What a scientist looks like: Portraying gender in the scientific media

Women’s representation in science, technology, engineering, and medicine (STEM) has lagged despite an increase in the number of women graduating from these programs (Coleborn 2014). Described as a leaky pipeline (Goulden et al. 2009), the passage from receiving a PhD to attaining tenure is markedly harder for women. Women in science are paid less (Moss-Racusin et al. 2012) and are less likely to be cited (Larivière et al. 2013) or recommended for peer review (Lerback and Hanson 2015) compared with men. Despite a growing recognition of this disparity by society, and efforts towards affirmative action, this bias persists. Are women self-selecting out of the pipeline?

Social learning theory suggests that new patterns of behaviour can be acquired by observing the behaviour of others (Bandura 1971). Scientists are strongly influenced by scientific culture in their institutions and abroad. Media represents an important source of role models for scientists (Chimba and Kitzinger 2010). However, role models for women scientists in the media have been problematic, as women are either lacking (Miller et al. 2015) or portray traditionally “feminine” roles that emphasize motherhood or sexuality (Chimba and Kitzinger 2010). Media representations of women in science inform the nascent identity-building of young women—and this identity is central to their ultimate career choices (Mendick and Moreau 2013).

Elite journals such as Science and Nature carry significant weight in their ability to influence the scientific culture, including gender stereotypes. These journals feature a range of articles that reflect what scientists value, who they are, and even what they look like. These journals provide a weekly snapshot into scientific culture.

We were interested in what this snapshot looked like with respect to gender representation. Are influential journals perpetuating a biased image of the scientist? It has been suggested that the lack of women in STEM is partly attributed to the lack of female role models (Coleborn 2014). Science journals provide a model of the archetypal scientist—if that archetype is predominantly masculine, then the scientific culture itself may be contributing to gender disparity.

Here, we asked whether there was a gender bias in the representation of women and men within the pages of Science and Nature. We examined three months of journal issues and found unequal proportions of male and female representation in authorship, content, and visibility in photographs.

Although Science and Nature routinely publish analyses and general interest articles about gender and STEM, our results suggest that the journals themselves may inadvertently perpetuate a biased image of what constitutes a scientist.

The bias is particularly dramatic for corresponding authors in the scientific literature. It is tempting to ascribe the lack of women as corresponding author to gender disparity among faculty members in the institutions themselves (Bain and Cummings 2000; Sheltzer and Smith 2014). Simply put, there are not as many women who could assume the role of principal investigator. However, although women are underrepresented among faculty members, their numbers in Canada are higher than what is portrayed in the journal. For example, in 2009, women comprised up to 30% of full time faculty in STEM at Canadian universities (Hango 2013). This is significantly more than primary authorship would suggest.

Whatever societal hurdles are preventing these women from publishing to the same extent as their male peers, the movement of women into senior academic roles is a slow transition (Surawicz 2016). It behooves journals to ensure equitable gender representation among authors because bias against female authors has been widely documented (Tregenza 2002). Innovative editorial workflows (such as double blind reviews (Budden et al. 2008)) may help improve the visibility of women authors.

The bias among authors of general interest articles is less extreme than primary authorship, which suggests that the journals are doing better at gender parity in the workforce. From the masthead listed on their website, the gender breakdown of editors and staff writers at Nature at the time of writing was approximately 42 women to 27 men. Science does not limit their masthead to writing staff, but also includes technicians, managers, and analysts. The gender ratio of the credited team members was 98 women to 79 men at the time of writing. The fact that more women are on staff at Science and fewer articles have female authors suggests that institutional barriers remain for women getting a piece to print.

Both journals featured significantly more male scientists than female as the subject of their general interest articles. This type of article is highly visible, spanning multiple pages and containing numerous photographs of the featured scientist. Whether or not this bias reflects a bias from the journal in selecting predominantly male candidates, or arises from reported modesty among women scientists, is unknown (Breithaupt 2001; Lerback and Hanson 2015). Whatever the root cause, an equal representation of men and women scientists in the media would help encourage more women to participate in STEM.

This bias was magnified for advertisement features, which featured male scientists almost exclusively in both journals. It is important to note that this is paid content. The intent of these advertisements is to blend in with journal content, and it is difficult for a typical reader to differentiate these ads from true articles (Crook 2004). Such strong bias may reflect the culture of the companies, and not the culture of science itself; yet, this matters little to the average reader. Ultimately, advertising decisions are an important part of the brand identity. Editorial boards have the ability to influence the nature of their advertisements and to demand content that does not alienate half of their readership.

Almost unilaterally, men were the subject of single person photographs in both journals. The only category where a bias was found in favour of women was in photographs associated with advertisements in Science (Fig. 3). The use of women in advertising is not surprising, as women models have historically been used to sell products to both women and men (Bower 2001). A recent study showed that when women were portrayed more often than men in the Spanish media they appeared as “decoration” (González et al. 2017). The preference for women models was not true for Nature. Whether this reflects a cultural divide among British and American cultural norms is not clear. Most of the photographs featuring women in science were some form of self-promotion, so it may be that Science is targeting a female demographic. The bias towards males in advertisement feature photographs reflects the bias in the subject of these articles. Evidence of this bias was more obvious in some issues than others. Issues featuring a spotlight on Chinese researchers typically contained pages of male-dominated advertisement features, causing a dramatic shift in the proportional representation of gender in this subcategory.

The lack of female representation in Nature and Science presented here may have implications for the lack of parity in the STEM workforce. Not only are women scientists being heard less, they are also being seen less. A young woman flipping through a current issue may consciously or subconsciously affirm beliefs science is a male activity (Miller et al. 2015).

The presence of positive female role models can change this belief (Coleborn 2014). Female representation at conferences and in the workplace consistently leads to increased female participation in science (Murphy et al. 2007; Young et al. 2013; Miller et al. 2015; Sardelis and Drew 2016). Conversely, exposure to gender stereotypes decreases women’s self-esteem (Sanchez and Crocker 2005) and job success (Reuben et al. 2014; Schuster and Martiny 2017), and can result in depression (Sanchez and Crocker 2005; Panayiotou and Papageorgiou 2007). As early as six, girls are less likely to believe that members of their own gender can be considered “really, really smart” when compared with the boys in their class and will even begin to avoid activities that they associate with intelligence (Bian et al. 2017, p. 389).

It is the responsibility of publishers to not only ensure that their publications do not promote outdated stereotypes, but they must actively promote equality among sexes within their pages. Many of the explicit biases limiting women’s participation in STEM have been removed, yet parity remains elusive perhaps due to implicit biases among the women themselves (Ceci and Williams 2011). Changing implicit gender stereotypes changes requires “repeated and varied” exposure to counter stereotypes (Miller et al. 2015, p. 639). Journals have an important role in expediting this change—editorial content does not need to wait for the numbers to catch up. Equitable and authentic representation in the scientific media may be the simplest and quickest way to address the “leaky pipeline”.

MMH was funded by NSERC (Discovery Grant).