Exploring the marginalisation of women in popular cinema using quantitative network tools

Lead Research Organisation: University of Manchester
Department Name: Social Sciences


The aim of this fellowship is to use statistical tools to explore the gender biases in mainstream Hollywood cinema. There is consistent year-on-year evidence that women occupy only one in three speaking roles in popular cinema. Moreover, scholarship on post-feminism has highlighted the need for analyses which move beyond discussion of the presence of a "strong female character" within a narrative, and consider the presence of female characters in the context of their relational positioning among the other characters in the narrative. Without this relational focus, debates remain limited to individualistic notions of empowerment that depoliticise women's narratives, and thus frequently overlook how Hollywood narratives isolate female characters amongst men and deny women opportunities for collective empowerment. The importance of this issue is clearer than ever given the waves that the #MeToo movement has made and continues to make, a movement which has highlighted both the severe ingrained issues related to gender and power in Hollywood, as well as the radical potential that comes from female voices and stories that are collective rather than individual. The fellowship will build on the platform of my doctoral research in order to produce a number of publications and research outputs which will disseminate key findings and arguments from my novel data on the gendered distribution of dialogue in blockbuster cinema to diverse audiences. In particular, I aim to use emerging statistical tools including relational event models and entropy-based approaches to multivariate network data to analyse gendered patterns of interaction in these narratives. These tools allow us to explore questions surrounding not only how much women speak, but also how often they speak with other women, what they speak about, and by what mechanisms and patterns their dialogue is characterised. Alongside these goals, I aim to explore how these findings from Hollywood cinema can be related to the context of the UK film industry, where there is also a strong need for quantitative research addressing the various representational inequalities that persist in the sector despite years of "diversity talk" and efforts at addressing these imbalances.


10 25 50
Description The fellowship enabled achievements in all four objective areas outlined in the research proposal:

Produce publications in order to help establish track record: Two journal articles were prepared, along with one book chapter. The first journal article, which provides an overview for a cultural sociology audience of the use of network analytical approaches for studying narrative texts, was rejected after review and will be resubmitted in a different journal; the second article is still awaiting final input from co-authors, and contains important methodological developments through its implementation of relational event models, as well as valuable substantive findings about the nature of the gender bias in mainstream film narratives; the book chapter is fully written and with the publisher, and extends prior work developed in doctoral work in two important ways by focusing more directly on race and ethnicity as well as gender, and by exploring male homosociality. Together, these publications advance the work developed in my doctoral research and extend its application, reach and utility to the wider academic community. Furthermore, key tools related to the work and the broader approach of studying narratives as networks were released as an R package called charinet, providing a first-of-its-kind open access software package for character interaction network analysis.

Build networks to develop impact opportunities and inform and support their further development: I presented my work at the 2021 European Conference of Social Networks, and was a co-author on papers presented at Networks 2021. Most importantly, the fellowship gave me a platform from which to secure a subsequent postdoctoral position working on a large multinational project which is directly policy-facing, and thus has clear pathways to impact in the area of screen sector gender inequalities.

Further training to improve their research and related skills: I completed many hours of online training in statistical and data science skills via Datacamp. This training and up-skilling directly contributed to my ability to secure a strong postdoctoral research position following the fellowship, and enhanced my employability in the fields of digital humanities and computational social research.

Developing funding proposals: Together with my fellowship mentor, I submitted an ESRC funding proposal for work which would further develop the research line established in the fellowship. The funding proposal was positively reviewed but unsuccessful.
Exploitation Route The software package developed in this fellowship can enable researchers in the social sciences and digital humanities interested in analysing networks as narratives to access key tools for doing so. Researchers can use the package to extract character interactions from screenplays, analyse dynamic network centrality scores, and produce visualisations to aid analysis of texts. In this way, the fellowship has generated tools which can facilitate more research on the relational dynamics of narrative texts and how these are linked with representational patterns and inequalities. Moreover, through one key publication developed in this fellowship, researchers can find a transparent and reproducible implementation of a relational event model for analysing dynamic interaction sequences, which is a methodological contribution that will help make these methods accessible and more widely understood. Academic and non-academic stakeholders interested in the gender inequities in cultural narratives can benefit from the findings of this research, which finds some evidence that women are no less likely to interact with each other than we would expect if there were no gender bias once we account for the opportunities to interact. This suggests that when women are written into the same scene, they are likely to interact, and that the lack of interactions between women in popular narratives is therefore likely stemming from women not being written into enough scenes together and thus not being given opportunities to interact with one another. This finding can be taken forward by policy actors looking to improve on-screen representation outcomes.
Sectors Creative Economy