in the world ´ (Bordons et al. 2003; Gonzalez-Brambila and Veloso 2007; Mauleon and Bordons 2006;
Prpic 2002) have been similar, considering both science as a whole and individual scientiﬁc disciplines.
Results from the existing literature are more nuanced when it comes to comparing the scientiﬁc impact of men’s and women’s work. Some studies have indicated similar levels of impact of men and women’s publications (Bordons et al. 2003; Gonzalez-Brambila and ´ Veloso 2007; Long and Fox 1995; Mauleon and Bordons 2006; Zuckerman 1991 citing Cole and Zuckerman 1984) and even, occasionally, a higher impact of women in certain scientiﬁc disciplines (Long 1992; Borrego et al. 2010). Other studies have shown that women’s patents had a higher impact than men’s (Bunker Whittington and Smith-Doerr 2005). These studies give credence to the often invoked hypothesis that women focus more on research quality, while men focus on the quantity of publications (Sonnert and Holton 1995). Another set of studies showed that articles written by women obtain, on average, ˜ fewer citations than those of their male counterparts (Penas and Willett 2006; Turner and Mairesse 2005) or take longer to reach their maximum number of citations (Ward et al.
1992). Bordons et al. (2003) were able to estimate the scientiﬁc impact of research (via the impact factor of journals) for three groups of Spanish researchers, ﬁnding similar impacts between men’s and women’s work in two of the three disciplines examined.
In terms of research funding, Stack (2004) showed that a smaller proportion of women receive ﬁnancial support (mainly through research grants): 37.7% compared with 43.3% for men. Similarly, Feldt (1986) found that male adjunct professors from the University of Michigan received more money for their laboratories and had access to better installations than their female colleagues. Similar results were obtained in an analysis of the status of women faculty at MIT (MIT 1999). Based on National Science Foundation (NSF) and National Institutes of Health (NIH) data, Fox (1991, p. 202, quoting Zuckerman 1987) concluded that both sexes receive a number of grants proportional to the number of proposals submitted, which was thus posited as the source of the observed disparity.
Sources and methods
containing the required information. The list of university1 and clinical researchers used in `re ´veloppement e ´conomique, this study (N = 13,636) was obtained from the Ministe du de ´bec (MDEIE) and the three provincial granting de l’innovation et de l’exportation du Que councils.2 In addition to their date of birth and sex, each individual in the list was ascribed a broad ﬁeld of research (health, NSE or SSH), based on their respective departments and the nature of their research.
` ´ Data on research funding came from the Systeme d’information sur la recherche uni´ ´ versitaire (SIRU maintained by the Ministere de l’education du loisir et du sport du Quebec (MELS)). The SIRU database includes grants awarded by the various granting councils as well as scientists’ research contracts, and was compiled for the 2000–2008 period.
Research projects involving many universities were attributed to each institution based on its fraction of professors involved, instead of simply assigning each project to the principal investigator. We have also excluded grants from the Canadian Foundation for Innovation and its provincial counterparts, since the majority of these grants are for infrastructure and are not directly linked to the research project itself.3 Finally, in order to keep only professors who are involved in research, only those having obtained some type of funding, irrespective of its source—government, industry, etc.—at least once during the 2000–2008 period (N = 9,074) were considered. When limiting the analysis to professors whose age could be determined the size of our sample reduced to 7,064.
Bibliometric data on scientiﬁc publications were found from the Thomson Reuters Web of Science (WoS), which annually indexes the articles published in approximately 11,000 journals across all disciplines within health, NSE and SSH. Although this database indexes several different types of documents (journal articles, letters to the editor, reviews, etc.), only articles and review articles are considered here, since they are generally accepted as the main instruments for communicating original research (Carpenter and Narin 1980;
Moed 1996). The WoS does not, however, cover all work published by researchers from ´ Quebec (or anywhere else, for that matter), since some are disseminated through nonindexed national journals, or other types of documents such as conference proceedings, grey literature and books. WoS limitations affect our examination of SSH in particular: the objects of SSH research tend to be more ‘local’ in nature and a larger proportion of their publications thus tend to be in local or national journals. These limitations are ampliﬁed in the case of non-English-speaking countries (Archambault et al. 2006). In addition, SSH researchers publish more books and book chapters than their colleagues in health or NSE ` (Lariviere et al. 2006), which translates into a lower coverage of their scientiﬁc production within WoS.
Attributing articles to a given university researcher from our list is a more complex process, since there is no unique code associated with each individual within the WoS.
´ ´ There are 15 universities in Quebec: Bishop’s University, Concordia University, Universite Laval, Uni´ ´ ´ ´ ´ ´ ` ´ versite McGill, Universite de Montreal, Universite de Sherbrooke, Universite du Quebec a Montreal, ´ ´ ` ` ´ ´ ` ´ ´ ` Universite du Quebec a Trois-Rivieres, Universite du Quebec a Chicoutimi, Universite du Quebec a
´ ´ ´ ´ ´Rimouski, Universite du Quebec en Outaouais, Universite du Quebec en Abitibi-Temiscamingue, Institut ´ ´ national de la recherche scientiﬁque, Ecole nationale d’administration publique, Ecole de technologie ´ superieure.
´ ´bec (FRSQ), Fonds que ´cois de recherche sur la socie ´ et la ´be ´te Fonds de la recherche en sante du Que culture (FQRSC) and Fonds que ´cois de la recherche sur la nature et les technologies (FQRNT).
´be Similarly, we also sought to limit the impact of other types of infrastructure grants not explicitly indicated as such (unlike those from the CFI) and assigned to a single researcher, but which, in fact beneﬁt an entire research group. We have therefore excluded researchers whose funding, for a given year, was greater than three times the standard deviation of the distribution of all funding received in a year.
Sex differences in research funding, productivity and impact
Therefore, given the large number of duplicate names, in order to correctly attribute articles to researchers, each paper with the ‘correct’ surname and initial had to be manually and individually validated.4 After this validation process, at least one article was successfully attributed to 8,485 Quebec researchers, a number which reduces to 6,231 when ﬁltered according to the availability of age data (as discussed above).
As shown in Fig. 1a, the percentage of women in each ﬁeld differs considerably: 36% in SSH, 30% in health and only 14% in NSE. In fact, the larger proportion of women in the social sciences is directly linked to the strong growth of these disciplines in the 1960s, attracting more women than men (Warren and Gingras 2007). In addition, health and SSH disciplines are often characterized by a greater focus on ‘care’, which is known to play a role in attracting women to certain career paths (Cockburn 1988; Collin 1986; Witz 1992).
The age distribution of men and women also differs considerably: while women account for about 40% of professors under 35 years of age, they represent less than 20% of those over 60 years of age. Women are also, on average, 3 years younger than men (CREPUQ 2010), and this tendency is visible in all three groups of disciplines under study here. More speciﬁcally, the average birth year of men is 1954 compared to 1957 for women in health sciences, 1954 compared to 1957 in SSH, and is 1955 compared to 1960 in NSE. On the whole, we see here that the age difference between men and women is quite similar in SSH and health sciences (women are about 3 years younger), but is greater in NSE where women are younger by about 5 years.