In the judiciary case study, why is "owns a luxury car" treated as a sensitive quasi-identifier?

Because research links luxury-car ownership to criminal behavior, it both narrows down individuals (quasi-identifier) and risks revealing the sensitive outcome.

The scenario: the Judiciary holds records mixing convicted criminals and law-abiding citizens, and wants to share them with an analyst training a model to spot criminal-behavior patterns. The privacy bar is absolute — no one should be able to determine with certainty whether a specific individual was convicted.

"Luxury car owner" is dangerous because domain knowledge says it's a statistically strong predictor of the sensitive attribute (Convicted). That gives it a double role: it narrows the population (few people in a ZIP own one — a strong quasi-identifier) and it correlates with the secret, so leaving it precise leaks the conviction outcome indirectly.

Tip: An attribute's risk depends on domain context. A "harmless" field becomes dangerous the moment it correlates with the thing you're protecting.

Go deeper:

• Data re-identification (Wikipedia) — how correlated attributes leak the sensitive outcome.