The ethics of asking again
Longitudinal VAWG data requires asking women about violence more than once. But repeated questioning is not neutral. It can retraumatize, it can signal to a woman that her previous answer was not believed and in some settings it can increase risk if a partner notices the pattern of visits. How do you decide when re-asking is justified and when the burden outweighs the value of the additional data point? And who makes that call, the researcher, the ethics committee, or the woman herself?
Rhoda, I think the key distinction is between scientific value and ethical justification. Re-asking about violence should not be justified simply because another data point would improve statistical power or longitudinal completeness. The additional measurement should have a clear scientific purpose, and the potential value of that information must be weighed against participant burden and safety.
The decision therefore should not rest with only one actor. The researcher and ethics committee should establish in advance when repeat measurement is scientifically necessary, proportionate, and accompanied by appropriate safeguards; however, the woman retains the right to decline, pause, or withdraw from the questioning at any point. In practice, that means privacy, trained interviewers, informed consent at each relevant encounter, safe procedures for contact and follow-up, and a clear referral/support pathway where needed. WHO guidance similarly emphasizes participant safety, privacy, preparedness for harm, and monitoring potential harms arising from violence research.
For me, the practical test is: What new information are we genuinely trying to learn, can it be obtained with less burden, and does the expected scientific value justify asking again? If the answer is weak, the additional question may not be ethically justified simply because the study design allows repeated measurement.
Dr. Magoba's distinction between scientific value
and ethical justification is the right frame, and I'd add one layer from the
practice side: in a peer-led model, the person doing the asking changes what
"re-asking" even means. When a woman is asked again by a
stranger-researcher, the question carries the full weight Rhoda describes — did
you not believe me, why are you back. When she's asked again by someone she
already has an ongoing relationship with — a peer who has earned trust through
consistent, non-extractive contact — the re-ask can read less as suspicion and
more as continuity of care.
That's not an argument for looser ethics; it's an
argument that the relationship the question sits inside is itself a
variable researchers should be accounting for, not just the question's phrasing
or timing. At Purpose Rwanda, our peer agents don't "re-ask" in the
research sense at all — disclosure resurfaces naturally across an ongoing relationship
rather than through a scheduled follow-up instrument. That's a different
data-generating process than longitudinal VAWG research, and I don't think it
transfers cleanly. But it does suggest the ethics-committee/researcher/woman
triad Dr. Magoba names might need a fourth consideration: whether the mode
of asking again — cold instrument vs. embedded relationship — changes the risk
calculus as much as the content of the question does.
Charles, I appreciate this important addition. I agree that the mode and relationship within which re-asking occurs should be considered part of the ethical calculus, rather than treating the question itself as the only relevant variable.
A trusted peer relationship may create a different environment for disclosure, particularly where continuity, familiarity and psychological safety have already been established. However, as you rightly note, naturally resurfacing disclosure within a supportive relationship is not necessarily equivalent to deliberately repeating a standardized research measure. The latter still requires a clear scientific rationale, appropriate consent and safeguards against distress, coercion or unintended harm.
I therefore think the fourth consideration you propose fits well alongside the researcher–ethics committee–participant triad: not only whether we should ask again, but who asks, how, why, and within what relationship or context. These factors can influence both participant safety and the nature of the data generated.
Ultimately, continuity of relationship may reduce some risks, but it should not automatically lower the ethical threshold for repeated measurement. The scientific purpose, proportionality, participant autonomy and safeguarding requirements should remain explicit in the study design.
@Henry Magoba, I want to push on one phrase you used: "continuity of relationship may reduce some risks, but it should not automatically lower the ethical threshold." I agree it should not automatically lower it but I wonder whether the threshold itself is the right metaphor. A threshold implies a fixed line that either is or is not crossed. What @Charles Bawate and @Nonvicks are both describing sounds more like a shifting risk profile: the embedded relationship lowers the risk of retraumatization but raises the risk of consent erosion; the cold instrument does the reverse. If that is right, then the job of the ethics committee is not to set one threshold and apply it uniformly, but to require researchers to name which specific risks their design amplifies and which it mitigates, and to justify why that trade off is acceptable for this population in this context. Does that framing feel workable to you, or does it risk making every protocol a bespoke negotiation with no shared standard?
@Rhoda Nakhosi In longitudinal Violence Against Women and Girls (VAWG) research, repeated questioning is an active intervention that carries real risks of retraumatization, participant mistrust, and physical safety hazards if a perpetrator notices a pattern of visits. Re-asking women about violence is ethically justified only when the additional data point directly alters an immediate safety plan, enables access to critical care, or evaluates whether an intervention is actively reducing risk over time; if a follow-up query exists purely for academic completeness or panel maintenance without offering tangible benefit, the burden outweighs the data's value. To minimize harm, methodological approaches must employ low-friction screeners, low-profile safety protocols, and explicit memory-anchoring language so women know their prior disclosures were believed rather than forgotten. While Institutional Review Boards set baseline ethical ceilings and researchers are responsible for trimming unnecessary variables, ultimate authority must belong to the woman herself through dynamic, ongoing consent. She retains the absolute right to skip modules, decline follow-ups, or withdraw entirely at any point in the study without forfeiting her access to support services or referral networks.
@Nonvicks Ochieng, your criterion of re-asking is justified only when it directly alters a safety plan, enables access to care, or evaluates whether an intervention is reducing risk or is admirably clear, but I want to test its edges. Longitudinal VAWG research often exists precisely because we do not yet know whether an intervention is reducing risk; that is what the follow-up is trying to establish. Under your criterion, would you permit a second wave of data collection whose primary purpose is to answer that question, even if no individual woman receives a tangible benefit from that particular interview? Or does the absence of immediate individual benefit mean the burden can not be justified regardless of the population level value? I am genuinely unsure where I land on this, and I would value your read.