A Commentary on Percival Carter and Obenauer (2025): From Compliance to Conscientização

Percival Carter and Obenauer (2025) propose diagnostic tools for workplace discrimination that represent a genuine advance beyond compliance-driven approaches. Their Root Cause Analysis, Threat Modelling, and Bug Bounty frameworks treat discrimination as an organizational design problem requiring systematic diagnosis rather than symbolic training programs. Yet embedded within their recommendations is an assumption that carries significant theoretical costs: the caution that including those “close to incidents” in diagnostic processes “risks introducing biases” (p. 21). This commentary argues that what the authors frame as methodological safeguard constitutes theoretical loss, forfeiting access to knowledge that only marginalized positions provide.

Standpoint theory, developed by Collins (1990) and Harding (1991), fundamentally recasts Percival Carter and Obenauer’s framing. Where the authors treat proximity to discrimination as requiring methodological safeguard, standpoint theory reveals this exclusion as theoretical loss. People in marginalized positions understand how systems function in ways that people in dominant positions cannot, precisely because they must navigate barriers that remain invisible to those who benefit from the system. Far from introducing bias, this situated knowledge provides access to information about system functioning that would otherwise remain hidden.

The theoretical costs of excluding marginalized knowledge are specific to each of Percival Carter and Obenauer’s tools. For their Root Cause Analysis, excluding those who experience discrimination risks systematic blindness to classes of causes. RCA teams composed primarily of organizational experts may identify causes visible from their structural position while remaining blind to causes apparent only when navigating the system from below. The result is a mis-specified causal model that addresses some factors while leaving others unexamined. For their Threat Modelling, the cost is incomplete vulnerability mapping. The authors recommend having “diverse employees” map vulnerability points, but diversity alone fails to guarantee that employees feel safe disclosing what they perceive or that their perceptions will be weighted appropriately. A threat model produced under these conditions will reflect vulnerabilities that dominant group members perceive while underrepresenting vulnerabilities experienced by those with less power to speak. For their Bug Bounty programs, the cost is biased detection coverage. If marginalized employees are less likely to report due to power differentials or prior negative experiences, bug bounty programs will systematically over-detect issues salient to relatively privileged reporters while under-detecting harms experienced by those facing the greatest barriers.

Percival Carter and Obenauer’s concern about bias from including those “close to incidents” is understandable. Individual grievances can sometimes overwhelm systematic analysis. However, the question should be reframed: the issue is how to structure inclusion, so that situated insight informs systematic analysis, rather than whether to include marginalized knowledge at all. Educator Paulo Freire (1970) distinguished between the banking model of education, which treats people as empty containers waiting to receive expert knowledge, and problem-posing approaches that treat people as capable of generating knowledge about their own conditions. The authors’ caution about including affected employees risks reinstating the banking model’s epistemic hierarchy within ostensibly participatory diagnostic processes.

Research on employee voice (Morrison, 2014), whistleblowing (Near & Miceli, 1985), and psychological safety (Edmondson, 1999) offers design principles that address Percival Carter and Obenauer’s legitimate concerns about bias while avoiding the theoretical costs of exclusion. Organizations might distinguish roles for affected individuals (co-analysts, informants, or reviewers), establish procedural safeguards demonstrating responsiveness to prior reports and protecting from retaliation, and use staged elicitation techniques that build trust before soliciting vulnerable accounts. These structures channel marginalized knowledge into systematic analysis while protecting those who share their experiences.

A brief illustration clarifies these extensions. Consider credential evaluation in higher education, where assessment systems can systematically undervalue knowledge acquired outside dominant educational traditions. An RCA team investigating why internationally educated professionals are disproportionately placed in entry-level programs would identify different root causes if the team included those who experienced the assessment process and could identify where rubrics failed to recognize their expertise. Similarly, a threat model of credential evaluation would be incomplete without input from those who navigate it and perceive barriers invisible to its designers. This example demonstrates how Percival Carter and Obenauer’s tools would function differently with marginalized knowledge at the center (Table 1).

Table 1.

Extending Diagnostic Tools to Include Marginalized Knowledge

Diagnostic tool	Authors’ design	Cost of exclusion	Extension for inclusion
Root cause analysis	Expert teams use 5 Whys and fishbone diagrams; affected individuals excluded to prevent bias	Systematic blindness to causes visible only from marginalized standpoints; mis-specified causal models	Include affected individuals as co-analysts or reviewers; use staged elicitation to move from individual accounts to systemic patterns
Threat modelling	Diverse employees map vulnerability points using STRIDE framework	Incomplete vulnerability mapping; threats perceived by marginalized employees underrepresented due to silence	Establish psychological safety norms; provide anonymous input channels; weight marginalized perspectives in analysis
Bug bounty programs	Compensate reporters for identifying system weaknesses	Biased detection coverage; harms experienced by those with less power to speak go unreported	Demonstrate responsiveness to prior reports; protect from retaliation; invite reporters as partners in remediation

Note. Authors’ design summarized from Percival Carter and Obenauer (2025). Extensions informed by Collins (1990), Freire (1970), Edmondson (1999), Morrison (2014), and Near and Miceli (1985).

Percival Carter and Obenauer’s core insight is correct: organizations need systematic approaches to diagnosing discrimination rather than defaulting to generic training programs. Their tools represent a genuine contribution to moving beyond compliance. This commentary has argued that realizing the diagnostic potential of these tools requires explicit attention to whose knowledge informs their application. When the authors caution that including those “close to incidents” risks introducing bias, they inadvertently frame epistemic privilege as epistemic liability. The theoretical costs are specific: systematic blindness in Root Cause Analysis, incomplete mapping in Threat Modelling, and biased coverage in Bug Bounty programs. Freire (1970) used the Portuguese term conscientização to describe the process by which people become aware of the conditions that affect them and take action to change those conditions. Moving from compliance to genuine problem-solving requires this transformation: recognizing that those closest to the problem often hold the clearest view of its causes and solutions.