classifying data masking correctly

Is data masking a synonym for anonymisation, or or is one of several methods to anonymise data? I read alot about it but now i got a bit confused. Maybe it's because people use it as a synonym, or it really is.

Data Masking isn't a synonym for Data Anonymisation.

Data Anonymisation is making data such that it can't be uniquely mapped to a given entity.

It is quite difficult because even after most Data Anonymisation techniques have been applied to a given data set, if relationships between certain elements in the data remain, these can be used in conjunction with other non anonymised data sets to associate data with a given entity even when the original data set has no direct means for entity association.

Data Masking is a technique used to attempt Data Anonymisation, specifically an attempt of the Data Perturbation type.

Other techniques used in Data Anonymisation include

• Data Aggregation (Making new summary data from old data that results in loss of entity association precision both directly and via data relationships
• Data Reduction (Removing data from the data set, either on a record basis or variable basis or by re-coding that variable or variable combination to ensure loss of entity association precision both directly and via data relationships)
• Data Perturbation (Modifying the data in the data set)
• Synthetic microdata (Generation of an entire new data set based on the some of the statisical properties of the old data set but with completely synthetic data)

Additional references are available here.