Which technique involves replacing identifying data with a reversible surrogate?

• A. Generalization
• B. Pseudonymization
• C. Masking
• D. Differential privacy

Answer: (B)

Replacing identifying data with a reversible surrogate captures the essence of pseudonymization. In this approach, direct identifiers such as names or IDs are substituted with pseudonyms or tokens. The mapping between the surrogate and the real identity is stored separately and protected, so authorized users can re-identify if needed. This lets you link records belonging to the same entity across datasets without exposing the actual identity in the data itself.

Generalization and masking operate differently: generalization reduces precision by grouping values into broader categories, which isn’t about reversible linking to individuals. Masking hides or alters values, often without a straightforward way to restore the original data. Differential privacy adds random noise to outputs to protect individual privacy, rather than creating reversible surrogates for identities.

For example, in a medical database, names might be replaced with codes like P-001 while the key that maps P-001 to a real name is kept securely separate, enabling re-identification only under strict authorization.