Beyond the Registry

For most experts, there is no record to match.

A credible physician sample rests on a primary-source match — an NPI entry, a medical-council listing — and Paper No. 132 sets out how that is done. But the physician is the easy case. The nurse, the pharmacist, the hospital administrator, the payer medical director, and the enterprise IT buyer — the audiences research most often needs — sit in no authoritative, queryable register. For them, "verified" cannot mean "matched," and treating it as though it does is how an unqualified respondent quietly becomes a data point.

Three reasons the registry runs out

First, most professions are not centrally registered — or are registered only locally, by state or country, with uneven public access and coarse specialty detail. Second, B2B identity is not a credential a body issues: it is a role, a level of decision authority, and a set of firmographics, none of which any register holds. Third, even where a register exists, a match proves the credential exists — not that this respondent holds it; a real licence number can be borrowed, bought, or typed by someone it does not belong to.

Fig. 01 — No master key: each box opens only on its own proof. For audiences with no central register, credential confidence is established one converging line of evidence at a time · Photo: Unsplash

No. 132 lives at the top of its own evidence ladder, where a primary source is available. This paper begins precisely where that rung is missing — and asks what "verified" can responsibly mean when no one can be looked up.

Corroboration replaces the lookup.

Where no register settles a claim, confidence comes from convergence — several independent kinds of evidence that are each fallible on their own but hard to fake all at once. Peer-reviewed work on fraud in internet research reaches the same conclusion: there is no single sufficient safeguard, so layered, corroborating checks are required rather than any one test.1

Aim for calibrated confidence — not the illusion of certainty.

Without a register, validation produces confidence, not proof — and more validation is not always better. Past a point, verification adds friction, drives off legitimate respondents, and intrudes on privacy; the research literature treats that balance between data quality and participant burden as a genuine tradeoff to be managed, not maximised.2 The discipline is matching validation depth to what the study is deciding, and stating the confidence actually reached.

"Verified" means corroborated to a stated confidence — not looked up.

For the physician, a registry makes verification a lookup; for everyone else — the partially-registered, the employer-credentialed, and the role-only B2B audience — it is a discipline of converging evidence. Locate the audience on the registry-availability spectrum; corroborate with independent streams rather than trusting any one; treat the result as a calibrated confidence level, not a binary stamp; and validate in proportion to the decision, stating plainly where confidence cannot reach. Do that, and a sample no database can confirm still rests on evidence instead of assertion — the rigour the ICC/ESOMAR Code, the EphMRA Code, and the ISO 20252 framework let buyers ask for in consistent terms.3