Why Registry Coverage Needs More Than a State Count
Why Registry Coverage Needs More Than a State Count
A registry API can search many jurisdictions and still leave a screening team with unanswered questions. The useful question is not only which jurisdictions are included? It is also what did each jurisdiction return, what identifiers support a candidate match, and what should a reviewer do when the available information is limited?
For developers and screening platforms, coverage is a property of the entire review workflow—not just a number on a product page.
Jurisdiction coverage is not field-level coverage
Sex offender registries operate under different state and territorial rules. The information made public, the categories of registration displayed, and the details available for comparison can vary by jurisdiction. A response with a name and location does not provide the same basis for review as one that also includes a date of birth or another distinguishing identifier.
That distinction matters when an application presents results in a uniform interface. A single schema makes responses easier to integrate, but it should not make unequal source information look equally conclusive. If a field is absent, the interface should show that it is unavailable rather than implying that a comparison was performed and found no conflict.
The same applies to a search that returns no candidates. A reviewer needs to know whether the relevant jurisdictions answered the request and whether the search had enough identifying information to be useful. “No candidates returned” is a narrower statement than “this person has no registry record.”
Separate search status from candidate status
Model these as two different questions:
- Search status: Which jurisdictions were included, and did each return a usable response?
- Candidate status: For each returned record, which attributes agree with the query, which conflict, and which could not be compared?
Combining them into one green checkmark creates ambiguity. A system could return a plausible candidate while one jurisdiction remains unavailable. It could also return no candidates from the jurisdictions that answered. Those outcomes call for different review decisions.
Here is a conceptual response sketch for an internal application—not a representation of any vendor’s actual API fields:
{
"search_summary": {
"jurisdictions_requested": ["Jurisdiction A", "Jurisdiction B"],
"jurisdiction_outcomes": [
{ "jurisdiction": "Jurisdiction A", "status": "answered" },
{ "jurisdiction": "Jurisdiction B", "status": "needs_review" }
]
},
"candidates": [
{
"jurisdiction": "Jurisdiction A",
"match_basis": {
"name": "similar",
"birth_date": "not_available_for_comparison"
},
"review_status": "unresolved"
}
]
}
The point is not to prescribe a particular data model. It is to keep an unanswered jurisdiction visible and prevent a similar name from silently becoming an identity conclusion.
Make the evidence inspectable
A candidate score can help prioritize review, but it cannot replace the underlying match basis. Reviewers should be able to see which attributes contributed to a candidate match and which attributes were unavailable. A name-only match deserves different treatment from a name-and-date-of-birth match, even if both appear in the same results list.
Per-source provenance is equally important. Preserve the jurisdiction associated with each result and a citation or reference that lets a reviewer inspect the supporting record. When the interface combines results from multiple jurisdictions, do not collapse distinct records merely because names look alike. Conversely, do not assume two similar records represent two different people without checking the available identifiers.
Keep search outcomes, candidate evidence, and reviewer decisions distinct in the audit trail. A useful record shows what the system returned and what a person subsequently concluded. It should not rewrite an uncertain search result into a definitive finding after the fact.
Define what happens when coverage is incomplete
Treat incomplete coverage as an operational state, not an exception to hide. Decide in advance whether the application will pause a workflow, request another review, or display a limited-result notice when a jurisdiction cannot be evaluated. The right choice depends on the use case, but the condition should remain visible to whoever makes the next decision.
For example, a trust-and-safety team reviewing an existing marketplace account might queue an unresolved candidate for manual examination rather than automatically changing account status. Its reviewer should see both the candidate’s match basis and any jurisdiction-level limitations. The interface should discourage copying a candidate’s name into an enforcement note as though identity had already been established.
Applicant screening has an additional boundary. A standard registry search or monitoring result is not a consumer report and should not be used to decide whether to hire or rent to someone. CRAs and screening platforms building those decisions need an appropriate FCRA workflow, including the applicable applicant-facing process. An API integration should not erase that distinction by routing every use case through the same decision button.
Test the workflow with inconvenient cases
Happy-path tests—an exact name and date-of-birth match with every jurisdiction answering—will not reveal the most consequential interface errors. Test a common name with no additional identifier, a candidate with an unavailable birth date, conflicting attributes, no candidates from answering jurisdictions, and a search with a jurisdiction requiring review.
For each case, ask what a reviewer will actually see. Can they distinguish “not available for comparison” from “does not match”? Can they find the jurisdiction and supporting citation? Can they leave a candidate unresolved without the system recording a pass or a confirmed match? Do downstream notifications preserve that uncertainty?
Coverage is useful only when its limits survive the trip from API response to human decision. Offendersearch is one option for teams evaluating registry search with match basis and per-source provenance; its API documentation is a place to assess the integration against those review requirements.
About the author
- The official Offendersearch account. Search US sex offender registries and criminal records by name or location. https://offendersearch.app
Comments
Join the discussionNo comments yet. Be the first to comment.