Doctor of Philosophy
School of the Arts, English and Media
Parker, Rhiannon B., The representation and production of visual gender bias in anatomy images and its effects on student attitudes, Doctor of Philosophy thesis, School of the Arts, English and Media, University of Wollongong, 2016. https://ro.uow.edu.au/theses/4873
Background: Gender bias in medical discourse is persistent and influences the way medical professionals treat and diagnose patients, and thus has a meaningful impact on health. Medical education provides a pathway for improving the attitudes, beliefs and knowledge that future healthcare practitioners hold about gender. However, it also has the ability to contribute to the construction of gender bias. Research on the role that images, a critical part of medical education curricula, play in contributing to gender bias has been limited. Further, no research on images from medical education has adopted a framework that distinguishes between the sites of the image, its production and its audience. An account of these three sites provides a comprehensive understanding of not only what messages are portrayed in medical images but also of how these messages are constructed and perceived.
Aims: The purpose of this thesis was to examine the visual representations of gender in medical education through the medium of anatomical textbooks. The aims were threefold: a) to determine how gender bias is visually represented in anatomical textbooks; b) to investigate how the context of image creation influences the construction of gender bias in medical illustrations; and c) to determine what effects gender-biased images have on the implicit and explicit attitudes of anatomy students.
Methods: This research was comprised of three separate studies and employed a mixed methods approach in order to determine the representation, production and effects of gender bias in medical images. Study 1 included an extensive content analysis on the representation of gender in 6004 narrative and conceptual images from anatomical textbooks used in Australian Medical Schools. The study focused not just on gender ratios but also on other forms of gender bias that might be present. In Study 2, 83 illustrators from the Association of Medical Illustrators and the Medical Illustration Sourcebook participated in an online interview that gathered data in relation to the contextual factors that influence the construction of gender in medical images. Participants provided detailed information on their experiences with clients, reference materials, and their own work, with a particular focus on gender bias. The data were analysed using a mix of quantitative and thematic analysis. In Study 3, a randomised control trial was conducted with 457 students studying anatomy at the University of Wollongong to assess the effect that gender-biased images from anatomical textbooks had on their implicit and explicit attitudes. In this study participants were randomly assigned to a treatment (gender-biased images) or control (non-gendered images) priming task before completing an Implicit Association Test and a questionnaire. The data was analysed using multiple regression.
Results: Study 1 found that females comprise only 36% of all images in anatomy textbooks, but 57% of all sex-specific images. Further, other forms of bias were found to exist in the visualisation of stereotypical gendered emotions, roles and settings; in the lack of ethnic, age and body type diversity; and in the almost complete adherence to a sex/gender binary. Study 2 indicated that multiple levels of proximal and distal context had an influence on how gender was represented in medical images, including the intrapersonal processes of the illustrator, the immediate physical and social context of the image and the broader institutional and societal contexts. A number of contextual themes were also identified in the data including the sexualisation of nudity, the use of gendered stereotypes, the impact of social networks, the limitations with diversity and pathology, the representation of average bodies and the tension between resource accuracy and accessibility. Lastly, Study 3 revealed that, compared with control images, gender-biased images significantly increased students’ implicit gender bias (Cohen’s d = .33) but had almost no impact on explicit attitudes.
Conclusions: This thesis provides evidence of the continued and complex gender bias present in images from anatomical textbooks. The research has revealed that the context of image production is multileveled and that there are a range of factors that influence the inclusion of gender bias in images. The finding that gender-biased images have an impact on the implicit attitudes of future healthcare providers is significant as this can contribute to perpetuation of gender bias in medical discourse and could have negative healthcare outcomes.