Removing names does not make records anonymous: set the de-identification standard in the contract

A 2019 study in Nature Communications estimated that 99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes, so a scrub that only removes names and direct identifiers leaves people identifiable from what remains.

Dated 23 Jul 2019Source checked 15 Oct 2026Page out of reach at the last check, 16 Oct 2026Basis: strong

In the source's own words

“Using our model, we find that 99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes.”

Where it was read

Rocher, Hendrickx and de Montjoye, Estimating the success of re-identifications in incomplete datasets using generative models, Nature Communications 10, 3069 (2019), checked 23 Jul 2019
Everything read from this source (1)Archived copy, 15 Sep 2026

Figures it counts in

It is shown on its own, not counted into a figure.

Cited in these 3 guides

Every record of its kind