Bearing Witness as Evaluators: When to Help, When to Stay Away
Two days ago I wrote about filming an ICE arrest from my window. The emotional energy that followed was immediate and overwhelming. I wanted to do something. I wanted to know: Can I use my skillset as an evaluator in this moment? Should I? What are the risks?
For evaluators working in humanitarian settings, these questions are familiar. But for those working domestically in the US, this is new terrain. We are trained to observe, to record, to analyze. When communities are organizing under threat, when the state has made explicit its interest in that organizing, our professional practices carry risks we need to think carefully about: to ourselves and to the communities and community members we might work with.
Each evaluator will make their own decision about how to respond. This is not about prescribing one right answer. It is about being intentional and thoughtful. About understanding what we might offer, what we might cost, and whether communities are better off with or without our involvement.
The starting point is this: the communities protecting their own in Minneapolis don’t need evaluators to validate their work or design their learning systems. They are already learning, adapting, building knowledge that keeps people safe. They have a learning system. It is called survival.
But there are things evaluation skills can support, if communities ask: facilitation of after-action learning, training in pattern recognition, capacity building that lets communities do this work themselves. And evaluators can contribute in ways that have nothing to do with evaluation skills: showing up, providing resources, offering solidarity. The question is not whether evaluators have useful skills. The question is whether our involvement creates more risk than value, and whether communities have actually asked for our help.
The choice to engage directly depends on your ethical framework, your capacity for risk, and most importantly, whether communities have invited your involvement.
Understanding the Risks We Pose
Evaluators are not neutral observers, nor are we outside the dynamics we study. We are present with our worldviews and our presence changes things. Our questions shape what gets said. Our documentation creates records that did not exist before. Our analysis makes implicit patterns explicit. In contexts where communities face state scrutiny, this matters. The choice is not between engaged and objective research. The choice is about what our research serves and who it might affect.
Our professional accountability runs in multiple directions. Evaluators are accountable to funders, to institutions, to professional standards, to review boards. We write reports. We present findings at conferences. Sometimes we publish. This means knowledge does not always stay where it originated. It circulates. It becomes accessible to audiences we may not have anticipated. Even when we anonymize, we reveal patterns. Even when we aggregate, we expose structure.
Before considering engagement, evaluators need to understand the risks our professional practices create, for ourselves and for the communities we might work with. Here are some things to be aware of:
Our documentation can create vulnerability. We are trained to write things down, to keep records, to document processes and outcomes. This is how we prove our work has value, how we demonstrate rigor. But when communities operate under threat, documentation becomes a different thing. Interview transcripts could be subpoenaed. Case studies could map who is involved and how they coordinate. Evaluation reports describing organizing strategies could become guides for how to disrupt them. We think we are capturing learning. We need to consider whether we might be creating evidence.
Our presence can signal attention. When researchers show up, it can mark something as significant. When evaluators start asking questions, it suggests something noteworthy is happening. When academics write, it draws eyes. The fact of evaluation itself can make work more visible, and visibility in this context is not always safe.
Our work can teach unintended lessons. When we write about what makes organizing effective, we are also writing about what would make it less so. When we identify strengths, we might also be identifying vulnerabilities. When we describe how trust is built, we might be describing how it could be undermined. The same analysis that helps communities could potentially help others understand how to interfere.
Before Considering Involvement: Sometimes the Answer Is Don’t
Some questions worth asking before engaging:
Does my involvement make this work more secure or more exposed?
Can I work without creating documentation that could be accessed later?
Am I prepared to never publish about what I learn?
Will community safety take precedence over my professional needs?
Am I ready to walk away if our presence creates risk?
Why do I think my expertise would be helpful here?
Am I part of this community, or am I external to it?
Most importantly: Was I asked? Or am I inserting myself?
If you’re internal, a community member who happens to have evaluation skills, you’re navigating different questions. You’re already embedded. You know the people. You understand the risks because you share them. But you still need to be careful. Your expertise can shift power dynamics even within your own community. Be aware of whether your skills are being offered as peer support or positioning you as the expert who knows better. Communities under threat need solidarity, not hierarchy.
If you’re external, the calculus is different. Have you been specifically asked? If not, be very cautious about offering. Communities protecting themselves from state enforcement are operating under threat. Your presence, even well-intentioned, creates risks. They may not know what an evaluator is or what evaluation skills could offer, but that doesn’t give you permission to insert yourself. The stakes are their safety, not your opportunity to help.
For many evaluators, both internal and external, the right answer will be: do not engage. Not because the work is not important. Not because your skills would not be useful. But because the risks your involvement creates, the professional structures you operate within, or the costs of working without documentation may outweigh what you can offer. Recognizing when you should stay away is not a failure. It is clarity about limits.
Guidelines for Engagement
These are collective risks we face as evaluators. What follows are individual choices each evaluator must make for themselves.
If communities have asked for your involvement, and you have decided the risks are manageable, here are principles that might guide ethical engagement:
Security comes before everything else. Before learning. Before documentation. Before accountability. Before impact. Before your professional reputation. If there is any possibility that your involvement creates risk through documentation, through visibility, through the questions you ask, through the connections you make visible, then you should not be involved. Communities define what security means for them.
Whatever you do should be lawful. You are supporting constitutionally protected activity, whether that’s facilitating learning, observing patterns, or helping people understand their rights. If your work is ever questioned, you should be able to explain clearly that you were supporting lawful organizing. This framing matters for your protection and others involved. There are times when there is a tougher calculus between what is lawful and what is moral or ethical. That must be a tightrope you are willing to walk.
Work without creating records when possible, or create only the records communities control. This means working differently: collecting only the data you actually need, not everything you could collect. Destroying data you no longer need rather than keeping it indefinitely. The less that exists, the less can be accessed later. This runs counter to evaluation’s archive-everything instinct, but in contexts of threat, minimalism is protection. In practice, this might mean no written documentation at all. It might mean writing things down, learning from them, and then destroying them. It might mean communities keep the records and you never see them. It might mean everything stays oral. There are ways to support learning without creating records: facilitating oral debriefs, asking questions that help people notice their own patterns, teaching analytical frameworks people can apply mentally, helping groups process their own observations together. The learning happens. It just does not get written down. If you cannot work this way, you cannot do evaluation work with communities under threat.
Democratize your skillset and spread it. If you are offering skills, you should be teaching them, not retaining them. The goal is for communities to be able to do this work without relying on you as the expert. This might mean training people in facilitation, not facilitating for them. Teaching pattern recognition methods, not being the one who recognizes patterns. Showing how to structure after-action learning, not running those sessions indefinitely. The timeline should be: teach, support while they practice, step back. If you are part of the community, you shift from expert to peer. If you are external, you make yourself unnecessary.
Don’t let your expertise dominate the agenda. You can contribute methodological insight about what questions might be useful, what analysis might reveal patterns. But your training does not mean you know what matters most. The community decides what learning serves their needs. Communities are not monoliths. Ensure all voices are heard, not just those who speak loudest or agree with your perspective. You should not push for certain types of analysis simply because they are methodologically interesting. You should not advocate for documentation because it would be valuable for the field. If you are external, you support the learning process on their terms. If you are part of the community, you offer your expertise as one perspective among many, not as the framework that should guide everyone else.
Accept that you may never see results, especially if you are external. You may not know how the learning was used. You may not know what decisions were made based on insights you helped facilitate. You may not know what difference your involvement made. You may not get to write about this. You may not get to present it. You may not get to use it as evidence of impact. If you are part of the community, you will likely see the results because you are living them, but you still may not be able to document or claim credit for your contribution. If you need external validation, if you need to prove your value through visible outcomes, if you need credit for your work, then this may not be work you can do. The work often requires invisibility.
Be prepared to forgo professional advancement. Working without documentation, without publications, without visible proof of impact means missing opportunities to build your evaluation career in traditional ways. This is easier if you are internal to the community and share in the community benefit. As an external evaluator, it may be more difficult to bear. The threat is active. The scrutiny is real. Documentation created now could be accessed now. But threats do not last forever. There may be a time after the threat recedes, after communities are safer, when documentation becomes possible. When communities might want their work recorded. When lessons learned could be shared to help others facing similar threats. That future documentation matters. But that work happens later, if communities decide it is safe and valuable. You do the work without documentation today. You might get to write about it eventually. But you cannot count on that. Most evaluators cannot do this. Not because they lack skill. But because the professional structures we operate within do not support this kind of work. Our careers depend on visible accomplishments. Our credibility depends on documented impact. Our livelihoods depend on proving our value.
When Not to Engage
Red flags that you should stay away:
You feel entitled to be involved because you have expertise.
You need this work for your career, your portfolio, your research.
You cannot imagine working without documentation.
You would be accountable to anyone other than the community.
Your presence would require justification to people outside the network.
The community has not explicitly invited you, but you think they should want your help.
You want to study this because it is interesting, not because you were asked to support it.
You believe evaluation is inherently helpful and communities just need to understand its value.
The harm you might cause:
If your involvement creates risk, if your presence drew attention, if your questions exposed something that should have stayed hidden, if your approach compromised security in ways you did not anticipate, you will not know until it is too late.
And you will have no way to make it right, because you will not be there when the consequences unfold. You might never know you caused harm. Or you might learn later that your help was not helpful, that your presence made things worse, that communities paid the price for your involvement.
You carry that possibility. And you have no insurance against it. No professional standards that guarantee your approach was sound. No peer review that confirms you did it right. Just the weight of knowing that you might have harmed people who were already vulnerable, and you did it while trying to help.
Sometimes the most ethical thing an evaluator can do is recognize when our work does not belong. There are spaces where we do not belong, work we should not touch, learning processes that should happen without us. Being willing to say “I should not be involved in this” is a form of professional integrity.
Knowing your limits is better than overstepping them. And sometimes, staying away is the most important contribution we can make.
Conclusion
One of the most important thing evaluators can do right now is say clearly, in public and in professional spaces, whose actions are lawful and whose are not. When colleagues express concern about communities organizing to protect their own, we can name that this organizing is constitutionally protected. When people suggest that encrypted communication is suspicious, we can point out that privacy is a legal right. When evaluation conversations treat community protection networks as objects of study rather than exercises of democratic rights, we can redirect that framing.
What communities are doing is constitutionally protected. What ICE is doing violates constitutional rights. Saying so requires no special evaluation skills. It creates no vulnerability for communities. It just requires honesty about what is happening and willingness to say so.
Yesterday I filmed an ICE arrest from my window because it felt like the only right thing to do. Today I am still asking: What does bearing witness require of me? The answer might be offering skills if asked. It might be providing resources. It might be showing up. It might be staying away.
But it definitely includes saying clearly: the people watching are exercising their rights. The people being watched are violating them.
Anthralytic helps mission-driven teams clarify and amplify their impact through evaluation practice informed by Buddhist philosophy and systems critique.
