Negotiating Openness under Authoritarian Risk: Feminist Open Data Sharing in Hong Kong

1. Introduction

In 2021, UNESCO published a report titled “UNESCO Recommendation on Open Science,” emphasizing the significance of open science in the contemporary globalized world. Observing the impact of large-scale global crises like COVID-19, there is an urgent need to facilitate the sharing of scientific knowledge, data, and information (UNESCO 2021, 3). This aligns with the broader theme of knowledge commons, which “seeks to resist these enclosures by challenging the grounds of exclusion as being illegitimate from a human flourishing point of view” (DeSouza 2023). Consequently, sharing research data has become an emerging norm across academia.

Indeed, advocacy for open data can be traced back to the 1950s when the World Data Center system was established. The 1980s saw the growing open access practices in natural science in the United States, Europe, and the Soviet Union; by the 1990s, it had become an international movement (Mauthner and Parry 2013, 50). In the new millennium, China established the Scientific Data Sharing Program in 2002. Subsequently, Western countries, including Canada, the United States, Norway, Switzerland, Germany, and Slovenia, introduced similar open data policies facilitated by higher education institutions, state authorities, and non-governmental organizations (NGOs) (Uhlir and Schröder 2007; Bishop and Kuula-Luumi 2017). In 2007, the Organization for Economic Co-operation and Development (OECD) published “Principles and Guidelines for Access to Research Data from Public Funding,” further promoting transnational open data sharing across Europe.

While discussions on open data predominantly focus on Science, Technology, Engineering, Mathematics, and Medicine (STEMM), there is a growing interest within Humanities and Social Sciences (HSS). Examples include Qualibank in the United Kingdom and the Qualitative Data Repository at Syracuse University in the United States (Corti and Fielding 2016). However, directly applying existing STEMM-derived practices and principles of data sharing to HSS may pose problems. In scientific disciplines, open data is narrowly defined as data “that can be used, studied, and modified without restriction, and which can be copied and redistributed in modified or unmodified form either without restriction, or with restrictions only to ensure that further recipients can also do these things” (Murray-Rust 2008; italics added). This concept is based on Robert Merton’s ideas of scientific ethos, including universalism and communality (Christensen et al. 2019, 13–22). However, an excessive emphasis on “without restriction” overlooks the importance of protecting research participants through confidentiality, reflecting an “ethics of care.”

Indeed, there has been literature discussing the challenges of sharing sensitive qualitative data as exemplified by feminist scholars (Westbury et al. 2022). We acknowledge the diversity of data in HSS and agree with the advocacy for openness regarding less-sensitive data, including statistical summaries, documentation of data collection methods, and codebooks (van der Zee and Reich 2018). This article’s focus, however, is on data sharing of highly sensitive data in an authoritarian and politically precarious setting. They may include interview transcripts, personal photos, and diaries containing personal narratives, particularly in cultural studies, sociology, and anthropology. Hence, we will explore the ways in which feminist and CARE-influenced approaches to data sharing can be operationalized in these regions.

We are particularly interested in data sharing under the politically precarious and authoritarian conditions in Hong Kong. Since the enactment of a national security law (NSL) in June 2020, Hong Kong’s political and civic environment has undergone significant change. The law criminalizes offenses defined in broad terms and carrying severe penalties, such as secession, subversion, terrorism, and collusion with “foreign forces.” In 2024, an additional national security framework was introduced under Article 23 of the Basic Law through the Safeguarding National Security Ordinance, which further extends the range of activities subject to national security provisions.¹ In light of these evolving conditions, it is timely to consider, from a theoretical and ethical perspective, how the principles of data sharing might be practiced responsibly within research conducted in “sensitive” contexts. We believe that the case of Hong Kong provides a unique perspective in the discussion on data sharing practices, supplementing existing discussions by suggesting operational practices of data sharing in politically precarious and authoritarian conditions.

In this article, we will reflect on two research studies carried out by the authors on the 2019 Anti-Extradition Bill (Anti-ELAB) movement and the asexual community in Hong Kong. Through paralleling these two distinct experiences, we challenge a dependence solely on the prevailing FAIR principles (Findable, Accessible, Interoperable, and Reusable) and consider a productive way to approach questions of data sharing via a conversation between data sharing expectations, the CARE principles (Collective benefit, Authority to control, Responsibility, and Ethics), and a feminist ethics of care. We propose that open data sharing may help marginalized communities, specifically in politically precarious or authoritarian contexts. However, visibility usually comes with risks and surveillance. Hence, issues of authorship of knowledge, positionality of researchers, and agency of the research participants have to be considered seriously.

Drawing from our experiences researching social movements and the asexual community in Hong Kong, we argue that open data can enhance the visibility of marginalized communities. Through opening raw data, participants can directly “speak their stories,” rather than merely serving instrumental roles in academic writing. However, we must remain aware of the tension between visibility and precarity, as greater visibility might lead to increased vulnerability and potential attacks. Thus, we explore how open data sharing can balance visibility with considerations of precarity and information justice, especially crucial for marginalized groups in authoritarian contexts where openness may pose significant legal and social risks. Ultimately, we question whether open data sharing is primarily tailored to the Global North and whether implementing such practices is impractical in the Global South, where ongoing political struggles complicate their feasibility. Balancing these considerations is a critical question for all researchers.

The subsequent sections will first situate open data sharing within a feminist ethics of care. We will then reflect on our research experiences with the asexual community and social movements in Hong Kong. Finally, we suggest steps to operationalize feminist and CARE-informed approaches to data sharing in politically precarious and authoritarian contexts, encouraging researchers to reflect critically on their positionality during the openness process.

2. Open Data from a Feminist and Decolonial Perspective

Following Natasha Mauthner et al. (1998), this article treats qualitative social science data as materials produced with and among people. Materials such as interview transcripts, field notes, audio, images, and diaries are context-dependent, whereas their meanings are co-constructed. As Mauthner et al. argue, even rich contextual background information cannot substitute for the biographical, social, and temporal conditions under which the accounts are produced. Because meaning in qualitative work is relational and historically situated, archiving alone cannot make a dataset whole or reparable for new theory building. Hence, treating archived interviews or field notes as free-standing facts risks a naively realist stance that sidelines reflexivity. Emma Gilby et al. (2022, 7) similarly remind us that data are “collected rather than generated.” They are organized and recontextualized to advance cultural conversations and explained as part of the analytic process. These insights, encouraging respect for the researched community, show why data sharing should be governed by differentiated access controls rather than by undifferentiated release.

While the educational promise of openness is well documented (Uhlir and Schröder 2007; Bishop and Kuula-Luumi 2017), these claims alone are too coarse for qualitative contexts because they presume data are detachable objects. In this article, we narrow the focus from open data in general to potentially more restricted forms of data sharing, defined as making research data available to others for verification and secondary analysis (DuBois et al. 2023). Often, the principle of open data is operationalized through tiered or controlled access based on how sensitive the data are, with requests reviewed, users signing data use agreements, access sometimes provided in secure settings, and reuse limited to approved purposes without redistribution (UK Data Service, n.d.; ICPSR, n.d.).

Open data discourse is not uniform; open data discourse is heterogeneous rather than monolithic. Qualitative researchers, nonetheless, warn us that collapsing these differences into a single ideal of “letting the data speak” risks a naive technological determinism.² Stefania Milan and Emiliano Treré (2019) sharpen this point by urging us to move beyond data universalism. When discussing data from the Global South, researchers should treat “Souths” as plural and thus foreground decolonial thinking and recenter agency in data practices and activism, ultimately cultivating alternative data imaginaries oriented to epistemic justice. Reading against this backdrop, information justice becomes less a distributive problem and more a question of which infrastructures we build, where, and for whom.

This leads to an infrastructural and ethical reframing of “access.” If openness is plural, then literacy must extend beyond reading files to reading infrastructures. Jonathan Gray et al. (2018) define data infrastructure literacy as the capacity to account for, critique, and intervene in the socio-technical systems through which data are produced and circulated. Infrastructures are relational ensembles, so sharing is an intervention in infrastructure, not merely a transfer. This infrastructural lens reframes “access” as a set of decisions shaped by community agreements, within which tiered or controlled access finds its place.

Ethically, a consent form alone is not sufficient. Annette Markham et al. (2018) argue that ethics constitutes an ongoing and situated practice throughout the entire data life cycle, encompassing collection, curation, access, and reuse and guided by the principles of contextual integrity, deflationary perspective, and an ethics of care as practice (Nissenbaum 2011; Luka and Millette 2018). This approach aligns with Carolyn Ellis’s (2007, 4) relational ethics, which she describes as a requirement posed to researchers “to act from our hearts and minds, acknowledge our interpersonal bonds to others, and take responsibility for actions and their consequences.” The notion of participatory and community-based research proposed by Katherine Tamminen et al. (2021) is also relevant here.³ They state that consent should be meaningful, ongoing, and reversible, with opportunities for participant review and redaction as risks change. It also clarifies why the common insinuation that refusal to share signals poor quality is untenable in politically sensitive settings (Wicherts et al. 2011), where “neutrality” can mask normative commitments (Harding 2004).

Evidence from qualitative practice reinforces governance over blanket mandates. Indeed, what qualitative researchers themselves report underscores the need for governance rather than blanket mandates. Alex Broom et al.’s (2009) findings show that many fieldworkers view qualitative inquiry as relational, context bound, even “artistic” and fear that archiving secondary-analysis-ready files distorts the craft. The way through this impasse is not to discard sharing but to govern it. Particularly for digital traces and vulnerable groups like LGBTQIA+ communities, a feminist ethic of care to prioritize safety, dignity, and careful de-/re-identification is advocated (Dieterle 2021).

Feminist and Indigenous contributions offer operational governance tools that align with this safety and dignity-first approach. The “Feminist Data Manifest-No” (Cifor et al. 2019) centers equity, accountability, accessibility, Indigenous sovereignty, disability justice, and community control. It recognizes “refusal” as ethically legitimate where harms outweigh benefits and reframes data as a relationship rather than a market product. Within this framework, Indigenous data sovereignty shifts from treating Indigenous peoples as a “population” to recognizing peoples with authority to govern data about their lands and lifeways (Kukutai and Taylor 2016). Inspired by feminist and Indigenous principles, Stephanie Carroll et al. (2020) developed the CARE principles to complement the existing FAIR principles by centering the repository rules of collective benefit and authority to control. Together, these literatures justify tiered access and conditional reuse that honors community decisions.

These reconsiderations of data sharing as necessarily situated and ethically contingent suggest their commonality with many aspects of feminist methodologies. We do not point to a unified notion of feminist methodologies; instead, it is important to clarify that feminism should be understood as “a perspective not a method” (Reinharz and Davidman 1992, 240).⁴ It is neither monolithic nor homogeneous but rather a continually evolving approach that demands ongoing reflection on the validity of research as both process and knowledge production (Gatenby and Humphries 2000, 90). Feminist methodologies examine how knowledge about social life is produced and whether it reflects the lived realities of marginalized subjects within androcentric methodological traditions that obscure gendered structures of experience. It interrogates who is authorized to know, what constitutes valid knowledge, and how epistemic practices are bound to the ontological relationship between knowing and being (Landman 2006, 431). Although this article emphasizes a critical lens guided by feminist methodologies, we feel these approaches are relevant to a wide range of research contexts. Thus, we draw on the concept of “feminism for the 99%” (Arruzza et al. 2019), suggesting feminist methodologies can apply broadly across various fields and research topics, offering insights to reflect on established research processes, ethics, and norms.

Multiple feminist scholars, particularly those working with Indigenous and ethnocultural communities, have emphasized standpoint approaches that recognize how marginalized social positions generate distinct epistemic insights by exposing structural power relations obscured from dominant perspectives.⁵ These approaches call for rigorous positional reflexivity to both amplify marginalized voices and sustain methodological integrity within feminist inquiry.⁶ Building on this, relational methodology has been advanced as one of the process-centered frameworks that foregrounds culturally sensitive and ethically accountable researcher-participant relationships, situating these relationships at the core of feminist methodological development (Allen et al. 2021, 178).

The ethics of care thus emerged as a pivotal concept in this reorientation toward relationality. In her landmark work In a Different Voice (1982), Carol Gilligan critiques traditional justice-based ethical theories grounded in rules, rights, and universality that are often associated with patriarchal moral reasoning. She traces instead the development of an “ethic of care” that values context, interdependence, relationships, and responsibilities toward concrete others.⁷ Virginia Held (2006, 43) encapsulates the ethics of care as the ways in which we build relations of care, concern, and mutual responsiveness to need across personal and social contexts, offering a means to evaluate human relationships and to envision more ethical forms of society.⁸ Feminist researchers have incorporated the ethics of care into research methodologies, emphasizing the dynamic and situated responsibilities that emerge between researcher and researched. For instance, Joan Tronto’s (1993) political theory of care identifies key dimensions of care, including attentiveness, responsibility, competence, and responsiveness, as central to both political and research ethics. A caring researcher, according to this framework, must remain attentive to participants’ needs, take responsibility for mitigating harm, exercise competence in ethical protocols, and remain responsive to feedback and dynamics in the field. Similarly, Rosalind Edwards and Melanie Mauthner (2012) argue that feminist research ethics requires attention to the “specificity of lived experience” and the co-creation of research encounters that position research subjects as active agents and knowers.⁹ In this sense, an ethics of care in feminist research extends beyond gender-related topics. It operates as a foundational ethical guideline for all qualitative inquiry involving human subjects. It helps researchers avoid the re-inscription of dominant discourses and maintain reflexive accountability that keeps voice, agency, and context at the center of ethical research practice.¹⁰ Hence, feminist methodologies reveal that the discourse of open data sharing must shift from universalizing transparency to situated, relational, and community-driven openness.

The geopolitical stakes identified earlier in this article reveal a gap in how the application of this continuum of ethical approaches to qualitative data sharing can be operationalized in the context of precarious or authoritarian settings. Based on the literature reviewed in this section, we will reflect on our research experience in Hong Kong and consider open data sharing within the ethics of care in order to aid the formulation of practical recommendations for researchers.

3. Social Movement Research

One of our co-authors, Lucas Wong, conducted research on the Anti-ELAB movement in Hong Kong between 2020 and 2021.¹¹ The primary research question explored the connections between previous social movements (the Umbrella Movement in 2014 and the Mong Kok Incident in 2016) and the Anti-ELAB movement. Through 13 focus group interviews involving young adults of 18 to 35 years old, who were key supporters of Hong Kong’s localist movement over the last decade, the project examined protesters’ perceptions of Edward Leung, a young activist, and his localist advocacy. The focus particularly centered on the affective responses to the documentary Lost in the Fumes. Research participants were asked to evaluate the relationship between their acts and their perception of political figures, and thus their responses may include personal accounts of their acts during the movement. Since the NSL took effect in mid-2020, these accounts have become highly sensitive and could expose research participants to legal risks they did not foresee when interviewed before the law’s implementation.

In Hong Kong, the promise of confidentiality in qualitative research sits in a structural tension with lawful powers that may compel disclosure of research data to state authorities (including both the police and the courts). The Personal Data (Privacy) Ordinance (PDPO) contains a “legal proceedings” exemption that removes the usual purpose-limitation barrier when the use or disclosure of personal data is required or authorized by law or by a court order. This exemption does not itself create a power to compel; compulsion flows from specific legal instruments (e.g., a sealed court order, subpoena, production order) and procedural rules. But once such authority exists, PDPO no longer blocks the use/disclosure for that purpose.

Following ethical research protocols, all interview data have been deleted after research publication. However, given the increasingly homogeneous political discourse and suppression of dissent in Hong Kong, a theoretical reflection on the suitability of making interview transcripts openly accessible becomes both timely and urgent. Doing so could allow local citizens to better understand young protesters’ motivations, potentially fostering deliberative discussions across political divides. This section highlights three main questions that inform a feminist data share approach.

4. Asexual Community Research

The research of another co-author of this article, Yumi Wong, explores the lived experiences of asexual and aromantic (A-spec) individuals in Hong Kong. Through collecting personal narratives of touch, affect, and everyday relational practices, the project examines how socio-cultural ideologies surrounding sex and romance work to marginalize and constrain the bodies and relational capacities of those who do not experience sexual or romantic attraction. Using a community-based narrative approach, the research is grounded in semi-structured interviews with A-spec individuals. These conversations cover participants’ journeys of identity, experiences of intimacy, and visions for future relationality. Given the sensitive content including personal and potentially traumatic experiences, participant protection is of central ethical priority. Participants give informed consent, and all data are anonymized and stored separately from personal identifiers on encrypted devices. Sensitive details are shared only with double-confirmed permission, and unprocessed data are deleted after the consent period ends. The study also considers the potential value of creating an open archive of A-spec narratives. Diverse representation across the spectrum is essential not only for increasing A-spec visibility but also as an expression of community agency to foster awareness and support social change. While open data offers possibilities for amplifying marginalized voices and democratizing knowledge production, it also raises complex ethical and practical challenges. In the context of this project, four major areas of tension are identified.

5. Toward a Feminist Approach to Open Data Sharing

These reflections prompt critical consideration of whether an implementation of data sharing under politically precarious and authoritarian conditions requires more cautious reconsideration. Ethical feminist research must remain sensitive to participants’ contextual realities, prioritizing relational ethics, participatory agency, and protecting vulnerable populations above open data’s universalizing aspirations.

We therefore propose five key questions for researchers concerning feminist open data sharing. They are inspired by scholarship in feminist and critical data studies:

1. How can openness be reimagined as a relational, situated, and accountable practice?

2. How can mutual benefits between researchers and researched communities be ensured?

3. How might open data sharing inadvertently perpetuate inequality or trauma?

4. How should community rights over data ownership and control be respected?

5. Who is entitled to reuse sensitive data and under what ethical conditions?

Our recommendations intentionally synthesize prior strands mentioned in section 2 and attempt to reconsider their relevance in politically precarious and authoritarian contexts. In particular, they operationalize (1) Milan and Treré’s (2019) call to move beyond data universalism and to treat the “South(s)” as plural and agency-centered; (2) Gray et al.’s (2018) “data infrastructure literacy,” which reframes access as participation in, and intervention on, socio-technical infrastructures; (3) “ethics-as-methods” heuristics that treat care and contextual integrity as ongoing practices rather than one-off approvals (Markham et al. 2018); (4) the commitments in the “Feminist Data Manifest-No” (Cifor et al. 2019) to meaningful, revisable consent and community control; and (5) Indigenous data sovereignty logics that foreground authority to control and collective benefit (Kukutai and Taylor 2016). We adapt these into concrete governance moves for qualitative researchers under authoritarian risk, as our Hong Kong cases require.

An ethics of care perspective challenges the false dichotomy that positions “open data sharing” on one side and “privacy/confidentiality” on the other. Instead, it reimagines openness as a relational and situated practice. Echoing what other scholars have called a “radical empathy approach” to data sharing or “care-full data management,” we follow De Lucas et al. (2025) in expanding the concept of openness. Openness is not merely about maximizing access; it is also about situationality, relationality, and accountability. It can mean fostering open dialogue about research processes, sharing aggregated insights, or creating platforms where communities retain control over access to their own narratives.

Additionally, we support Carroll et al.’s (2020) CARE principles. Initially focused on Indigenous data, these principles advocate for “Indigenous control over Indigenous data,” positioning data practices within Indigenous cultural frameworks for community benefit, thereby decolonizing open science. The CARE principles may apply to diverse datasets other than Indigenous data. Collective benefits—the “C” in CARE—are not abstract public goods that justify sacrificing minority interests; rather, they emerge directly from relationships between researchers and researched communities. Researchers and researched communities should mutually benefit from open data sharing. For instance, how can A-spec or student protesters benefit from open data? While open data discourses often emphasize social benefits broadly, an information justice perspective highlights potential inequalities exacerbated by data openness.

The Collective Conditions for Reuse (CC4r) practice may provide insight into ethical open data reuse. While CC4r licenses generally permit open use, they can impose limitations or specific requirements for certain user groups whose reuse might harm the researched communities (Snelting and Weinmayr 2024, 7–8). For instance, consider how heterosexual men might use queer narratives or how state authorities might exploit student protesters’ personal narratives. This approach challenges the unrestricted openness advocated in open science (Murray-Rust 2008). Given the involvement of human subjects, their narratives, and researchers’ insider/outsider positions within communities, debates around data openness must consider community roles, past traumas, authorship of knowledge, and rights regarding data openness.

Thus, we advocate six steps for feminist open data sharing, designed to maintain openness as context-sensitive and negotiated, ensuring marginalized communities primarily benefit from research decisions guided by care, empathy, and justice.

1. Seek explicit and situated consent for open data use: Clearly explain to participants whether and how their data may be archived, reused, and circulated; allow them to veto or set conditions for sharing.

2. Establish a community advisory board: Form a representative group from the researched community to guide decisions on data collection, interpretation, archiving, and reuse.

3. Enable participants to review and redact their data: Before archiving, invite participants to review transcripts or materials, edit sensitive parts, decide how they want their narratives labeled, and specify any restrictions.

4. Use standard protective practices like anonymization and data use agreements: Protect identities through technical means but acknowledge that these are not always sufficient without relational ethics.

5. Embed contextual information with datasets and require engagement with it: Alongside the data, include research journals, consent forms, codebooks, and reflexive field notes that explain how the data was collected, consented, and interpreted; require that any secondary analysts read and reflect on this embedded context, including power dynamics and ethical dilemmas, before reuse.

6. Implement tiered access or controlled repositories: When appropriate, limit who can access the data through permission-based systems rather than full public release.

Our six-step workflow translates established principles into actionable procedures for sensitive data sharing under authoritarian risk. First, consent is situated and revisitable in the spirit of the FRIES model mentioned in “Feminist Data Manifest-No” (freely given, reversible, informed, enthusiastic, and specific) and is aligned with people-oriented governance (Cifor et al. 2019; Carroll et al. 2020). When data sharing is considered, researchers must explicitly seek consent and allow participants to veto sharing or specify reuse conditions, thereby rejecting a one-size-fits-all release (DuBois et al. 2018). Second, a community advisory board operationalizes “authority to control” from Indigenous data sovereignty models (Kukutai and Taylor 2016; Carroll et al. 2020) by giving exposed communities binding power over archiving and secondary access so that each data decision advances community rights, collective benefit, and cultural values, in line with Milan and Treré’s (2019) recentering of agency. Third, participant review, redaction, and self-labeling enact ethics-as-methods (Markham et al. 2018), enabling contributors to remove high-risk passages, rather than being fixed by external categories.

Fourth, recognizing the limits of technical masking flagged by Mauthner et al. (1998), anonymization is paired with formal data use agreements that condition reuse and exclude foreseeable harmful actors (DuBois et al. 2018), resisting data extractivism while addressing concrete risks.¹³ Fifth, context-rich “packets” (field journals, consent pathways, reflexive notes, glossaries) with mandatory engagement apply Gray et al.’s (2018) infrastructural inversion and respond to Mauthner et al.’s (1998) warning that “background” cannot make data whole by transforming context from an optional appendix into an interpretive obligation; including such metadata ethically situates the materials when openness would otherwise abstract them from relations of power. Sixth, tiered access and controlled repositories with conditional licensing implement CARE logics (Carroll et al. 2020) together with CC4r-style constraints and the Global South(s) agenda of anti-universalism (Milan and Treré 2019), so that openness is negotiated relative to power, risk, and vulnerability. This shows that, at times, limiting access, requiring special permissions, and using permissioned tiers can help uphold the spirit of consent as social and political conditions evolve (Campbell et al. 2023).

In our Hong Kong cases, these steps could work as an integrated, care-centered workflow. Recognizing consent as a continuous negotiation, A-spec narrators may choose to tighten or relax data use permissions and update their self-descriptions at multiple points, whether prior to publication, during data dissemination, or even post-release. Decisions about archiving and reuse are reviewed by a community advisory board that evaluates collective benefit and risk before any secondary analysis proceeds. Prior to deposit, participants read their transcripts and may redact or rephrase passages that carry legal, social, or identifiability risks. Technical protections such as anonymization are paired with explicit data use agreements that specify acceptable purposes and forbid reuse by other hostile actors. While such measures cannot fully prevent state demands for research data, removing direct identifiers and limiting granular indirect identifiers would increase the difficulty for authorities to trace specific participants. Each dataset travels with a required “context packet” (field notes, consent rationales, positionality statements, codebooks, and glossaries), and secondary users must attest that they have engaged with it before analysis. Finally, Anti-ELAB materials are released only through permissioned and tiered access so that public visibility does not become participant vulnerability.

In terms of interpretive authority, if we advocate for open data sharing as a means to democratize knowledge (McLeod and O’Connor 2021), then the primary audience entitled to interpret and benefit from data about marginalized communities should be the communities themselves. Knowledge should not be unilaterally controlled by external researchers or institutions. Authority must be reimagined in ways that empower research subjects and prioritize their rights, voices, and values.

It is equally important to recognize that research subjects are not merely “vulnerable populations” in archives, but knowledgeable, agentic individuals who should guide decisions about their own data. At the 2023 workshop “Being Open About the Unopenable,” researchers working with LGBTQIA+ communities, refugees, and survivors of violence emphasized that ethical openness must remain context-dependent and that data governance should be led by community self-determination (De Lucas et al. 2025). Community advisory boards, for example, can help ensure that decisions about data interpretation and reuse are inclusive, reflective, and co-produced with community input. Involving community representatives in early decisions about data handling, post-study archiving, or research design may also help align data practices with participants’ comfort and expectations.¹⁴ These collaborative practices not only enhance ethical alignment but also improve anonymization and metadata quality, as participants can flag what they consider sensitive or misrepresentative.

Another crucial concern is how to structure secondary data analysis in a way that preserves the relational context of the original data and avoids extractive practices. Interestingly, there has been discussion about the lack of contextual information. And some scholars believe that secondary analysis could be carried out even without that contextual information.¹⁵ However, we believe that contextual information is indispensable in secondary analysis, particularly in topics related closely to personal trauma and emotions like personal gender identification and social movement involvement. This is to ensure that the research remains faithful to the research participants and to acknowledge their roles as “co-producers” of that knowledge.

A practical solution is to accompany shared datasets with extensive contextual information, such as research journals, consent forms, codebooks, and reflexive notes alongside the data. Decades of feminist research have documented dilemmas around “power and privilege, and the overall quality of the relationships between researcher and researched” in fieldwork (Miller et al. 2012). Narratives detailing how consent was obtained, the power dynamics at play, and any ethical dilemmas encountered during the research can also be included in repositories. By embedding ethical context into the data, we invite future researchers not only to analyze but also to reflect on whose voices are being amplified and whose might be overlooked in the process of secondary analysis. Anyone reusing the data must engage with the human context and relational labor that gave rise to the numbers or narratives they are working with. This approach helps shift the paradigm: Instead of treating data as static, abstract objects, we begin to understand them as relational products born out of dialogue, trust, and shared responsibility. In doing so, we ensure that the drive for data accessibility does not come at the cost of those whose stories are being told.

6. Conclusion

This article has examined the operationalization of data sharing in a politically precarious and authoritarian context. Central to the argument is the critique of positivistic norms of objectivity and unrestricted openness, which often overlook ethical responsibilities, participant protection, and power dynamics, particularly in research involving marginalized communities. Through autobiographical analyses of research experiences within Hong Kong’s asexual community and social movements, we highlight the complex interplay of visibility, precarity, and information justice in implementing data sharing. Our key argument is that openness must be reimagined as community governed, context embedded, and contingent, and our contribution is an operationalized care-centered workflow that we demonstrate as both necessary and feasible in Hong Kong’s A-spec and Anti-ELAB cases under politically precarious, authoritarian conditions.

Positionality and reflexivity emerge as crucial themes in reconciling these tensions. Autobiographical reflections that acknowledge the researcher’s embeddedness within socio-political contexts are fundamental not just as methodological tools but as ethical imperatives. Reflexivity, understood as “a radical consciousness of self in facing the political dimensions of fieldwork and the construction of knowledge” (Nagar 2014, 82), allows researchers to critically interrogate their own power and privilege and navigate the ethical complexities of data sharing. We emphasize a “speaking-with” model, where reflexivity and positionality evolve dynamically, shaped by continuous engagement with participants’ voices and lived experiences (Nagar 2014, 83–85). This approach resists the simplistic binaries of the researchers versus the researched, or Global North versus Global South, emphasizing instead situated solidarities that are sensitive to local struggles and global power asymmetries. Consequently, feminist research does not merely critique the discourse of open data sharing but actively proposes alternative frameworks that foreground relational ethics, community authority, and mutual benefit.

The implementation of feminist open data sharing demands nuanced ethical strategies, such as explicit, situated consent; community-led governance; and comprehensive contextual documentation to facilitate responsible reuse. Recognizing the complex, sometimes contradictory nature of visibility, particularly in authoritarian contexts, we advocate a context-sensitive openness that prioritizes the safety and autonomy of marginalized communities. Ultimately, this article underscores the necessity for data sharing to evolve beyond positivist, universalist frameworks. Through rigorous reflexivity and careful attention to positionality, feminist methodologies provide a critical lens that ensures openness serves ethical, relational, and justice-oriented ends, reshaping not only how data is shared but also whose voices are amplified and protected in scholarly and public discourses. Therefore, we invite researchers considering open data to also reflect on their own positionality and their relations with the communities.

Open Peer Review Reports

Open peer review reports for this article are available at the following location: https://doi.org/10.17613/vtm91-2t748