The disclosure gap. What happens between a woman saying it and a form recording it?
When a woman discloses violence but the disclosure does not fit the categories on your instrument because the perpetrator is a family member not listed, because the incident does not map to the timeframe, because she describes it in words your coding frame does not carry, what do you do with it? Do you record the divergence, or does it get absorbed into the nearest available category? I would like to hear how people handle the moment where a real account meets a fixed form.
This question follows me from over a decade
teaching Gender and Development into the practice now. When a disclosure gets
folded into the nearest available category, that's not simplification — it's an
epistemic erasure: the instrument's categories encode someone's prior
assumption about what violence looks like, and every account that doesn't fit
is evidence the assumption was incomplete, not that the disclosure was
imprecise.
At Purpose Rwanda, we've faced the same tension
in peer-led recovery disclosure. What's worked is treating divergence as data,
not noise: our peer agents record the account in the person's own words
alongside the coded field, flagged as divergent, rather than forcing a forced
fit. It doesn't fix the coding frame — but it stops the instrument from
silently overwriting what it can't hear, and those flags are exactly what
should revise the categories over time.
A fixed form encodes a theory of violence whether
or not anyone chose one on purpose. The fix isn't better categories — it's
building the ongoing capacity to notice when reality keeps exceeding them.
@Rhoda Nakhosi When a real account of violence meets a fixed form, absorbing the disclosure into the nearest available category creates a double harm: it invalidates the survivor's lived reality and creates sanitized datasets that blind programs to how harm actually occurs in their community. In direct-service delivery, the immediate priority must always be care first and coding second, meaning the survivor's self-identified needs and safety strategy are driven by her exact narrative rather than the constraints of an administrative tool. To handle divergence without sacrificing data fidelity, effective programs implement a dual-layer recording model that pairs standard forced-choice options with mandatory, unconstrained free-text fields and explicit "divergence flags." Instead of quietly shoehorning non-partner familial abuse or unlisted timeframes into ill-fitting IPV categories, capturing these accounts verbatim preserves the contextual nuance required for tailored care while aggregating crucial qualitative evidence. When researchers and implementers actively track these divergent flags, the "mismatch" ceases to be an administrative error and becomes a vital feedback loop, proving that the instrument itself needs to be adapted to reflect the real-world complexity of the community it serves.
-->