Civil Society Repression for Women's and Ethnic/Racial Minorities

• In Event: Lightning Round: Research Design

Abstract

A government’s human rights practices can directly influence the amount of foreign investments, development aid, and cooperation received from liberal states (Dupuy et al. 2016; Barry et el. 2013). As a result of this, states have an incentive to preserve and conceal their human rights violations in order to protect their international reputations (Meernik et al. 2012). Civil Society Organizations (CSOs) are a key component in generating political costs for human rights abuses because they are able to report human rights conditions to international actors and domestic constituencies (Dai 2005; Meernik et al. 2012; Krain 2012). Governments are able to curtail CSO efforts to publicize government-perpetrated human rights violations by placing restrictions on them, including limiting their access to funding, travel etc. (Bromley, Schofer, and Longhofer 2020; Christensen and Weinstein 2013; Poppe and Wolff 2017). Therefore, the experiences of civil society organizations are important to understand, because this describes the extent to which CSOs can hold repressive regimes accountable for human rights violations (K. Bakke et al. 2020; Simmons 2009).
There is limited exploration in the literature on how the experiences of different organizations can vary based on the identity associated with the civil society organization (Kiyani and Murdie 2020; A. Murdie and Peksen 2015). If minority groups are being targeted based on their identities, this could influence the types of human rights abuses that governments are likely to be held accountable for. Following a similar process to other datasets on civil society repression, I create original data that notes when minority civil society groups are being repressed or targeted by the state in instances reported in the United States Human Rights Reports. I conduct a unique technique that begins with the human coding of the texts of these reports to indicate if the text represents an instance in which the state is repressing civil society. Once this sample of data is hand coded, multiple machine learning models are trained on this hand coded data to identify civil society repression allegations. The best performing model is then used to identify civil society repression allegations in the rest of the sample.
I compare the levels of repression and experiences in both women’s and race identity-based groups to broader civil society groups overall in order to investigate the research question, Do women and race identity-based civil society groups experience more repression than civil society groups that are not attached to a minority identity?

Author

• Julianne Windham, Arizona State University