Glossary · Measurement and evidence
k-anonymity
The property that any individual record in a published set is indistinguishable from at least k minus one others on the attributes exposed. A hash-prefix lookup gives k-anonymity to the asker, and a publication floor gives it to the sites behind a rate.
What k-anonymity means
The property that any individual record in a published set is indistinguishable from at least k minus one others on the attributes exposed. A hash-prefix lookup gives k-anonymity to the asker, and a publication floor gives it to the sites behind a rate. It can bound how much a single query or row reveals; it cannot protect against an attacker who already holds the other k minus one, which is a limit worth stating beside any published floor.
What k-anonymity can and cannot support
| It can support | It cannot support |
|---|---|
| The property that any individual record in a published set is indistinguishable from at least k minus one others on the attributes exposed. A hash-prefix lookup gives k-anonymity to the asker, and a publication floor gives it to the sites behind a rate. | It can bound how much a single query or row reveals; it cannot protect against an attacker who already holds the other k minus one, which is a limit worth stating beside any published floor. |
Terms this definition uses
hash-prefix lookup · publication floor
Related terms in Measurement and evidence
The vocabulary for saying how strongly something is known.
signed verdict · hash-prefix lookup · publication floor · data snapshot · evidence level · false positive · denominator note · percentile rank · population baseline · cohort asymmetry · self-audit invariant · unscored measurement · dated series · change receipt · regression · finding age · graded through a wall · self-audit · placeholder in record · fetcher disagreement · own-host loop · score version · row review · double-counted defect · cosmetic row · gated row · log file analysis · AI referral attribution · finding fingerprint · finding state · headroom · fix list · render gap · lab data · Lighthouse · RUM · LCP · CLS · INP · topicality · intent coverage · canary · citation rate · presence rate · answer stability
Questions about k-anonymity
What is k-anonymity?
The property that any individual record in a published set is indistinguishable from at least k minus one others on the attributes exposed. A hash-prefix lookup gives k-anonymity to the asker, and a publication floor gives it to the sites behind a rate.
What does k-anonymity not show or guarantee?
It can bound how much a single query or row reveals; it cannot protect against an attacker who already holds the other k minus one, which is a limit worth stating beside any published floor.
Which area of the glossary does k-anonymity belong to?
Measurement and evidence: The vocabulary for saying how strongly something is known.
← hash-prefix lookup · publication floor →
See it in the full glossary · 688 terms across 19 areas.