Skip to content
Outrich Data

Glossary

Re-identification

Re-identification is working out which person a record describes after identifying details have been removed. It usually happens by linking the remaining fields to another source of information, or by recognizing distinctive content. The risk of re-identification is the central measure of whether de-identification has worked.

Three routes account for most re-identification. Linkage matches quasi-identifiers in the dataset against an outside source that includes names. Singling out finds a record that is unique on some combination of fields. Inference deduces a fact about a known person from the group they fall in. Free text adds a fourth route: a message may describe a situation that a colleague or customer would recognize at once.

Risk depends on the recipient as well as the data. A dataset released publicly can be attacked by anyone with any auxiliary data. A dataset licensed to one buyer under a contract that forbids re-identification attempts, with access controls, presents a narrower risk. Sellers reduce risk by removing direct identifiers, generalizing quasi-identifiers, sampling text for review, and binding the buyer by contract not to re-identify. Laws treat the question differently. HIPAA, the GDPR and the CCPA each set their own test for when the remaining risk is low enough.

No dataset of individual records has zero risk. The aim is to reduce the risk to a level the applicable law accepts and to document how that was done, so the decision can be explained later.

Find out if your data qualifies.

Eleven questions, about five minutes, no call required. You get an answer on screen and a written follow-up.

esc