Executive summary
This investigation aimed to uncover potential disparities in the treatment of male and
female journalists, politicians, and activists on social media in Zimbabwe. It focused on
the platform, X (formerly Twitter). It investigated three questions:
1. What is the Share of Voice (SOV) across gender groups?
2. Are women subjected to harsher, more gendered language?
3. Have these interactions had a chilling effect on female voices?
A pronounced gender SOV disparity emerged as a key finding, indicating a substantial imbalance in visibility and engagement between genders. The volume of content related to men vastly exceeded that related to women, underlining a significant gender gap in digital discourse participation and representation.
The study documented instances where female journalists, Haru Mutasa, Violet Gonda, and Shingai Nyoka, faced attacks from various political and social movements. However, when broadening the analysis to measure sentiment differences between gender groups, the results did not reveal clear disparities. If anything, men sometimes received slightly
more negative commentary. The investigation also examined the frequency and intensity of gender-based attacks using specific gendered insults and found no significant disparities. It is worth noting the subtleties of gendered insults though. It is our belief that these are too subtle to be picked up by current AI models. AI models that are are sensitive to gendered language might be a future advocacy area.
The study also explored the possibility of female commentators being disproportionately ‘silenced’ over time or deterred from engagement due to negative interactions on X/Twitter.
A time series analysis found that, indeed, female journalists in particular are posting less over time. This might be due to a chilling effect based on their experience on social media although this study does give reasons for this reduction.
In summary, this research finds that female voices and perspectives are sorely under-represented in social media discourse. While we were not able to pick up gendered language quantitatively, there is evidence of a chilling effect among female journalists, exacerbating the problem over time.
